0% found this document useful (0 votes)
11 views354 pages

Engineering Mathematics

The document is a preface to a book on engineering mathematics. It discusses the need for engineers, especially electrical engineers, to have knowledge of certain branches of mathematics that are not adequately covered elsewhere. It aims to provide supplementary material to address these gaps and enable effective application of mathematics to engineering problems.

Uploaded by

Patrick Horvath
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
11 views354 pages

Engineering Mathematics

The document is a preface to a book on engineering mathematics. It discusses the need for engineers, especially electrical engineers, to have knowledge of certain branches of mathematics that are not adequately covered elsewhere. It aims to provide supplementary material to address these gaps and enable effective application of mathematics to engineering problems.

Uploaded by

Patrick Horvath
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 354

ENGINEERING MATHEMATICS

A S E R I E S O F L E C T U R E S D E L I VE R E D

A T U NI O N C O L L E G E

C H AR LE S P R O TE U S TE I NME TZ , A M
S . .
,
PH D . 0

P AS T R S I N T
P E DE

AM E R I C A N IN ST I TU T
E OF LE C TRI C AL N I N RS
E E G EE

T H I RD DT
E I ION
R E VI S E D AN E N LAR D
D GE

E IG H T H I M R SS I N
P E O

N E W YO R
LOND O N :
""
M C G R AW—H I LL B OO
: 3 70
6 dz 8
C O M PAN Y , I N C
S E VE NT H AVE NUE
B O U VE R I E ST .
, E . c . 4
.

1 917
C O P Y RIG H T , 1 9 1 1 , 1 9 1 5

M C G R A W -H I L L B O O " A ND 1 9 1 7 B Y T HE
,

C O M P A N Y I NC
, .

P RI NT E D I N TH E U N ITE D S TAT S
E OF A M RI A
E C
P R E FA C E T O T H I R D E D I T I O N .

I n p re p aring the third edition o f E ng i ne e rin g M athemati cs ,

besides revision an d corre ction o f the p revio u s text consi dera


,

ble new matter has been added .

The chain f ractio n has been reco gni z ed and disc u ssed as a
convenient method o f n u merical re p resentati on and a pp ro xi
mation ; a para g raph has been devoted to the diO p h ant i c e qu a
tions and a section added on eng ineerin g re p orts disc u ssin g
, ,

the di ff erent pu rp o s es f or w hich en g ineerin g re p orts are made ,

and the corres p on ding character and nat u re o f the re p ort in ,

its bearing on the s u ccess an d reco gnition o f th e e n gin eer s ’

"
w ork .

C H A R L E S PR O T E U S S T E I N M E T Z
.

C AM P M O HAW ,

S eptem b er l s t, 1 9 1 7 .
P R E FA C E T O FI R S T E D I T I O N .

TH E follow ing work emb odies the s u bject matter of a lecture -

c ourse which I have given to the j u nior and s enior electrical


engineering st u dents o f Union University for a number of
years.

I t is generally conceded that a fair knowledge of mathe


maties is necessary to the engineer and especially the electrical
,

engineer For the latter however some branches o f mathe


.
, ,

m at i cs are o f f u ndamental importance as the algebra of the ,

general n u mber the e x ponential and trigonometric series etc


, ,
.
,

w hich are seldom adequ ately treated and o ften not ta u ght at
,

all in the u s u al te x t books o f mathemat ics or in the college


-
,

co u rse o f analyt ic geometry and calc u lu s given to t h e engineer


ing st u dents and there f ore e lectrical engin eers often possess
, , ,

little know ledge of these subj ects As the result an e lectrical


.
,

engineer even if he possess a fair kno w ledge o f mathematics


, ,

may o f ten find d iffic u lty in deal ing with problems thro u gh lack ,

of fam iliarity w ith t h ese b r a nch e s o f mathematics which have ,

become of importance in e lectrical engineering and may also ,

find diffic u lty in looking u p in f ormation on these s u bj ects .

In t he same way the c ollege st u dent when beginning the ,

st u dy of electrical engineering t heory after completing h is ,

general co u rse o f mathematics freq u ently finds himself sadly


,

deficient in the knowledge of mathe matical subj ects of w hich ,

a complete familiarity is requ ir ed for e ffective understanding


of electrical e ng mee rm g theory It was this e x perience which
.

led me some years ago to start the course o f lectures which


is reprod u ced in the f ollowing pages I have thus attempted t o
.

bring together and discuss e x plicitly with numero u s practical


,

applications all those branches of mathematics which are of


,

special importance to the electrical engineer Added thereto .

Vl l
v i ii P R E F AC E .

are a nu mber of s u bj ects wh ich e x perience has shown me


to be important for the e ff ect ive and e x peditious e x ecution o f
e lectrical engineering calc u lations Mere theoretical kn owledge
.

o f mathematics is not su fficient f or t he engineer but it m u st ,

be accompanied by abili ty to apply it and derive results — to


carry o ut n u merical calc u lat ions I t is not su fficient to kno w
.

h ow a phenomenon occ u rs and how it may be calculated b ut


, ,

very o ften there is a wide gap bet w een this knowledge and the
abilit y to carry out the calc u lat ion ; indeed f requ ently an ,

attempt to apply the theoret ical knowledge to derive numerical


resu lts leads even in si mple problems to apparently hopeless
, ,

complicat ion and a lmost endless calc u lation so that all hope ,

of getting reli able results vanishes Th us considerable space .

has been devoted t o the discussion o f methods o f calculation ,

the use of curves and the ir eval u at ion and other ki ndred ,

subj ects requ isite f or e ffe ct ive engineering work .

Thus the follow ing work is not intended as a complete


course in mathematics b ut as s u pplementary to the general
,

college cours e o f mathematics or to the general know ledge o f


,

mathematics which every enginee r and really every educated


man should possess .

I n ill u strat in g the mathematical d isc u ssion practical ,

e x amples usually taken from the field of electrical engineer


,

ing have been given and disc u ssed These are s u ffi ciently
,
.

num erous that any e x ample deal ing w ith a phenomenon


with which the reader is not yet fam iliar may be om itted and
taken up at a later time .

As appendix is given a descriptive o u tline of the intro


duction to the theory of funct ions s ince the e lectrical engineer
,

should be familiar with the general relat ions between the


di fferent fu nct ions which he meets .

ing ,
” “
I n relation t o Theoretical E lements of E lectrical E ngineer
Theory and Calc u lat ion o f Alternat ing Cu rrent Ph e
no m e na ,
and Theory and Calc u lation of Transient E lectric
Phenomena the follow ing work is inte nded as an introduct ion
,

and e x planat ion o f the mathematical side and the m ost e fficient ,

method of study appears to me to start with


,
E lectrical
,

E ngineering Mathemat ics and after e ntering its third


,

chapt er t o take u p the reading o f the first section o f Theo


,

re t ica l E lements and then parallel the study of E le ctrical
,
P R E F AC E .
ix

M E E

lectr ical ”
E ngineering athematics Theoretical lements of ,

E ngineering and Theory


,
and C alc u lation o f A lternat i ng

Cu rrent Phenomena together w ith selected chapters f,


ro m

Theory and Calculation o f Transient E lectric Phenomena ,

and a fter this once more systemat ically go thro u gh all fo u r


,

books .

CH A R L E S P S T E I N M E T Z . .

S C H E N E C

I AD Y ,

N Y . .
,

D e c e m be r ,
19 10 .

P R E FA C E T O S E C O ND E D I T I O N .

I N preparing the second edition o f E ngineering Mathe


m at ics , besides revision and correction o f the previo u s te x t ,

considerable new matter has been a dded m ore partic u larly ,

with regard t o periodic c u rves I n the form er edition the .

st u dy of the wave shapes prod u ced by vario u s harmonics ,

and the recog ni tion of the harmonics from the wave shape ,


hav e no t been treated s i nc e a short discussion o f wave shapes
is given in Alternating Cu rrent Phenomena
,

S ince h o w .

,

ever the periodic f u nctions are the most important in elec


,

t ri cal engineering it appears necessary to consider their shape


,

more e x tensively and this has been done in the new edition
,
.

The symbolism o f the general number as applied to alter ,

nating waves has been changed in con formity to the decision


,

o f the I nternational E lectrical Congress of T u rin a discussion ,

of the logarithmic and semi logarithmic scale of c u rve plot -

ting given etc ,


.

C H A RL E S P S T E I NME T Z . .

De ce m be r 1 9 1 4 ,
.
CO NTE NTS .

PR E FA C E

C HAPTE R I . THE GE NE R AL NUMBE R .

A TH E SY S T E M O F NU M B E R S
. .

1 Additi on a nd S u btracti on
. . O ri g innu m b e rs Count ing a nd
Of .

m e a su ri n g . Additi on . S u btracti on a s re ve r se O pe rati o n o f


addi ti on . .

2 . L i m itati on o f s u btracti on S u bdivi si o n o f t h e ab s o l u t e nu m b e r s


.

i nt o pos itive and ne gative .

3 . Ne gative nu m b e r a m a the m atica l c on c e pti on l i ke t h e i m agi n ary


nu m be r C a se s w he re t h e ne g ative num b e r h a s a ph ys ica l
.

m e an i ng a nd ca se s whe r e it h as no t
,

4 . Mu lti p lic ati on a nd D ivi s i on Mu lti plicati o n as m u lti p l e addi .

ti on D ivi si o n a s it s r e ve r se O pe r ati on L i m itati on o f divi


. .

s i on

5 . Th e fracti on a s m ath e m atica l c on c e pti on Ca se s w h e re it h as a .

phys ica l m e ani ng a nd ca s e s w h e r e it h as no t


,

6 . I n v o l u ti o n a nd E v ol u ti on I n v o l u ti on a s m u lti p l e m u lti p lica


.

ti on E vo l uti on a s it s re ve r se O pe rati on Ne gative e x po


. .

ne nt s .

7 . Mul ti ple i nvo l u ti on le ad s t o no ne w ope rati on .

8. Fra cti ona l e x pone n t s .

9 . I rrati on a l Num b e r s L i m itati o n o f e v ol u ti on


. E ndl e ss d e ci m a l .

fracti on Rati onality o f t h e irrati onal n um b e r


. .

10 .
Q u adrat u re nu m b e r s Mu lti ple val ue s o f r oo t s S q u ar e r oo t o f
. .

ne gative q uantity r e p re s e nting q uadrat ure nu m b e r o r r o ta ,

ti on by 90 °

11 . Co m pari s on o f po sitive ne gative a nd q u adrat u re nu m be r s


,
.

Re alit yo f q u adrat u r e n u m b e r Case s w h e re it h a s a phys ical.

m e an in g a nd c as e s w h e r e it h as no t
,

12 . Ge ne ra l Nu m b e r s R p re se n ta ti on O f t h e p lane by t h e ge ne ra l
. e

num be r I t s re lati o n t o r e ctangular c oo rdi nate s


.

13 . L i m itati on O f al ge bra by t h e g e ne ral nu m b e r Roo t s o f t h e un it . .

Num be r O f s u c h r oo t s a nd the ir re lati on ,

14 . The t w o r e ve r se O pe r at i ons o f i n vol u ti on


C ON TE N TS .

P A GE

15 . L ogarith m at io n . Re l at i on
b e tw ee n logari thm and e x pone nt o f
i nvol u ti on . Re d u c t i on to o the r ba se L og arith m o f ne g ativ e
.

q u anti ty
1 6 Q u at e rn i ons
. Ve c to r calcu l u s o f sp ace
.

1 7 Sp ac e r oto r s and the ir r e l ati on


. Supe r al ge braic n at u re o f sp ac e
.

a nalys i s .

B AL G E B R A O F T H E GE N E R AL NU M B E R O F C O M P L E X Q U A N TIT Y
. .

Re c tangul ar a nd Po lar Coo rdi n a te s


Pow e r s o f j Ordi n ary o r r e al and q u a d rat u r e o r i m agi nary
.
,

nu m be r Re lati ons
. .

1 9 Con c e pti on o f ge ne ral nu m b e r b y p o i n t o f p lane in r e ctang u l ar


. .

c oo rdin ate s ; i n pol ar c oo rdi nate s Re lati on b e twe e n re ct .

angul ar a nd po lar fo r m
Additi on a nd Su b tracti on Al ge braic a nd ge o m e trical additi on
.

and s u b tracti on Com binati on a nd r e s o l u ti on by parall e lo


.

gram law
2 1 D e notati ons
.

2 2 S i g n o f v e c t o r ang l e
. Conju gate and ass o ciate num be r s Ve c
. .

to r analysi s
2 3 I n st an c e o f s te am path O f t u rbi ne
. .

2 4 Mu lti pl i c ati on
. Mu lti pl ic ati on in r e ctangu lar c oordi nate s
. .

2 5 Mu lti p li c ati on in p ola r c oo r d i nate s Ve ct o r and ope rat o r


. .

2 6 Phys ica l m e an i ng o f r e su lt o f al ge braic O pe rati on


. Re p r e se n ta .

ti on o f re su lt
2 7 L i m itati on o f app licati on o f a lge braic O pe r ati ons t o p hys ic al
.

q u anti ti e s and o f th e graphical r e pre se ntati on o f t h e re sul t


, .

G ra p hic al r e p r e s e nt ati on o f al ge braic O pe rati ons b e t w e e n


c u rre nt v ol tage a nd i m pe dan ce
,

Re pr e s e n tati on O f v e c to r s a nd o f O pe rat o rs
2 9 D ivi s i on
. . D ivi si o n in r e ctangular c oo rdinate s
30 D ivi s i on in pol ar c oo rdi n ate s
.

3 1 I nv ol uti on a nd E v ol u ti on
. Use o f p o la r c oo rdi n ate s
.

3 2 Mu lti ple va l ue s o f t h e r e su l t o f e v o l u ti on T h e ir lo c a
. ti on i n t h e .

pl ane o f t h e ge ne ra l nu m b e r Polyp h ase a nd 77 ph a se syste m s


. .

O f nu m b e r s

33 The n va l u e s o f V1 a nd th e ir r e lati on
. .

3 4 E v ol u ti on i n r e ct angul ar c oo rdi nate s


. Co m pl e x ity o f r e su lt
.

35 Re d u cti on o f p r o d u ct s a nd fra c ti ons o f ge ne ral n u m b e r s b ypo lar


.

re pre se ntati on Instanc e .

36 E x pone ntia l r e p re se n ta t i on s o f ge ne ra l n u m b e r s
. Th e di ffe re n t .

fo rm s o f t h e ge ne ral nu m b e r
37 I n st an c e o f u se o f e x pone n tia l fo r m i n s o l u ti on o f di ffe r e n tia l
.

e q u ati on
CON TE N TS .
xii i

P A
GE

38 . L ogarit hm at io n . Re s ol u ti on of th e logarithm Of a ge ne ra l
nu m b e r

C HAPTER II . THE POTE NTI AL SE R I ES AND EXPONE NTI AL


FUNC TI ON .

A GE N E R AL
. .

39 . i nfinite se rie s o f powe rs o f x


Th e .

40 App r ox i m a ti on by se ri e s
.

4 1 Al te rnate a nd o ne —s id e d a pp r ox i m ati on
.

42 Conv e rge nt and dive rge nt se rie s


. .

43 R ange o f c on ve rge n c y
. Se v e ra l s e ri e s o f d i ffe r e nt ran ge s f o r
.

s am e e x p r e ss i on

44 D i s c u ss i on o f c on v e rge n c y i n e ng i ne e ring a pp licati on s


45 Us e o f se ri e s f o r app r ox i m a t i on O f sm al l te r m s
. I n s tan c e o f .

e l e ctric c ir c u it

46 B i no m i al t heo r e m for d e v e l op m ent i n se ri e s


. I ns t a nc e o f i n .

d u c t i ve c irc u it .

47 Ne ce ss ityo f d e ve l o pm e n t in se ri e s
. I nst an c e o f a re o f hype rb ol a
.

48 I ns tan c e o f n um e rical c a lc u l at i on o f lo g ( 1 + 27)


.

B D IFF E RE N T I AL E Q U A T I O N S
. .

49 C haracte r o f m o s t di ffe r e n tia l e q u a t i o ns o f e l e ctri c a l e ngi ne e ri ng


. .

The ir typ ic al fo r m s
d
£ =
y . Sol ut i on by r
se i e s , by m e tho d o f Inde te rm i nate co

e ffi cie nts .

= az . Sol uti on by i nde te r m inate c o e ffici e nts


I nte grat i on constant a nd te rm i nal c onditi ons
I n v ol uti on o f s ol u ti o n E x pone nt i al f un c ti on
.

I ns tance o f ri se o f fie l d c u rr e n t in d ir e c t c u rr e n t shun t m o t o r
E va l u ati on o f i n d u ctance a nd n u m e ri c a l c a l c ul ati on
,

I ns tance o f c on d e ns e r d i s char ge t hr o ugh re si st an c e

Solu ti on of
g ay by i nde te r m inate co e ffic i e nts by e x pone nt ia l,

f unct i on .

Sol u ti on b y t ri gono m e tric f uncti ons


Re l ati on s b e tw e e n t ri g ono m e tric f unct i on s and e x pone nt ia l f unc
t i ons w i th i m aginary e x pone nt and in ve r se ly ,

I ns tance o f c on d e n se r d i s ch arge thr o ugh i n d u c tan c e Th e t w o .

i nte grati on constants and te rm in al cond iti ons


E ffe ct o f re si st ance o n t h e d i scharge Th e ge ne ra l di ffe r e n tial
.

e q u a t i on .
xiv CON TE N TS .

P A
GE

Sol u t i on
o f t h e ge ne ra l diffe r en ti al e q uati on by m e ans o f t he
e x pone n tia l fun cti on b v t h e m e th o d o f i nde te r m in ate
,

I ns tan c e o f c ond e ns e r d i s charge t hr o u gh r e si s tanc e a nd in d u c


tan ce E x pone ntial solu ti on a nd e val u ati on o f constants
. .

I m a gi n a ry e x pone n t s o f e x pone nt ial f un cti on s Re d u c ti on to .

tri gono m e tric functi ons The O s cillatin g func ti ons


.

E x pl anat i o n o f t ab l e s o f e x pone nt i al f unct i on s)

C HAPTER III . TRIGO NOME TR I C SE RIES .

A TR IG O N O M E T R I C
. FU N C TI O N S .

66 . De finiti on o f tri gono m e tric f un cti ons o n circ le a nd ri gh t triangle


67 . Si gn O f f unc ti on s i n di ffe r e nt q u adra nt s
68 . Re lati on s b e t w e e n s in c o s t an a nd c o t
, ,

69 . Ne g ati ve su pple m e nta ry a nd c o m p l e m e nt a ry angl e s


,

70 . An gle s ( x i x ) a nd

71 . Re l ati o ns be t w e e n tw o ngle s
a ,
a nd b e tw e e n a ng l e a nd d o u b le
angle
7 2 D i ff e r e n tiati on and i n te gr a t i on o f tri g ono m e t r ic f un cti on s
.

De fi nit e i nte gral s


7 3 Th e bi no m ial r e l at i on s
.

7 4 Polyp ha se r e lati on s
. .

7 5 T ri g ono m e tric f o rm ul a s O f t h e triangl e


.

B T RIG O N O M E T R I C S E R I E S
. .

76 Co nst an t tra ns i e n t a nd pe ri o di c phe no m e n a


.
,
Un ival e n t pe ri .

o d i c f unc ti on re pr e se nte d b y tri g ono m e tric se ri e s .

7 7 Al te rn atin g s ine w av e s and d i s t o r te d w av e s


.

7 8 E va l uat i on o f t h e Con s t ant s fr o m I n stantane o u s Va l ue s


. Cal .

c ul ati on o f c o nstant te rm o f se ri e s
7 9 C a l c ul ati on O f co s co e fficie nt s
.
-

80 C alc ulat i on o f s in c o e fficie nt s


.
-

8 1 I n stan c e o f calc ulati ng 1 1 t h h ar m o nic O f ge ne rat o r w ave


.

8 2 D i s cu ss i on
. I ns tance o f c o m p l e t e calc ul ati on o f p u l s ati ng c u r
.

r e nt w ave
83 Alte rn ati ng w ave s as s vmm e t ric a l w ave s
. Calc ul ati on O f sym .

m e trica l w a ve
8 4 Se p arati on o f O dd a nd e v e n har m on ic s a nd o f c on s ta nt t e r m
.

85 Se parati on o f s i ne a nd co s i ne co m pone nt s
.

8 6 Se p arati on o f w av e i nt o con s tant te rm a nd 4 c o m pone n t w ave s


.

8 7 D i s c u ssi on o f ca l c ul ati on
.

88 Me c h an i sm O f ca l cul ati on
.
C ON TE N TS . xv

89 . o f r e soluti on o f th e annual te m pe rature curve


I nstance
9 0 Con s tan t s a nd e q u ati o n o f te m pe rat u re w av e
.

9 1 D i s c u ssi on o f te m pe rat u r e w av e
.

C R E D U C T I O N O F TR IG O N O M E TR I C SE R I E S B Y PO LY P H A S E RE LA T I O N
. .

9 2 Me th o d o f s e p arating ce rtai n cla s s e s o f har m on ic s and it s


. ,

li m itati on
9 3 I ns ta nc e O f s e parating t h e 3 d a nd 9 t h h ar m o nic o f tran s f o rm e r
.

e x citi ng c u rr e nt

D CALC U LA TI O N O F TRIG O N O M E TR I C SE R I E S F RO M O T H E R TR IG O N O
.

M E T R I C SE R I E S .

9 4 I nstan c e o f c a lc ul ati ng c u rre nt in lon g di s tan c e tran sm i ssi on l i ne


.
,

d u e t o di sto r te d v ol ta g e w av e o f ge ne r ato r L i ne con s tan t s


. . 1 39
9 5 C irc u it e q u ati on s a nd c a lcul at i o n o f e q uati on o f c u rr e nt
.
, 1 41
9 6 E ff e cti ve val u e o f c u rre n t a nd c o m p a ri s on w ith t h e c u rr e n t
. ,

p r o d u c e d b y si ne wave O f vol tage . 1 43


9 7 Vol ta ge wa ve o f r e act ance in ci rc u it o f th i s d i s t orte d c u rr e nt
. 1 45

C HAPTE R IV . MAXI MA AND MI NI MA .

98 . Max i m a a nd m in i m a b y c u rve plo tting I nstance o f m agne ti c


.

r
pe m e a bi l i ty . Ma x i m u m po we r fact o r o f i nd u cti on m oto r a s
fun cti on o f lo ad
99 . I n te rpol at i on O f m a x i m u m v al u e i n m e t ho d o f c u rve plot ting .

E rr o r i n c ase o f unsym m e tri c a l c u rv e I nstan c e o f e ffic i e n c y


.

o f s te am t urbine no zzle D i s c u ssi on


- .

1 00 . Mathe m atical m e th o d Max i mum m i ni m um and infle x io n


.
,

po i nt . D i s cussi on
101 . I ns tan c e : Spee d o f i m p ul se t urbi ne whe e l f o r m a x i m um
e ffici e ncy . Cu rre nt i n transfo rm e r f o r m a x i m u m e ffi ci e nc y .

1 02 . Effe ct Of i nte r m e d iat e variab le s I nst an c e : Max i m um p owe r


.

in re si st an c e shunti n g a con s tan t r e si s t an c e i n a c o ns tant c ur


re nt cir cu it
1 03 . Sim pl ificati on o f c a l c ul at i on by s uppr e ssi on o f unne c e ssary te r m s
e tc. Inst an c e
1 04 . I ns tance : Ma x i m u m no n i n d u ctive l o ad o n i n d u ctive t ran sm i s
-

s i on line . Ma x i m u m c u rr e n t in li ne
1 05 . D i s c u ssi on O f physic al m e aning o f m ath e m atical e x t re m um
Instance
1 06 . I nstance : E x t e rn a l r e actance givi ng m a x i m u m O u t pu t o f a lt e r
nato r at c onstant e x te rnal r e s i s tan ce a nd c onstan t e x citati on
D i s cussi on
1 07 . Ma x i m um e ffic i e n cy o f al te rn at o r o n no n in d u c t ive l o a d D is
-
.

c u ss io n O f physi c a l li m it ati ons


xv 1 C ON TE N TS .

1 08 . E xt re m a w i th ve r al i nd e pe nde nt variab le s Me tho d o f m at h


se .

e m at i ca l c a l c ulati on a nd ge o m e trica l m e an i n g
,

1 09 . Re si s tan c e a nd r e acta n c e o f l o ad t o g ive m a xi m um o u tp u t o f


t ransm i ssi on l ine at c ons t nt supp ly v o lta ge
,
a

1 10 . Di s cu ssi on o f physi c a l li m itati ons


111 . D e te rm i na i on o f e x tr e m a b y plo tting c u rve O f di ff re ntial qu o
t
e

ti e nt I ns tanc e : Max i m a O f c u rr e n t w av e o f alte rnat o r O f


.

di s t o rte d v olt age o n transm i ssi o n l ine


1 12 . Grap hical calc u lati o n o f di ffe r e n tia l c u rv e o f e m p iric al c u rve
fo r d e t e r m ining e x tr e m a
1 13 . I ns tan c e : Ma x i m u m pe r m e ability calc ulati o n
1 14 . Gr o upi ng o f batte ry c e ll s f o r m a x i m u m po w e r i n con st an t r e si s t
an c e
1 15 . Vol ta ge o f tran sfo r m e t o give m a x i m u m o u t pu t at cons tan t
r

l o ss
1 16 . Vo ltage of transfo r m e r at c ons tan t o u t p u t t o g ive m a x i m um
, ,

e ffici e ncy at f ull lo ad at half lo a d


, .

117 . Ma x i m u m val ue o f char gi ng c u rre n t o f c on d e nse r thr o ugh


i nd u ctive circ u it ( a) at lo w r e si s tan c e ; ( b ) at hi gh r e s i s tanc e .

1 18 . At wh at o u t pu t i s t h e e ffi c i e ncy o f a n i n d u c ti on ge ne rat o r a m a x

1 19 D i s c ussi on o f physical li m i tati ons Ma x i m um e ffi ci e n cy at co n


. .

s tan t c u rr e n t o u t p u t .

1 2 0 ME T H O D O F L E AS T S Q U A R E S
. Re l ati on O f n u m b e r o f O b s e rv a
.

t i ons t o n u m b e r O f c on s t ant s D i s c ussi on o f e rr o rs O f


.

o b se rvati on
1 2 1 Pr o babilit y ca l c ul u s a nd t h e m in i m u m s u m o f s q u are s o f t h e
.

e rr o r s .

1 2 2 Th e di ffe r e n tia l e q u ati o ns o f t h e s u m o f l e a s t s q u ar e s


.

1 2 3 In stan c e : Re d u cti on o f c u rve o f p ow e r o f i nd u cti o n m o t o r


.

r unni ng l i ght i nto t h e c o m pone nt lo sse s D i s c u ssi on o f


, .

re su lt s

I Z3A D i o p h a nt i c e q u a t i o n s
.

C HAPTER V . METHOD S O F APPR OXI MATION .

Fr e q ue ncy o f sm all quantiti e s in e le c tric al e ngi nee ri ng pr o b le m s .

I n s tan c e s .Appr ox i m ati on b y dr o pping te r m s o f hi ghe r O rd e r


Mu lti pl ic ati on o f t e r m s w it h sm all q u an titi e s
I ns tance o f c al c u l ati on o f po w e r o f dir e ct c u rr e n t s h unt m o t o r
Sm a ll q u antiti e s i n d e no m i n at o r o f fra cti ons
I ns tan c e o f c a lc ul ati on o f i n d u c ti on m o t o r c u rr e n t a s f un cti on
,

o f sli p
C ON TE N TS . vi i
X

P A GE

1 2 9 Use
. o f binom i al se ri e s in app r ox i m ati ons o f powe rs a nd r oo t s
a nd i n n u m e ric al cal c u lati ons

1 30 . I ns tan c e O f calc ul ati on o f c u rre n t in a lt e rn ati ng circ u it o f lo w


i nd u ctance I nstanc e o f cal cu l ati on o f s h o rt circ u it c u rre nt
.

o f alte rnato r a s f uncti on o f spee d


,

1 31 . Use o f e x pone ntia l se ri e s a nd l o g arith m ic s e ri e s i n app ro x im a

r i ati ons o f tri gono m e tric f unc t i ons


1 3 2 App ox m
.

1 33 Mc L a u rin s a nd T ayl o r s se ri e s i n app r ox i m ati ons


’ ’
.

1 34 T ab u lati on o f vari o u s i n finit e s e ri e s a nd o f t h e a pp r ox i m a t i ons


.

d e rive d from the m


1 35 E s ti m a t i on o f ac c u r acy o f a pp r ox i m ati on
. App licati on t o .

sho r t c ir c u i t c u rr e n t o f alt e rn at o r 2 00
1 36 E x p re ssi on s w hi ch a re a pp r ox i m ate d b y ( 1 + 3 ) a nd by ( 1 — s )
. 201 .

1 37 Math e m atical i ns tan c e O f a ppr o x i m a ti on


. 2 03
1 38 E Q U A T I O N S O F T H E TR A N SM I SS I O N L I N E
. I nt e g rati on o f t h e
.

di ffe re nti al e q u ati o ns 2 04


1 3 9 S u b s ti tu ti on O f t h e t e r m i nal c on d iti o ns
. 2 05
1 40 The app r ox i m at e e q u ati ons o f t h e tra n sm i ssi o n li ne
. 2 06
1 4 1 N u m e ri cal i ns t ance
. D i s cu ss i o n o f a c cu r a cy o f a p p r o x i m a t i o n 2 0 7
. .

1 4 1 A App rox i m at i o n b y cha i n fra ct i o n


. . 20 8
1 4 1 B App r ox i m at i on b y cha i n f ract i on
. 2080

C HAPTE R VI . E MPI RI C AL C URVE S .

A G E N E R AL
. .

1 42 . Re lati on
b e twe e n e m pirical c u rve s e m p irical e q u ati o ns and
,

rati onal e q u ati ons


1 43 Physical n at u re o f p h e no m e no n
.
Po i n t s at ze r o a nd at i n fin it y
. .

Pe ri o di c o r no n pe ri o dic Constant te r m s Ch ange o f c u rve


- . .

law Sc al e
.

B NO N PE R I DI C C U RVE S
.
- O .

1 44 Po te ntial Se ri e s
. I nstance o f c o r e lo ss c u rv e
.
-

1 45 R ati on a l and irrati on al u se o f p o te ntial s e ri e s


.
I ns tan c e O f fa n .

m o t o r t o rq u e L i m itati on s o f po te n tial se ri e s
.

1 46 P A R A B O LI C AN D H Y P E R B O L I C C U RVE S
.
Vari o u s s hape s o f p ara
.

b ol as and o f hype rb ol a s
1 4 7 T h e c hara c t e ri s tic o f p arab o li c and hype rb o li c c u rve s
.
I t s u se .

a nd li m itati on b y c on s tan t te r m s .

1 48 Th e l o garith m ic ch aract e ri s tic


.
I t s u s e a nd li m itati o n
.

1 4 9 E X .
P O N E N T I A L AN D L O G A R I T HM I C C U R V E S Th e e x p on e n ti al
.

f uncti on
1 50 : C haract e ri s tic s o f t h e e x pone ntial c u rv e th e ir u s e a nd li m itati o n
,

by c onstant te rm Co m pari s on O f e x pone nti al c u rve and


.

hype rb o la
151 .
Do ub le e xponenti al f uncti ons Vari ous shape s the reo f .

1 52 . E VAL U A TI O N O F E M P I R I C AL C U R VE S Ge ne ral p ri n ciple s o f .

i nve sti gati on o f e m p iri c al c urve s


153 . I ns tance : Th e v ol t am pe r e characte ri s ti c o f t h e t ungste n lam p
-
,

re d u ce d to p arab ol a w it h e x p one n t R ati onal i z e d b y


r e d u cti on to radi ati on law
1 54 . T h e vol t am pe r e c h aracte ri s ti c o f t h e m a gne tite are re d u ce d
-
,

t o hype rb ol a w ith e x pone nt


155 .
C hange o f e l e c tric c u r re nt with c h ange o f cir cu it c ond iti ons
re d u c e d to d o u b le e xp one ntia l f unc ti o n o f tim e

1 56 . R ati onal r e d u ct i on O f c o r e lo s s c urve o f p ar ag ra ph 1 44 b y


-
,

p ar a b ol a w it h e x pone nt

157 . R e d u cti on o f m a gne ti c c hara c t e ri s ti c f or h i ghe r d e n s iti e s t o, ,

hyp e rb ol i c c urve I ns t anc e o f th e in ve st iga t i on o f a hys


.

t eres is c urve o f s ili c on s te e l


C . PE R I O DI C C U R VE S .

1 58 . D is tortio n of s i ne ave b y har m o nic s


w
159 . T hird nd a fifth harm o nic Pe a k m u lti p l e
.
p e ak fl t t p a nd, , a o

s a wtooth

1 60 . C o m bi ne d e ff e ct of t h ird a nd fifth har m o nic


161 . E v e n har m o ni c s Une qu al s ha p e and l e ng th of half w aves
.

C o m bi ne d s e co nd a nd third har m o nic


1 62 . E ff e ct of hi g h har m o ni c s
1 63 . Ri p p l es a nd nod e s cau se d b y hi g h e r har m o nic s I nc m m e n . o

s ur a b le w av es

C H APT E R VII . NU M E RI C AL C AL C UL ATI O N S .

1 64 . M E TH O D OF C AL C U LAT I O N T ab u lar for m


. of calc u latio n
1 65 I ns ta nc e of
. tr ans m iss io n li ne r e gu latio n
1 6 6 E X A CT NE SS O F C AL C U L A T I O N
. . n r e es of e x m t ne s s : m gn a it u de ,

ap p ro xi m ate , e xact
1 67 Nu m b e r of d e ci m als
.

1 68 I N T E LL I GIB I LIT Y O F E N G I N E E R I N G D A T A
. . C u rv e p lotti ng for
ho wi ng s hap e of f u nctio n and for re cord
s ,
of nu m e rical val u es 2 8 3
1 6 9 S cal e of c urves
. Pr inci p l es. . 2 86
1 70 L g arith m ic a nd s e m i log arith m i c p a p e r a nd i t s p ro p e r
. o - e us 287
1 7 1 C o m p l e t e ness of r e cord
. 2 90
1 7 1 A E ng i ne e ring R e p ort s
.

1 7 2 R E LI AB I LIT Y O F N UM E R CI AL C AL C U LAT I O N S
.
Ne c e s ity o f r lia . s e

b i lit y in e ng in e e i ng calc latio ns r u .

1 7 3 M e tho d s of ch e c k i ng calc u latio ns


. C u rv e p lotti ng .

1 7 4 So m e fre qu e nt e rro rs
.

AP P E ND IX A . N O TE S O N THE T H E O RY O F FUNC TI O N S .

A . G E N E RAL FU N C TI O N S .

1 75 . I m p licit analyt ic f u nctio n . E p x licit a nal ytic f u nctio n . Re ve rse

f u nct i o n
E NG I NE E R I NG MAT H E MAT I C S .

C HAPTE R I .

TH E GE NE RAL NUMB E R .

A THE SYSTE M O F NUMB E R S


. .

Additi o n an d S u b tr a c t i o n .

I . From the operation of counting and meas u ring arose the


art of figuring arit h metic algebra and finally more or less
, , , , ,

the entire stru ct u re of mathematics .

Du ring the development of the h u man race t h ro u gho u t the


ages which is repeated by every child d u ring the first years
,

o f life the first conceptions o f n u merical values were vag u e


,

and crude : man y and f ew big and little large and small
, ,
.

L ater the ability to count that is the kno w ledge o f n u mbers


, , ,

developed and last of all the ability o f meas u rin g and even
, ,

up to day me as uring is to a considerable e x tent done by co unt


,

ing : steps knots etc , ,


.

From co unting arose t h e simplest arithmetical operati on


a ddi ti on Thus we mav co u nt a b u nch o f horses :
.

1 , 2 , 3 , 4, 5 ,

and then co u nt a second b u nch Of horses ,

1 , 2, 3 ;

now put the second bunch together w ith the first one into one ,

bunch and co u nt them That is after co u nting t he h o rs e s


,
.
,
2 E N GINE E R I N G M A THE M ATI CS .

o f the first b u nch we conti n u e to co u nt those of the second


,

bunch thus :,

1 ) 2 ) 3 ) 4 ) 5 1— 6 1 7 ) 8 ;

which gives addition ,

or in general
, ,

a + b = a

We may take a w ay again the second bunch of horses t hat ,

means we co u nt the entire b u nch of horses and then co u nt


, ,

o ff those we take away th u s :


1 , 2 , 3 , 4, 5 , 6 , 7 , 8 7 , 6, 5 ;

which gives su b tra cti on,

8
or in general
, ,

c
-
b = w

The reverse of putting a gro u p of thi ngs together with


another gro u p is to take a gro u p away therefore s u btraction ,

is the reverse of addition .

2 I mmediatel y we n otice an essential di fference between


.

addition and s u btraction which may be illu strated by the


,

following e x amples :

Addition : 5 horses + 3 horses gives 8 horses ,

Subtraction : 5 —
horses 3 horses gives 2 horses ,

Addition : 5 horses + 7 horses gives 1 2 horses ,



S u btraction : 5 horses 7 horses is impossible .

From the above it follows that we can al w ays add but we ,

cannot always s u btract ; s u btraction is not al w ays possible ;


it is not when the n u mber of t hings which we desire to sub
,

tract is greater than the number of things from which we


desire to s u btract .

The same relation O btains in measuring ; we may meas u re


a distance from a starting point A ( Fig f or instance in s t eps .
,

and then measure a second distance and get the total distance ,

fro m th e s t art in g p oi nt b y addition : 5 steps from A to B . .


TH E GE NE R AL N U MBE R . 3

then 3 steps from ,


B to C , gives the distance from A to C as ,

8 steps .

5 steps + 3 steps 8 steps ;

1 2 3 4 5 6 7 8

B C
FI G 1 . Additi o n .

or we may step O ff a distance and then step back that is


, , , ,

subtract another distance f or instance ( Fig , .

5 st eps —
3 steps = 2 steps ;
that is going 5 steps from A to B and then 3 steps back
, , , ,

from B to C b ri ngs u s to C 2 steps away from A


, , .

1 2 5

C 8
FI G 2 . . Su btraction .

Trying the case o f s u btraction w hich was impossible in the ,

e x ample with the horses 5 steps 7 steps = ? We go f rom the ,


-

starting point A 5 steps to B and then step bac k 7 steps ;


, , , ,

here we find that sometimes we can d o it sometimes we cannot ,

do it ; i f bac k o f the starting p oint A is a stone wall we cannot ,

step bac k 7 steps I f A is a chalk mark in the road we may


.
,

step ba c k beyond it and come to C in Fig 3 I n the latter case


,
. .
,

1 16 . 3 . S ub t r ction
a . Ne gative R e s u lt .

at C we are again 2 steps distant f rom the starting point j ust ,

as in Fig 2 That is
. .
,

5 ( Fig .

5 ( Fig .

In the case where we can s u btract 7 from 5 we get the same ,

distance from t he starting point as when we subtract 3 from 5 ,


4 E NGIN E E RING M A TH E M ATI CS .

b ut t he distance AC in Fig 3 w h ile the s a me 2 steps as .


, , ,

in Fig 2 is di fferent in character the one is to w ard the le ft


.
, , ,

the other to w ard the right That means we have two kinds .
,

o f distance u nits t h ose to the right and those to the left and
, ,

have to find some w ay to distinguish them The distance 2 .

in Fig 3 is to w ard the left of the starting point A that is


.
, ,

in that d i rection in w hich we step when subtracting and


, ,

it th us appears nat u ral t o distinguish it from the distance '

2 in Fi g 2 by calling the f ormer 2 while we call t h e dist ance


.
,

,

AC in Fig 2 : + 2 since it is in the direction f rom A in which


.
, ,

w e step in addi ng .

This leads to a s u bdivision o f t h e s vste m o f absolute n u mbers ,

1, 2,

into two classes positive n u mbers


, ,

+ 1, + 2, + 3,

and negative n umbers ,

— 1, —
2, —
3,

and by the introduction of negative n u mbers we can always ,

carry o ut the mathematical operation o f s u btraction :


c
— b=%

and i f b is greater than c a merely becomes a negative number


, , .

3 We m u st there f ore reali ze that the negative number and


.


the negative unit 1 is a mathematical fiction and not in
, , ,

u niversal agreement with e x perience as the absol ute n u mber ,

fo u nd in the operation o f counting and the negative number ,

does not alw ays represent an e x ist ing condition in practi cal -

e x peri ence .

I n the application of n u mbers to the phenomena of nature ,

we sometimes find conditions w here w e can give the negative


number a physical meaning e x pressing a relation as the ,

reverse to the positive n u mber ; in other cases we cannot do



7 = —2
this For instance 5 horses
. horses
,
horses has no
physical meanin g There e x ist no negative horses and at the
.
,

best we could only e press the relation by saying 5 horses 7


x —
,

horses is impossible 2 horses are missing ,


.
TH E G E N E R AL N UM B E R . 5

In the same way an ill u mination o f 5 f oot candles lowered


,
-
,

by 3 foot candles gives an ill u mination of 2 f oot candles thus


-
,
-
, ,

5 f oot candles
- —
3 f oot candles = 2 f oot candles
- -
.

If it is tried to lo w er the ill umination o f 5 f oot candles by 7 -

foot candles it will be f o u nd impossi ble ; there cannot be a


-
,

negative illumination of 2 foot—candles ; the limit is z ero ill u mina


tion or darkness
,
.

From a stri ng o f 5 f eet len gth we can cut O ff 3 feet lea ving , ,

2 feet b u t we cannot c ut o ff 7 feet leavi ng — 2 feet o f string


, , .

I n these instances the negative n u mber is meaningless , ,

a mere imaginary mathematical fiction .

If the temperat u re is 5 deg cent above f ree zing and falls . .


,

3 d e g it wi ll be 2 deg cent above f ree z ing


,
I f it f alls 7 deg
. . . .

it w ill be 2 deg cent belo w f ree z ing The one case is j ust as
. . .

real physically as the other and in this instance we may


, ,

e x press the relation th u s :


deg cent —3 deg cent deg cent
+5 . . . . +2 . .
,


+5 deg cent . . 7 deg cent . .
-
2 deg cent. .
;

that is in temperature meas u rements by the conventional


,

temperat u re scale the negative n umbers have j u st as m u ch


,

physical e x istence as the positive n umber s .

The same is the case w ith time we may represent future ,

time f rom the present as starting point by positive numbers


, , ,

and past time then will be represented by negative n um bers .

But we may eq u ally well represent past time by positive num


bers and f ut u re times then appear as negative n u mbers I n
,
.

this and most other ph ysical applications the negative n u mber


, ,

t h u s appears equivalent with the positive n um be r and inter ,

changeable : w e may choose any direction as positive and ,

the reverse direction then is negative Mathematically ho w .


,

ever a di fference e x ists between the positive and the negative


,

number ; the positive unit m u ltiplied by itself remains a pos , ,

it ive unit b u t the negative unit multiplied w ith itsel f does


, , ,

not remain a negative unit but becomes positive : ,

and not
6 E N GIN E E RIN G M ATH E M ATI CS .

S tarting from 5 deg northern latitude and going 7 deg . .

so u th brings us to 2 deg southern latitude which may be


,
.
,

e x presses thus ,

deg latit u de —7 deg latit u de —


+5 . . 2 deg latitude
. .

There f ore in all cases where there are two opposite dire c
, ,

tions right and le f t on a line north and so u th latitude east


, , ,

and west lon gitude f uture and past assets and liabilities etc
, , ,
.
,

there m ay be appli cation of the negative numbe r ; in other cases ,

where there is o nly one kind or direction co u ntin g horses , ,

meas u ring ill u mination etc there is no physical meaning , ,

which wo u ld be represented by a negative number There .

are still other cases where a meaning may sometimes be f ound


,

and sometimes not ; for instance i f we have 5 dollars in o u r ,

pocket we cannot take a w a y 7 d ollars ; i f we have 5 dollars


,

in the bank we may be able to draw out 7 d ollars or we may


, ,

not depending on o ur credit I n the first case 5 dollars


,

7 .
,

dollars is an impossibili ty whi le the second case 5 do llars 7 ,


-

dollars 2 dollars overdraft .

I n any case ho w ever we m u st reali z e that the negative


, ,

num ber is not a physical but a mathematical conception , ,

which may find a physical representation or may not depend , ,

ing on the physical condi t ions to which it is applied The .

negative number th u s is just as imaginary and j u st as real , ,

d epending on the case to which it is applied as the imaginary ,

number x/ — 1 and the only di fference is that we have be come


, ,

familiar with the negative n u mber at an earlier age where we ,

wer e less critical and thus have taken it f or granted become


.
, ,

familiar with it by u se and u s u ally do not reali ze that i t is


,

a mathematical conception and not a physical reality When , .

w e first learned it however it was q uite a step to become


, ,

accustomed to saying 5 — 7 = —2 and not simply —


5 7 is
, , ,

impossible .

Mu lti pli c a ti o n a nd D ivi si o n .

4 .we have a bunch o f 4 horses and another bunch of 4


If ,

horses and still another b u nch o f 4 horses and add together


, ,

the three b u nches of 4 horses each w e get , ,

4 horses + 4 horses 4 horses 12 horses ;


8 E NGIN E E RIN G M AT H E M A T I C S .

Thus it is seen that the reverse O peration of multiplication ,

or division cannot always be carried o ut


,
.

5 If we have 1 0 apples and divide them into 3 we get 3


.
, ,

apples in each group and one apple left over ,


.

10
rema i nder 1,

we may now cut the left over apple into -


3 eq ual parts in ,
w hich
case ,

In the same manner i f we have 1 2 apples we can divide , ,

into 5 by cutting 2 apples each into 5 eq ual pieces and get


, ,

in each of the 5 groups 2 apples and 2 pieces ,


.

1
2

To be able to carry the operation of division through for


all n u merical values makes it necessary to introd u ce a new
,

unit smaller than the original unit and derived as a part o f it


, ,
.

Th u s i f we divide a string o f 1 0 feet length into 3 eq ual


,

parts each part contains 3 feet and 1 foot is le f t ove r O ne


, ,
.

f oot is made up o f 1 2 inches and 1 2 inches divided into 3 gives ,

4 inches ; hence 1 0 feet divided by 3 gives 3 feet 4 inches


, .

D ivision leads u s to a new form of n u mbers : the fraction .

The fraction however is j ust as m u ch a mathematical con


, ,

ce pt io n which sometimes may be applicable and sometimes


, ,

not as the negative n u mbe r I n the above instance of 1 2


,
.

horses divided into 5 gro u ps it is not applicable


, , .

12 horses
2 g horses
,

is impossible ; we cannot have fractions of horses and what ,

we would get in this attempt would be 5 gro u ps each com '

prising 2 horses and some pieces of carcass .

Thus the mathematical conception of the f raction is ap


,

plicable to those physical quantities which can be divided i nto


smaller units but is not applicable to those w hich are i ndi
, ,

visible or i ndi vi du als as we usually call them


, , .
T HE GE N E R AL N UM BE R . 9

I nvo l u ti o n a nd E vo l u ti o n .

6 . If w e have a prod u ct o f several equal factors , as ,

4X4X4 64,
3
it is written as ,
4 64 ;

or in general
, ,

The operation of multiple multiplication of e qual f actors


th u s leads to the ne x t algebraic operation — i nvolu ti on; j u st as
the operation o f m u ltiple addition o f eq u al terms leads to the
operation o f multiplication .

The operation o f invol ution defined as multiple multiplica ,

tion requires the e x ponent b to be an integer n umber ; b is the


,

n u mber of f actors .

3
Thu s 4 has n o immediate meaning ; it would by definition
-

be 4 m u ltiplied — 3 ) times with itsel f .

D ividing contin u o u sly by 4 we get , ,

etc and if this s uccessive divi sion by 4 is carried


.
,

s till fu rther we get the following serie s


,

3
4 —
4X4
4 4
2
4 4X4
4 4

or in general
, ,
10

Thus
E NGIN E E RING IVA TH E M AT] C S

powers with negative e x ponents


" as
.

a
—b
are the
, , ,

1
reciprocals of the same po w ers with positive e x ponents b
°

7 . From the definition o f invol ution then follo w s ,

"
b n b -t- n
a x a z
a ,

beca u se a
b
means the product o f b eq u al f actors a and a
b
prod u ct of n eq u al factors a and a X e th u s is a prod u ct hav ,
" , the
.

ing b + n eq u al f actors a For instance .


,

= = 45
( 4 x 4 x 4)
3 2
4 x4 .

The q u estion now arises w hether by m u ltiple involution ,

we can reach any f u rther mathematical operation For instance .


,

may be written ,

( )
3 2 = 3 3
4 4 X 4

46 J

an d in the same manner ,

b n
(a )
'
b th
that is a po w er a is raised to the n power by m ultiplying
,

"
,

its e x ponent Thus also .


,

W )
t hat is the order of invol ution is immaterial
,
.

Therefore m u ltiple invol ution leads to no further algebraic


,

operations .

3
8 . 4 64

that is the product of 3 e qual f actors 4 gives 64


, , .

I nversely the problem may be to resolve 6 4 into a product


, ,

o f 3 eq u al factors E ach o f the f actors then will be 4 This


. .

reverse operation of invol u tion is ca lled evolution and is written ,

thus ,

x/ B
s
4 = 4;
or more general
, ,
TH E GE N E R AL N UM B E R . 1I
-

x/b th us is defined as that n u mber a, which raised to the po w er


,

b , gives 0 ; or in other
,
w ords ,
b

thu s f ar was defined only f or integer positive


I nvol ution
and negative e x ponents and the q u estion arises w hether po wers
, ,
I n

w ithfractional e x ponents as . ch or 0 75 , have any meaning .

Writing ,

I
it is seen that c? is that n u mber ,
w hich raised to the power b,
1
b
gives c ; that is c? is x/c and the operation o f evol ution th us
,

can be e xpressed as invol ution with f ractional e xponent ,

O bvio u sly then ,

I rrati o na l Nu m b ers .

I nvol u tion
with integer e x ponents as 6 4 can al w ays 43
9 . , ,

be carried o u t I n many cases evol ution can also be carried


.
,

o ut . For instance ,

while in other cases it cannot be ca rrie d o u t For instance


'

,
.
,
,
E NGI N E E R I NG M ATH E M ATI CS .

Attempting to calc u late ( 5 , we get ,

92 :

and find no matter how f ar we carry the calculation we never


, ,

come to an end b u t get an endless decimal fraction ; that is


, ,

no n u mber e x ists in o u r system of numbe rs w hic h can e x press ,

V2 but we can only appro x imate it and carry the approxi ma


, ,

tion to any desired degree ; some s u ch n u mbers as E have be en , ,

calc u lated up to several h u ndred decimals .

S u ch n u mbe rs a s V ’
which cannot be e x pressed in an y
finite form b ut merely appro x imated are called i rrat i ona l
, ,

nu mb e rs The name is j u st as wrong as the name negative


.

number ,
or imaginary n u mber There is nothi ng irrational .

abo u t V2 If we draw a square with 1 foot as side the len gth


.
, ,

o f the dia g onal is feet and the len gth of the diagonal o f
,

a sq u are obvi ously is j ust as rational as the length of the sides .

I rrational n u mbers th us are those real and e x isting num be rs ,

which cann ot be e x pressed by an integer or a f raction or finite ,

decimal fraction but give a n endless decimal fraction whi ch


, ,

d oes not repeat .

E ndless decima l fractions freque ntly are met when e x press


ing common f ractions as decimals These decimal representa .

tions of c ommon fractions ho w ever are pe ri odi c decima ls , , ,

that is the n u merical val ues period ically repeat and in this
, ,

respect are different fr om the i rrational nu m ber a nd can due , ,

to their periodic nat u re be converted into a finite common


,

fraction For instance


.
,

Le t

1 000x = 2 1 3 8 7 3 87 38 7. .

s u btracting ,

999x = 2 1 3 6 6 .

H ence ,
2 1 3 66 1 184 77
2
9 99 9 990
°

5 55 5 55
TH E GE NE RAL N U MBE R . 13

Qu a d ra tu r e Nu m b e r s .

10 . It is

b u t it als o 1 s

— 2)

Therefore V1 4 has t w o val u es ,


and and in ,

evol ution w e th u s first stri ke the interesting f eat u re that one ,

and the same operation with the same numerical val ues gives , ,

several di fferent res u lts .

Since all the positive and negative numbers are u sed up


as the sq u are roots of positive numbers the q uestion arises , ,

W hat is the sq u are root of a negative n u mber ? For instance ,

V— 4 cannot be —2 as — 2 sq u ared gives + 4 nor can it be + 2


, ,
.


solves itsel f into : Wha t is I? i
x i
We have derived the absol u te n u mbers f rom e xperience ,

f or instance by measuring distances on a line Fig 4 f rom a


,
.
,

st a rting point A .

0
—r
-5
i
-
-r
4 -3 -2 -
1 +1 +2 +3

+4 + 5
1
C A 8

FI G 4 . . Ne gative a nd P o itive
s Num be rs .

Then we have seen that we get the same distance from A ,

twice once toward the right once toward the left and this
, , ,

h as led to the subdivision of the numbers into positive and


negative numbers Ch oosing the positive to w ard the right
.
,

in Fig 4 the negative n u mber wo u ld be toward the le f t ( or


.
,

inversely choosing the positive to w ard the le f t wo u ld give


, ,

the negative toward the right ) .

I f then we take a n u mbe r as + 2 which represents a d is , ,

tance AB and m u ltiply by


,
A
we get the di s tance C = —
2
14 E N GIN E E R I NG M A TH E M A T I C S .

in opposite direction from A . I nversely , if we take AC ,


—2

and multiply by we get AB = + 2 ; that is , m u ltiplica


tion by — 1 ) reverses the direction t u rns it through 1 80 deg ,
.

I f we multiply + 2 by V: we get + 2 V — 1
a quantity , ,

o f w hich we do not yet know the meaning Mu ltiplying once .

more by V T I we get ,
V It = — 2 ; that is ,

multiplying a number + 2 t w ice by V — 1


gives a rotation o f
, ,

1 80 d e g and m u ltipli cation by


,
th u s means rotation by
half o f 1 80 d e g ; or by 9 0 d e g and + 2 V — 1 thus is the dis
, ,

FI C . 5 .

tance in the direction rotated 90 deg f rom + 2 or in q uadrat u re .


,


direction AD in Fig 5 and s u ch n u mbers as ,
1 th u s
are quad ra tu re nu mb e rs that is represent direction not to w ard
, ,

the right as the positive nor to w ard the le f t as the negative


, , ,

n u mbers b u t u p w ard or dow n w ard


,
.

— 1
For convenience of writing is us u ally denoted by
,

the letter j .

I I . J u st as the operati on of s u btraction introd u ced in the


negative n u mbers a ne w kind o f numbers having a direction ,

180 deg di ff erent that is in opposition to the positive nu m


.
, ,

be rs,
so the operation of evol u tion introd u ces in the q u adrat u re
n u mber as ,
a ne w kind of n u mber having a direction 9 0 deg ,
.
16 E NGINE E RING M AT H E M A T I C S .

n u mbers I n still other problems as when dealing w ith illu m i


.
,

nation o r with individ u als the negative numbers are not


, ,

applicable b u t only the absol u te or positive nu mbers


, .

J u st as m u ltiplication by the negative unit — 1 ) means


rotation by 1 80 d e g or reverse of direct ion so multiplication
, ,

by the quadrature unit j means rotation by 9 0 d og or change


, , ,

f rom the hori zontal to the ve rtical direction and inve rsely ,
.

G e n e r a l Nu m b e r s .

1 2 . By the
positive and negative numbers all the points of ,

a line co u ld be represented n u merically distances f rom a


chosen point A .

FI G 7
. . i
S mp le Ve ctor D iagram .

By the addition O f the q u adrature n u mbers all points of ,

the entire plane can now be represented as distances f rom


chosen coordinate a x es a: and that is any point P o f the ,

plane Fig 7 has a hori zontal distance OB = + 3 and a


.
.
, , ,

vertical distanc e BP = + 2j and ther efore is given by a


, ,
.

combination of the distances OB = + 3 and B P = + 2 j For ,


.

convenience t h e act o f combining t w o s u ch distanc es in q u ad


,

rat u re with each other can be e x pressed by the pl u s si gn ,

and the result of combination thereby e x pressed by OB +B P


3 + 2j .
TH E G E N E R AL N UM B E R . 17

Such a co mbination of an ordinary nu mber and a qu adra -

t u re number is called a gene ral nu m b e r or a com plex qu anti ty .

The q u adrat u re n u mber jb th u s enormo usly e x ten ds the


field o f u se fulness o f algebra by a ff ording a n u merical re pre
,

se nt at io n of two dimensional systems


-
as the plane by the , ,

general n u mber a + 3b They are especially u se f u l and impor


.

tant i n electrical engineering as most problems o f alternating


,

c u rrents lead to vector representations in the plane and there ,

fore can be represented by the general number a + 3b ; that is ,

the c o mbinati o n o f the ordinary n u mber or hori zontal distance


a and th e q u adrat u re n u mber or ve rt ical distance jb
,
.

FI G 8 . . Ve ctor D iagram .

Analytically , points in the plane are represented by their


t wo coordinates : the hori z ontal coordinate or abscissa x and , ,

t he vert ical coordinate or ordinate y Algebraically in the


,
.
,

ge neral nu mber a + jb both coordinates are combined (1 being ,

the a: coordinate jb the y coordinate


,
.

Th u s in Fi g 8 coordinates of the points are


.
, ,

P1 : x = + 3, =
y +2 P2 2 =
y
-
2,

P3 = —3 = +2 P4 : = —
3 = 2,
y y
-
: x ,
x

and t he points a re located in the plane by the n u mbers


18 E NG I N E E RING M ATH E M ATI C S .

1 3 . Sincethe sq uare root o f negative n u mbers has


a lre a d v

e x tended the system O f n u mbe rs by giving the quadrat u re


n u mber the q u estion arises whether still f u rther e x tensions
,

o f the system o f n u mbers wo u ld res u lt f rom higher roots of


negative q uantities .

For insta nce ,

The meaning o f V 1 we
t fin d in the same manner as that

A positive n u mber a may be represented on the hori z ontal


a x is as P
Multiplying a b y {l 1 gives a i l 1 whose meaning we do
— —
,

not yet kno w M u ltiplying again and again by i


. f— 1 we get afte r , ,

f o u r m u ltiplications a H ,

a ; that is in four steps we
-
,

have be en carried from a to a a rotation o f 1 80 d e g and -


, ,

1 80
V —1 thus means a rotat i on of — V
4

de g ,
therefore ,
C

1
4
is the point P I in Fig 9 at distance a from the coordinate.
,

center and u nde r angle 45 d e g which has the coordinates


, , ,

5 2
and y
=
i or is represented by the general number
, , ,

P1 = a

V —1
however m ay also mean a rotation by 1 35 deg to P 2

, .
, ,

since this repeated f o u r times gives


,
de g , ,

or the same as 1 80 d e g or it may mean a rotation by 22 5 deg


, .

or by 3 1 5 deg Thus f o u r points e x ist w hich represent a {I —1 ;


.
,
.

the points

a, P4

There f ore {l l is still a general n u mber cons ist ing o f an


,

,

ordinary and a q u adrat u re n u mber and th u s does not e x tend ,

o u r system of n u mbers any f u rther .


TH E G E N E R AL N UM B E R .
9

Inthe same manner 1 can be f o u nd ; it is that n u mber
,
,

which multiplied n times with itsel f , gives + 1 Th us it re pre


, .

360
sents a rotat i on by de g , or any m u lt i ple thereo f ; that is
n ,

the x coordinate is cos q>< the y coordinate sin q><

n 3 60
V +1 = cos — —
qX +j s 1n q><
n

where q is any integer number .

Ve ctor D ia gram a d ‘
FI G 9. . 1 .

There are therefore n di fferent values of a f/ + 1 w hich lie ,

equidistant on a circle with radi u s 1 as sho w n f or n = 9 in ,

Fig 1 0
. .

I4. I n the operation of addition a + b = c the problem is , , ,

a and b being given ; to find c .

The terms of addition a and b are interchangeable or


, , ,

eq u ivalent thus : ,
and addition there f ore has o nlv
one reverse operation subtraction ; c and b being given a is
, ,

fo u nd th u s : a c b and c and a being given b is f o u nd th u s :
=
, , , ,

b= c — a —
E ither leads to the same operation s u btraction
. .

The same is the case in m u ltiplication ; The


90 E NG I N E E R I NG . MAT HE MAT I CS .

factors a and b are i nterchangeable or e q u ivalent ; a >< b = b >< a


C C
and t he reverse operation division a = is the same as b=
d , ,
b

b i
I n invol u tion however o c t h e two numbers a and b
, ,
-
,

are no t i nterch angeable and a is not eq u al to b For instanc e ,
. .

and
Therefore invol ution has t w o reverse operatio ns
,

( a) c and b given a to be fo u nd , ,

or evolution ,

FI G 1 0 . . P oi nt s D te rm i ne d
e by

(b ) c and a given ,
b to be f o u nd
b = lo ga c;

or ,
lo ga rit h m at io n .

L o ga ri th m a t i o n .

I sLo garit hm at io n thus is one o f


. the reverse operations
of invol u tion , and the logarithm is the
e x ponent of invol ution .

Thus a logari th mic e x pression may be changed to an e x


n nt ial and inversely and the laws o f lo garit h m at i on are


p o e , ,

the laws which the e x ponents obey in involution


,
.

1 Powers of eq u al base are multiplied by adding the


.

e x ponents : Therefore the logarithm of a ,


TH E GE N E R AL N UM BE R . 21

product is the sum o f the logarithms o f the f actors thus lo g 0 >< d , o

— log c + log d
a a .

2 A po w er is raised to a power by m u ltiplying the e x ponents :


.

There f ore the logarithm of a power is the e x ponent times


the logarithm o f the base or the n u mbe r under the logarithm , ,

is raised to the power n by m ultiplying the logarithm by n : ,

log a c
n =n log a 0,

log , 1 = 0,
because d 1 I f the base a 1 log} c is positive o
.
, , ,

if c > 1 and is negative i f c < 1 but > 0 The reverse is the


, , ,
.

case if a < 1 Th us the logarithm traverses all positive and


,
.
,

negative val ues for the positive values o f c and the logarithm ,

o f a negative n u mber thus can be neither positive nor negative .

log — e ) = log c + lo g and the question o f finding


a a a

the logarithms of negative n u mbers thus resolves itself into


fi nding the val u e of log a

There are two standard systems of logarithms one with


the base e = 2 7 1 828 and the other with the base 10 is
. .

use d t he f ormer in algebraic the latter in numerical calcula


, ,

tions L ogarithms of any base a can easily be reduced to any


.

other base .

For instance to red u ce b = lo g c to t he base 1 0 : b = log c


, a a

means in the f orm o f involution : a c Taking the logarithm


,
b =
.

hereof gives b lo gl o a = lo g1 0 c hence


, , ,

lo g o 0 lo g l o 0
or lo go, c
lo gl o a

lo g e a

Thus regarding th e logarithms o f negative n u mbers


, ,
we
l
to consider only o g l o — 1 ) or log ,

x = lo then n
If j g, a

and since as will be seen in


,
C hapter II ,

cos x +j sin 23,

it follows that
cos x +j sin 2:

Re g ardi ng 6 , see Chapte r II, p 71 . .


22 E NGIN E E RING AI A TH E JII A TI C S .

x = 7z, or an odd mu ltiple thereo f and ,

10 ge ( I ),

where n is any integer n u mber .

Th u s lo garith m at io n also leads to t h e q u adrature n u mber


,
3 b u t to no further e x tension of the system of n u mbers .

Qu a t e m i o ns .

1 6 . Addition and subtraction , multiplication and division ,


involution and evol ution and lo gari t hm at io n thus represent all
the algebraic operations and the system of n u mbers in which
,

all these operations can be carried out u nder all conditions

is that of the general n u mber a + jb comprising the ordinary , ,

number a and the q u adrat u re number jb The n u mber a as .

well as b ma y be positive or negative may be integer f raction , ,

or irrational .

Since by the introdu ction o f the quadrature number jb ,

the application of t he system of n u mbers was e x tended from the


line or more general one dimensional q u antity to the plane
, ,
-
, ,

or the two dimensional q u antity the q u estion arises whether


-
, ,

the system o f n u mbers co u ld be still f u rther e x tended into ,

three dimensions so as to represent space geometry While


,
.

in electrical engineering most problems lead only to plai n


figures vector diagrams in the plane occasionally space figu res
, .
,

would be advantageo u s i f they co u ld be e x pressed algebra


ically . E specially in mechanics this wo u ld be of importance
when dealing with forces as Vectors in space .

I n the q u aternion calcul u s methods have been devised to


deal with space problems The q u aternion calc u l u s however
.
, ,

has not yet found an engineering application comparable w ith


that of the general n u mber or as it is f req uently called the , , ,

com ple x qu a nti ty The reason is that the q u aternion is not


.

an algebraic q u antity arid the la w s of al gebra do not u ni fo rm lv


,

appl y to it .

1 7. With the rectangu lar coordinate system in the plane ,

Fig 1 1 the x a x is may represent the ordinary nu mbers the y


.
, ,

a x is the q u adrature numbers and multiplication by ,

represents rotation by 90 deg For ins tance i f P I is a point .


,
24 E NGIN E E RING M ATH E M ATI CS .

Algebrai cally this is not possible , since each o f the three quan
tities is -
1, and —1 — 1 and not
,

FI G 1 2
. . Ve ctor s i n S pac e , ihk = 1 .

Ifwe now proceed again from x in positive rotation but


, ,

first t u rn in the r e plane w e reach by m u ltiplication with k


,

z as seen in Fig 1 3

the negative 2 a is
x , ,
Fu rther multiplica
. .

Ve ctors in Sp a c e khf
,
:

tion by It brings us to + y and m u ltiplication by j to — x and


, ,

in this case the resu lt o f the three s u ccessive rotati ons by


TH E GE N E R AL N UM BE R . 25

90 de g , in
the same direction as in Fig 1 2 b ut in a di ff erent .
,

order is a reverse ; that is represents


, There f ore , ,

k j
h = — 1,
and hence ,

= —
jh k l
c hj .

Thus in vector analysis o f spa ce


, ,
we see that the fundamenta l
law o f algebra ,

a Xb b X a,

does not apply and the order of the factors of a produ ct is


,

not immaterial but b v changing the order o f the factors o f the


,

product jhk its S ign was reversed


,
Thus common f actors can .

not be canceled as in algebra ; for instance i f in the correct e x ,

pressio i
n h k ,
—khj we should can cel by i h and k as could be
, , ,

done in alge b ra we wou ld get 1 — 1 which is O bvio usly wrong


, , .

For this reason all the mechanisms devised for vector analysis
in space have proven more di ffic u lt in thei r appli cation and ,

have not yet been used to any great e x tent in engineering


practice .

B . AL GE B RA OF THE GE NE RAL NU MBE R , OR C OM PL E X


QUANTITY .

R e ct ang u la r a nd Po la r C o o r d i nat e s .

The general n u mber or comple x q uantity a + jb is


1 8. , , ,

the most general e x pression to which the law s of algebra apply .

I t therefore can be handled in the same manner and under


th e same r u les as the ordinary n u mber o f elementary arit hm eti c .

The only feature which m u st be kept in mind is that j 2


1 and -
,

where in multiplication or other operations occu rs , it is re


placed by its val ue — 1 Thus for instance ,
.
, ,
'

( a jb ) (c id) ac jad jb c b
3 d
2

— b
ac j d
a jb c d
( ac b d) j ( a d b e) .

Here f rom it follows that all the higher powers of i can b e


eliminated thu s : ,

2= —1
.

I I
0

I
l ) 7

s= 6 = —1 7
i + 33 i . i
9 =
i +i .
e tc
26 E NGI N E E R I NG Iii ATH E M A TI C S

I n d istinction f rom the general nu mber or comple x q u antity ,

the ordinary n u mbers + 61 and are occasio nally called—a


, ,

sca la rs or real n u mbers The general n u m ber thus consists


,
.

o f the combination o f a scalar or real n u mber and a q u adrat u re


n u mbe r or imaginary n u mber
,
.

Since a q u adrature number cannot be eq u al to an ordinary


n u mber it follo w s that i f t w o general n u mbers are equal
, ,

thei r real components or ordinary n u mbers as well as their ,

q u adrature n u mbers or imaginary components must be e qual ,

thu s if ,

a =o and b=d .

E very e quation w ith general n u mbers th u s can be resolved


into t w o e q u ations one containing o nly the ordi nary num be rs
, ,

the other only the q u adrat u re num bers For i nstance i f .


,

33
_
5 3l)
.

x=5 and = —3
y .

1 The best way of getting a conception o f the general


9 .

number and the algebraic operations with it is to consider


, ,

the general n u mbe r as representing a point in the plane Thus .

the general n u mber may be considered as


representing a point P in Fig 1 4 w hich has the hori z ontal
,
.
,

distance from the y a x is and the vertical


,

distance f rom the x a x is ,

The total distanc e of the point P from the c oordinate center


0 then is

and the angle which thi s distanc e OP makes with the


,
2: a xi s ,

is given by
TH E GE N E R AL N UM B E R . 27

I ns tead of repre senting the general nu mber by the tw o


components a and b in the f orm a + jb it can also be re pre
, , ,

sented by the two qu antities : the distanc e o f the point P from


the c enter 0 ,

and the angle between this distance and the 2: a x is ,

FI G 1 4
. . Re ctang lar and Polar Coordi nates
u .

Then re f erring to Fig . 1 4,

a = o cos d and b=c si n 0 ,

and the general nu mber a + jb th u s can also be w ritten in the


f orm,

(
c c os d +j si n

The form a + jb e x presses the general n u mber by its


rectangu lar components a and b and corresponds to the rect ,

angular coordi nates o f analytic geometry ; a is the x coordinate ,

b the y coordinate .

The fo rm c (co s 0 + j sin 0 ) e x presses the general number by


what may be called its polar components the radius 0 and the ,
28 E NGIN E E RI NG M AT H E M AT I C S .

angle (9 and corresponds to the polar coordinates of anal yti c


,

geometry 0 is freq u ently called the radi u s vector or scalar


.
,

d the phase angle o f the general number .

While usually the rectangu lar form a + jb is more con


ve nie nt sometimes the polar form c ( co s d + j sin 0 ) is pre f erable
, ,

and transformation from one form to the other therefore fre


q u ently applied .

Ad diti o n and Su btr a c ti o n .

20 . If j is represented by the point P I ;

this point is reached by going the hori z ontal distance d l = 6


and the vertical distance b 1 = 2 5 I f a 2 + jb 2 = 3 + 4j is re pre
. .

sented by the point P 2 this point is reached by going the


,

hori z ontal distance a 2 = 3 and the vertical distance b 2 = 4 .

The sum of the two general n u mbers + jb 2 )


( 6 + 2 5 7) + 4j) then is given by point P 0 which is reached
, ,

by going a hori z ontal distance equal to the sum of the h e r


i zont al distances of P I and P 2 : and a
vertical distance equal to the s um of the vertical distances o f
P I and P 2 : hence is given b V the ,

general nu mber
00 + 150 =
( a 1 + 0 2 ) + f(b 1 + 52 )
= 9 + 65
i

G eometrically point P 0 is derived from points P I and P 2


,
,

by the diagonal C FO o f the parallelogram 0 P 1 P 0 P 2 constructed ,

with OP I and C P 2 as sides as seen in Fig 1 5 ,


. .

Herefrom it follo w s that addition of general nu mbers


represents geometrical combination by the parallelogra m law .

I nversely i f P 0 represents t h e nu mber


,

( Lo + ] Oo “"
9 l

and PI represents the nu mber

al + jb 1 = 6 + 2 5 3,

the di fference of these n u mbers will be represented by a point


P2 which is reached by going t he di fference o f the hori z ontal
,
TH E GE NE R AL N UM B E R . 29

distances and o f the vertical distances o f the points P0 and


PI . P 2 thu s is represented by

a2 = a — a = 9
o 1

bz bo —b =4
1 .

Therefore the di ff erence o f the t w o general n u m b ers ( a o + ib o)


,

and (d l + jb 1 ) is given by the general n u mber :

as seen in

Addition a nd Su btract i on of V ctorse .

This di fference a 2 + jb 2 is represented by one side OP 2 o f


the parallelogram OP I P OP Q which has C P 1 as the other side
, ,

and OP O as the diagonal .

Su btraction of general n u mbers thus geometrically represents


the resol ution o f a vector OP O into two components OP I and
OP 2 by the parallelogram law
,
.

Herei n lies the main advantage o f the u se of the general


nu mber in engineering calculation : I f the vectors are represented
by general n u mbers ( comple x q u antities ) combination and ,

resolution of vectors by the p arallelogram law is carried O iI t by


30 E NGIN E E RING M A T H E M A TI C S .

simple addition or s u btract i on of their general numerical valu es ,

that is by the simplest O peration of algebra


,
.

2 1 G eneral numbe rs are us u ally de noted by capitals , and


.

their rectangular components the ordinary n u mbe r and the ,

q uadrat u re num b e r by small letters th u s :


, ,

A= a1 + 711 2 ,

the distance of the point w hich represents the general numbe r A


f rom the coordinate center is called the a b solu te valu e radius ,

or scalar o f the general n u mber or comple x q u antity I t is .

the vector 0 in the polar representation o f the ge neral numbe r :


.

A = a (c os d + j si n

and is given by a = V (1 1 + a 2
2 2
.

The absol ute val u e or scalar of the general numbe r is usuallV


, ,

also denoted by small letters but sometimes by capitals and


. ,

in the latter case it is distingu i shed f rom the general n u mber b v


u sing a di ff erent t vpe for the latter or u nderlining or dotting ,

it thus
,

A = a 1 + ja 2 ; or + 311 2

or A= a 1 + ja 2 ; or A = a 1 + ja 2

or
al + ja 2 = a (cos d + i si m

al + ja 2 = A(cos 0 + j sin

22 The absol ut e val u e or scalar o f a general numbe r is


.
, ,

al ways an absolu te nu mber and p ositive that is the S ign of the


, , ,

rectangular component is represented in the angle 0 Th u s .

referring to Fig 1 6 .

A = a i + ja 2 = 4 + 3j ;

tan (9
d = 37 d e g ;

A = 5 ( c os 3 7 deg . +j si n 37 deg) .
E NGINE E R I NG M ATH EM ATI C S .

The e x pression

d = — 37 de g . or = 1 80 de g ;

b utsince a1 =a cos d is negati ve 4 co s 0 must be negative —


, , ,

hence d ,
1 43 deg is the correct val u e and
.
,

A = 5 (cos 1 43 deg sin 1 43 de g )


. +j .

— 5( deg + j si n 37 deg )
. .

The e x pression
A= a 1 + ja 2 = —4 — 3
]

de g ; or de g ;

b utsince al = a cos d is negative 4 cos 0 m u st be negative ,


-
, ,

hence deg is the correct value and


.
, ,

A = 5 ( cos 2 1 7 deg + j si n 2 1 7 deg ) . .

— 5( cos 37 deg .

j si n 37 deg ) .

The fo u r general n u mbers + 4 + 4 —


3j — 4 + 3 j and
, , ,
— 4
3j have the same absol u te value 5 and in their re pre
-
, , ,

se nt at io ns as points i n a plane have symmetrical locations in

the four quadrants as s hown in Fig 1 6 ,


. .

As the g eneral number A = a 1 + ja 2 finds its main use in


representing vectors in the plane it very f requently is called
'
,

a vector q u antity and the algebra of the general numb e rvi e fl


,
-

spoken o f as vector analysi s .

Since the general numbers A = can be made to


represent the points of a plane they also may be called plane ,

nu m b e rs while the positive and negative n u mbers


,
and —a ,

,
TH E GE NE R AL N UM B E R . 33

may be called the li nea r nu m b e rs , as they represent the pOint s


of a line .

E x a m pl e : S t e a m Pa t h i n a Tu r bi ne .

2 3 . As an e x ample o f a simple operation with general nu m


bers one may calc u late the steam p ath in a t w o - w heel stage
of an imp u lse steam turbine .

FI G 1 7. . Pa th of Ste am i n a Two whe e l Sta ge of a n I m pu lse


- Tu rbi ne .

Le t Fig 1 7 represent diagrammatically a tan gential section


.

through the b u c ket rings of the t u rbine w heels W1 and W2 .

are the two revolving wheels moving in the direction indicated


,

by the arrows with the velocity s = 400 feet per se e I are


,
.

the stationary intermediate b u ckets which turn the e x haust ,



steam f rom the first b u cket wheel W 1 bac k into the direction ,

req uired to impinge on the second b u cket wheel W2 The .

steam jet iss u es f rom the e x pansion no zz le at the speed s o = 2 2 00


34 E NG I N E E R I NG Il/I A T H E M A TI C S .

f eet per sec and under the angle


.
,
de g ,
against the first

b u cket wheel W 1 .

The e x hau st angles of the three s u ccessive ro w s o f buckets ,

W 1 I and W2 are respectively 2 4 d e g 30 deg and 45 deg


, , , ,
. .

These angles are calc u lated from the section of the b u cket
e x it required to pass the steam at its momenta ry velocity ,

and f rom the height of the passage requi red t o gi ve no steam


eddies in a manner whi ch is o f no interest here
,
.

As friction coe fficient in the b u c ket passages may be ass u med


that is the e x it velocity is l — kf = 0 88 of the entrance
,
.

velocity of the steam i n the buc kets .

FI G 1 8 . . Vec tor D i a gram of Ve loc iti e s of S te am i n Tu rbine .

Choosing the n as x a x is the direction of the tangential


-

velocit y o f the turbi ne wheels as y—a x is the a x ial direction , ,

the velocity of the s team supply f rom the e x pansion no zz le is


represented in Fig 1 8 by a vector OS O of length s o = 22 00 feet
.

per sec making an angle d o = 2 0 deg with the x—a x is ; hence


.
,
.
,

can b e e x pressed by the general numbe r or vector q uantity


S o = So (C OS 00 + j sin 00 )
— 22 00 cos 2 0 deg + sin
( j . 20 deg ) .

2 070 + 7 5 0j f t per sec . .

The velocity o f the turbine wheel W 1 is s 400 feet per second ,

and represented in Fig 1 8 by the vector OS in hori z ontal


.
,

dire ction .
TH E GE NE R AL N UM B E R . 35

The relative velocity with whi ch the steam enters the b ucket
passage of the first t u rbine w h eel W 1 th u s is :
SI $0 “ 8
(2 07 0 7 5 03) — 400
1 6 7 0 + 7 50 j_ft . per se e .

This vector is shown as 0 S 1 in Fig 1 8 . .

The angle 0 1 u nder which the steam enters the bu cket


,

passage th u s is given by
7 50
ta na ,
1 67 0
as deg .

This angle thus has to be given to the front edge o f t he


b u ckets of t he turbine w heel W 1 .

"
The absol u te val u e of the relati ve velocity of steam j et
and turbine wheel W 1 at the entrance into the buc ket passage
, ,

18

31 7 50 2
1 830 f t per
. se e .

In traversing the bucket passages the steam velocity de


creases by friction etc f rom the entrance value 3 1 to the
.
,

e x it Val u e
X0 . 88 = 1 61 0 ft per sec
.
:
.
,

and since the e x it angle of the b u c ket passage has been chosen
as de g the relative velocity with which the steam
,

leaves the first bu c ket wheel W 1 is represented by a vector


0 S 2 in Fig 1 8 of len gth
.
,
under angle 2 4 deg The .

steam leaves the first wheel in bac kward direction as seen in ,

Fig 1 7 and 2 4 deg thus is the angle bet w een the steam j et
.
,
.

and the negative x a x is ; hence -


deg is the
, .

vector angle The relative steam velocity at the e x it from


.

wheel W 1 can th u s be represented by the vector q uantity


S z = s z (c o s 0 2 + j sin 0 2 )
— 1 6 1 0 ( cos 1 5 6 deg + j sin . 1 56 deg ) .

— 1 47 0 + 655
j .

the velocity o f the t u rbine w heel W 1 is s = 400 t he


S inc e ,

ve locity o f the steam in space after leaving the fi rst turbi ne


,
36 E NGI N E E R I NG M ATH E MA TI CS .

wheel that is the velocity


, ,
w ith which the steam enters
intermediate I is ,

53
1 :
S2 + 8
A

1 47 0 + 6 5 53) + 400
— 1 07 0

and is represented by vector OS ; inFig 1 8 . .

The direction of this steam j et is given by

tan 03 :

de g ; or ,
1 80 deg .

The latter value is correct as cos 0 3 i s negative and sin 03 is , ,

positive .

The steam j et thus enters the intermediate under the angle


of d e g ; that is the angle 1 80 ,
deg in opposite .

direction The b u ckets O f the intermediate I th u s must be


.

curved in reverse direction to those of the wheel W 1 and must ,

be give n the angle deg at their front edge . .

The absolute value of the entrance velocity int o the inter


mediate I is
655 2
1 2 55 f t per sec
. .

In passing thro u gh the bucket passages this velocity de ,

creases by friction to the val ue : ,

s4 = 3 3 (1

k, ) 1 2 5 5 X 0 88 . 1 1 05 ft per sec
. .
,

an since the e x it edge o f the intermediate is given the angle


d
de g the e x it velocity o f the steam from the intermediate
,

is represented by the Vector OS ; in Fig 1 8 of len gth .


,

and angle 04 = 30 d e g hence , ,

S deg + j sin
( cos 30 . 30 deg ) .

= 955 + 55 0 ft per s e e
j . .

This is t he velocity with w hich the steam j et impinges


on the second t u rbine wheel W 2 and as this wheel revolves ,
TH E G E NE R AL N UM B E R . 37

with ve locity s = 400 the relative velocity that is the vel ocity
, , ,

with which the steam enters the bucket passages of wheel W2 is , ,

S5 —
S4 8
— 40 0
( 955 55 03)
555 + 5 5 0j ft per sec
. .
;

and is represented by vector OS 5 in Fig 1 8


~
. .

The direction of this steam j et is given by


55 0
tan 05 5 as d5 = 44 8 deg .

555
.

Therefore the entrance edge of the b u c kets o f the second


,

wheel W 2 m u st be shaped under angle 05 deg .

The absol ute value of the entrance velocity is

555 + 55 0 = 7 80
2 2
ft per s e c
. .

In
traversing the b u cket passages the velocity drops from ,

the entrance valu e S 5 to the e x it valu e ,


,

86 = s5 ( 1 — k) 6 90 ft per sec
, . .

e x it angles o f the buckets of wheel W2 has bee n


Since t h e
chosen as 45 de g and the e x it is in backward direction 06
, ,

1 80 d e g the steam jet velocity at the e x it o f t he


,

bucket passages o f the last wheel is given by the general number


S 6 = 3 6 ( cos 06 + j
sin 06 )
j
= 69 0 cos 1 3 5 de g + sin
( . 1 35 deg ).

— 487 + 487 f t per sec


j . .
,

and represented by vector 0 S 6 in Fig 1 8 . .

Since s = 4o0 is the w heel veloc ity the Velocity o f the ,

stea m a f ter leaving the last wheel W2 that is the lost , ,


or rej ected velocity is ,

$7 = ‘S6 + 8
— 487 + 487
)
3 + 400
487 i ft per sec
. .
,

and is represented by vector 08-


1 in Fig . 18 .
38 E NG I N E E R I NG M A T H E MA T I C S .

The direction of the e xha u st steam is given by ,

" e r
1 :
as deg ,
97

and the absolute velocity is ,

Mu lti pli c at i o n o f G e ne r a l Nu m b e rs .

2 4. I f A = a 1 + ia 2
and B = b 1 + jb 2 are tw o general O r , ,

plane n u mbe rs their product is given by m u lt iplication th u s :


, ,

AB ( C 1 + ja 2 ) ( b 1 + jb 2 )

a l b l + j0 4b 2 f jCI l

/

a nd since j 2 = —1
,

AB = (a 1 b 1 —
0 , 2 O2 ) + j (a 1 b 2 a zb l ) ,

and the product can also be represented in the plane by a point , ,

C = c 1 + jc2 ,

CI = ai b 1
“ 0 52 ,
2

0 2 = 0 1 O2 + G Q O1 .

For instance A = 2 + j m u l tiplied by


,
B=1 j gives
Cl = 2 >< 1

as shown in Fi g 1 9 . .

25 The geometrical relation bet w een the factors A and B


.

and the prod u ct C is better shown by u sing the polar e x p ression ;


hen ce s u bstituting
, ,

-
a1 =a cos a

a g = a S in a

which gives

tan a
40 E N GINE E RIN G M ATH E MA TI CS .

The multiplier B is occasionally called an o pe rator as it ,

carries out the operation o f rotating the direction and changin g


the length of the m u ltiplicand .

26 I n m u ltiplication division and other algebraic opera


.
,

tions with the representat ions o f physical q u antities ( as alter


nating c u rrents voltages impedances etc ) by mathematical
, , ,
.

symbols whether ordinary n u mbers or general n u mbers it


, ,

is necessary to consider whether the result of the algebraic


operation f or instance the prod u ct of two f actors has a
, , ,

physical meaning and i f it has a physical meaning whether


, ,

this meaning is s u ch that the prod u ct can be represented in


the same diagram as the f actors .

For instance 3 X 4 = 1 2 ; but 3 horses X 4 horses does not


,
'

give 2 horses nor 1 2 ho rse s b ut is physically meaningless


1 z
, ,
.

H owever 3 f t X 4 ft = 1 2 s q ft
, . Thus i f the n u mbers represent
. . .
,

FI G 20 . .

horses m ultiplic ation has no physical meaning I f they re pre


,
.

sent feet the prod u ct o f multiplication has a physical meaning


, ,

but a meaning which di ffers f rom that o f the f actors Thus .


,

if on the line in Fig 2 0 0 A = 3 feet 0 B = 4 fe et the product


.
, , , ,

1 2 square f eet while it has a physical meaning cannot be


, ,

represented any mo re by a point on the same line ; it is not


the point OC = 1 2 because i f we e x pressed the distances 0 A
, ,

and 0 B in inches 36 and 48 inches respectively the prod u ct


, ,

wo u ld be 36 X 48 = 1 7 2 8 s q in while the distance C C would be . .


,

1 44 inches .

27 I n all mathematical operations with physical quantities


.

it t herefore is necessary to consider at every step of the mathe


m at ical operation whether it still has a physical meaning
, ,

and if graphical representation is resorted to whether the


, ,

nature of the physical meaning is such as to allo w graphical


representation in the same dia gram or not ,
.

An instance of this general limitation of t he application of


mathematics to physica l q u antities occ u rs in the representation
of alternating c u rrent phenomena by general numbers or ,

comple x quantities .
T HE GE N E R AL N UM B E R . 41

An alte rnating c u rrent can be a vector 0 1 re pre s e nt e d b y


in a polar di agram Fi g 2 1 in which one complete revol ution
,
.
,

or 3 60 deg represents the time of one complete period o f the


.

alternating c u rrent This vector 0 1 can be represented by a


.

general n u mber ,

= 3
I l:1 + 352 , ’

where fl is the hori z ontal ,


73
2 the vertical component of the
current vector

FI G 2 1
. . Cu rr nt
e ,
E MF. . . a nd I m pe dance Ve ctor D ia gra m .

In the same manner an alternating E M F of the same fre . . .

que nc y can be represented by a vector OE in the same Fig : 2 1 ,

and denoted by a general number ,

An impedance can be represented by a general number ,


Z = r + jx ,

where ris th e resistance and x the reactance .

I f now we have two impedances 0 Z 1 and 0 Z 2 Z 1 = r 1 , ,

and Z 2 = r2 + s their prod u ct Z 1 Z 2 can be f ormed m at h em at


,

ically but it has no physical meaning


,
.
42 E NGIN E E R I NG ItI A T H E M A TI C S .

I f we have a c u rrent and a voltage I 151 ji g and E 6 1 ,

the prod u ct o f curre nt and voltage is the power P o f the alter


nating circ u it .

The prod u ct of the two general n u mbers I and E can be


formed mathematically I E and wo u ld represent a point C , ,

in the vector plane Fig 2 1 This point C how ever and the . .
, ,

mathematical e x pression I E w hich represents it does not give , ,

the po w e r P o f the alternating circuit since the power P is not ,

o f the same f req u ency as I and E and there f ore ca nnot be ,

represented in the same pola r diagram Fi g 2 1 which represents .


,

I and E .

I f we have a c u rrent I and an impedance Z in Fi g 2 1 ,


.

I = i 1 + 3732 and Z = r + 31: their prod u ct i s a voltage and as the ,

voltage 1 8 of the same f req u ency as the c u rrent it can be repro ,

sented I n the same polar dia gram Fi g 2 1 and thus i s given by ,


.
,

the mathematical product of I and Z ,

E 1 Z = (i 1 + 1752 ) (r
—i2 x
t
( l r j (i 2 r + i 1 x ) .

2 8 . Commo nly ,
in the denotation of graphi cal diagrams by
general numbers as the polar diagram o f alternating currents
, ,

those quantities which are ve ctors in the polar diagram as the


, ,

current voltage etc are represented by dotted capitals : E I


, ,
.
, , ,

w hile those general n u mbers as the impedance admittance etc , , ,


. .

whi ch appear as operato rs that is as m u ltipliers of one vector , , ,

for i nstance the c u rrent to get an other vector the voltage are
, , ,

represented alge b raicallv by capitals without dot : Z = r + jx


impedance etc ,
.

T h is limitation of cal culation w ith the mathematical re pre


s e nt at io n o f physical q u antities m ust c onstantly be kept in
mind in all theoretical investigations .

D i vis i o n of G e ne r a l Nu m b e rs .

2 9.The division of tw o general n u mbe rs A = a 1 + ja2 and ,

B = b 1 + jb 2 gives , ,

A d l + ja 2
B b l + 21)

This f raction contains the q uadrat u re n u mber in the numer ~

a tor as well as in the denominator T he q u adrat u re n u mbe r .


TH E G E N E R AL N UM B E R
.
" 43

can be eliminated from the denominator by m u ltiplying n u mer


ator and denominator by the conj u gate q uantity o f the denom
i at o b — b which gives
n r , l j z ,
:

—a b
(0 1 + 721 2 ) ( b l ‘

Ib zl ( albl + 0 2 52 ) + 10 s 1 i zl

I52 )
z
( b l + Ib 2 ) (b 1
— b l + 132 2

a lb l + a2b2 0 2 5 1 — 0 1 52
+2
51 + 52
2 2
(7 1 2
5
+ 2 2

f or instance ,

A 6 + 2 5j .

B 3 + 4j

( +
6 2 5 j) ( 3 4
j).

—4
(3 + 41) (3 f)
28

— O 66 .
j .

If desired the quadrat u re n u mber may b e eliminated from


,

the numerator and le f t in the denominator by m u ltiplying with


the conj ugate number of the numerator thus ,

(G I + 30 2 ) ( 61 1
0

(bl
2 2
(11 +02

f or instance ,

(6 (6
( 3 + 3) (
4 6 — 2 5
j) .

2 8 + 1 6 5j .

30 . as in m u ltiplication the polar re presen t ation o f


J u st ,

the general n u mber in divi s i on i s more perspic u o u s than any


other.
44 E NGIN E E RING M A TH E AI ATI C S .

L e t A = a ( cos
'
a +j sin a ) be divided by B = b (c os 8 + j ,
sin (I) ,

th u s :

"
A a ( cos a + j sin a )
I3 b ( cos fi+ j sin 3)
a ( cos a + j sin a ) ( cos fl — j sin fl )

b ( co s fl + j sin fl ) ( cos I j
? — sin Q)
a t( cos a cos fl + sin a sin fl ) + j ( sin a cos 8
,
— cos a sin fl )
b ( cos fl + sin fl )
2 2

—B +
, )
( a
j sin ( a

That is general n umbers A and B are divided by dividing


,

their vectors or absol u te val u es a and b and s u btracting thei r , ,

phases or angles a and B .

I nvo l u ti o n a nd E vo lu ti o n of G e ne r al Nu m b e rs .

31 . Since
invol ution is multiple multiplication and e volu ,

tion is invol u ti on with fractional e x ponents both can be resolved ,

into simple e x pressions by using the polar f orm of the general


number .

I f,
A = a 1 + ja 2 = a ( cos a sin a ) ,

na + j sin na ) .

For instance i f ,

A = 3 + 4j = 5 (co s 5 3 deg + j sin 53 d e g ) ; .

C = A4 = 5 4 cos 4 X 53
( deg + j sin 4 X 53 deg ) . .

= 62 5 cos 2 1 2 deg + sin 2 1 2 deg


( j ) . .

= 62 5 — cos 3 2 deg — sin 3 2 deg


( j ) . .

62 5 ( —
O S4S
.

O S3O j)
.

—5 29 — 33 1
j.

If, A = a 1 + j a 2 = a ( Cos a +j sin a ) then


,

C= <VA= A c os
g + j sin
n
T HE GE NE R AL N UM BE R . 45

32in the polar e x pression o f A we increase the phase


. I f, ,

angle a by 2 7x or by any m u ltiple of 2 7r : 2 grr w here q is any


, ,

integer number we get the same val u e o f A thu s :


, ,

A = a {cos ( a + 2 q7z) + j sin ( a


since the cosine and sine repeat a fter every 3 60 deg or ,
22 .

The nt h root however is di fferent : , ,

G= VA = W cos

We hereby get n di ff erent val ues o f C f or q= 0 , ,


1, 2 . n 1
= n gives again the same as = 0 S ince it gives
q q .

a + 2 n7r a

that is an increase of the phase angle by 360 d e g w hich leaves


, ,

cosine and sine unchanged .

Th u s the nth root o f any general number has n di ff erent


,

values and these val u es have the same vector or absol ute
,

2 7:
term VB
n
b ut d i ffer from each other by the phase angle and
,
7
its m u ltiples .

For instance let A = ,



52 9 —
33 1 j = 62 5 (cos 2 12 deg .

j sin 2 1 2 deg ) then .


,

212 3 60q 2 1 2 + 36
C = VA = Z

sin 53 )
5 ( cos 5 3 + j —
3 + 4j

5 ( cos 1 3 + j sin 1 43 ) 5 ( c os 3 7
4 = — sin 3 7 ) + 3j

5 ( cos 2 33 j sin 2 33 ) 5 ( cos 53 j sin 5 3 )
+ = — — —
3 4j

5 ( cos 3 2 3 + j sin 3 2 3 ) 5 ( cos 3 7 j sin 3 7 )
— 4
= 5 ( cos 5 3 + sin 53 =3+4
= 5 cos 4 1 3
( sin 4 1 3 ) j ) j
The n roots of a general n u mber A = a (cos a +j sin a ) di ffer
2 7:
f rom each other by the phase angles — or l /nt h o f 3 60 deg ,
.

and si nce they have the same absol u te val u e {7 a it follo w s that , ,

they are represented by n eq u idistant p oints o f a circle with


radi us ( 7 5 as sho w n in Fig 22 f or n = 4 and in Fig 2 3 for
,
.
, ,
.
46 E NGI NE E R I NG M A T H E M A TI C S .

n= 9 . a system o f 71 eq ual vectors di ff ering in phase from


S u ch .
,

each other by l /nth of 360 d e g is called a polypha se s yste m or


, ,

an n— p ha s e s ys te m The n .roots of the general number th us

give an n—phase system .

33 For . instance Q/ T =?
,

I f A = o ( cos a + j sin a ) = 1 this means : a = 1 a = 0 ; and.


,

hence ,

FI G 22
. . B oot s of a Ge ne ral Num be r , n = 4.

and the n roots o f the unit are

3 60

cos 2X +1 s 1n 2X
n

3 60
=n- 1 cos ( — l) —
sm —l
q n
n
( n )

However ,
3 60 '

cos q7 2 + j sin q +j sin
48 E NGIN E E R I NG MA TH E M ATI CS .

S q u aring both e q u ations and adding them gives , ,

(1 1
2
+0 2
2
(0 1 2

Hence

2 — 2
and since 01 62 0 1 ;

whi ch is a rather comp licated e x pression .

3 5 Wh en representing
. physical quantities by general
numbe rs that is comple x quantities at the end of the calc u la
, , ,

tion the fi nal res ult us u ally appears also as a general number ,

or as a comple x o f gene ral numbers and then has to be reduced ,

to the absol ute val u e and the phase angle of the physical q u an
tity This is most conveni ently done by reducing the general
.

n u mbers to their polar e x pressions For instance if the resul t .


,

of the calculation appears in the f orm ,

(0 1 V01

(d l + j6 2 )
by subst i tuting

and so on .


(
c cos a sin a ) b (c os fl + j sin
3
( cos r + j sin
d ( c o s 8 j sin 5 ) e ( cos
2 2
e +j sin e )

r/2 23 e )+j sin r/2
-
25 e )l
.
TH E G E N E R A L N UM B E R . 49

Therefore the absol ute val ue o f a f ractional e x pres s i on Is


,

t h e prod u ct o f the absol u te val u es o f the f actors o f the n u mer

ator divided by the prod uct o f the abs ol ute val ues o f the
,

factors o f the denominator .

The phase angle o f a f ractional e xpression is the su m o f


the phase angles o f the factors o f the n u merator min u s the su m ,

of the phase angles of the factors of the denominator .

For instance ,

1 4 + j s in 1 1 4) i

1 25 ( cos 3 7 + j s i n

1 14
2 x 3 0? + 45
3

+ f sin 2 X 307 + 45 + 2 x 37
"
2 63 + 3 sin 2 63 }
— O I 22
.

O 99 2 j }
.

O 7 4j
. .

36 . As will be seen in C hapt e r I I :


2 3 4
u u u

x
2
x6 2
8
cos x
_ 1
l
4


3 5 7
x x x
srn x = r
|3 | 5 If
Here f rom f ollows by s u bstituting
, ,
ft = 0, u = 0
j ,
cos 0 +j sin 0=

a nd the polar e x pression of the comple x q uantity ,

A = a ( cos a sin a ) ,

thus can also be written in the f orm


"
,

A ae
= i
,
50 E NG I N E E R I NG M A T H E M A TI C S .

where e is the base o f the nat u ral logarithms ,

1 1 1
— +
Z
.

Since any n u mber a can be e x presse d as a po w e r o f any


other n u mber one can s u bstitute
, ,

a = e

1
2222
6
where = lo 0. and the general n u mber thus can
a0 ge
1
,

also be written in the form ,

A s
ac + ia .

that is the general n u mber or c omple x q u antity can be e xpressed


, ,

in the f orms ,

A 0 1

( cos sin )

a a a


a5 5

The last two or e x ponential forms are rarely used as they


, , ,

are less convenient f or algebraic operations They are o f .

importance however since solutio ns of di fferential eq u ations


, ,

f req uently appear in this f orm and then are reduced to the ,

polar or the rectangular f orm .

3 7 For instance the di ff erential eq u ation o f the d ist ri b u


.
,

tion o f alternating c u rrent in a flat cond u ctor or o f alternating ,

magnetic fl u x in a flat sheet of iron has the f orm : ,

d y z
— 2
.
2 10 y;
dz
"V”
y As
and is integrated by = w here ,
,

— j ) cx
y Al e :
+ (1

This e x pression red u ced to the polar f orm is


, ,

y A1
= cx

j sin ex ) +A cx +j sin ex ) .
TH E G E NE R AL N UM B E R .
51

L o g ari th m ati o n .

38I n taking the logarithm o f a general numbe r the

"
.

p e nt ial e x pression is most convenient , thus :


on

loge ( a l + 3a2 ) = log ( cos sin )



e a. a +3 a


lo ge a e
log e a Ho g.
“ s
f

log e a + ja ;

or if
,
b base of the logarithm f or instance , ,
b = 10 , it IS

logb ( a 1 + ja2 ) = lo f = lo
j lo g ,
a
gb a e
gb a + a 5
;
or if b unequal
,
1 0, reduced to lo g l o ;

lo g l o a lo g l o 6
log , ( c l + ja 2 ) + 3a
lo g l o b 10 g 1 0 b

NO T E . mathematics for qu adrature u nit


In ,
1 is al w ays
chosen the symbol 15 Since ho w ever in engineering the symbol 1
.
, ,
.

is u niversally used to represent electric cu rrent f or the q u ad ,

rature u nit the symbol j has been chosen as the letter nearest ,

in appear ance to 75 and j th u s is al w ays u sed in engineering


,

c al cul ati ons t o denote the quadrature u it


n — 1 .
C HAPT E R I I .

PO TE NTI AL SE RI E S AND EX PO NE NTI AL FUNC TI ON .

A . GE NE RAL .

39 . An e x press ion such as

(1)

repres ents a fracti on ; that is the res u lt o f di vision and li ke


, ,

any f raction it can b e calc u lated ; that is t he f ractional f orm ,

e liminated b v di viding the num erat or b v the denomi nator t hus :


, ,

1 — 3:

Hence , the f racti on ( 1 ) can also b e e xpressed in the f orm


1

This is an infinite series o f s u ccessive po w ers o f r,


. or a po te n

ti al s er ie s .

'

I n t he same manner by ,
di fl d ing thro u gh ,
the expression
1 o


1 +x

can be reduced to the in finite series ,

1 — x3 +
P OTE N TI AL S E R I E S AN D E X P ON E N TI AL F UN C TI ON . 53

The infini te series ( 2 ) or ( 4) is another f orm o f rep rese nta


tion o f the e x pression ( 1 ) or j ust as the periodic decimal
fraction is another representation o f the common f raction
( for ins tance 0 6363 . .

4 0 As the series contai ns an infinite numbe r o f terms


.
,

in calc u lating num erical val u es from s u ch a series per fect


e x actness can never be reached : since only a finite n u mber of
terms are calc u lated the res u lt can only be an appro x imation
, .

By taking a su ffi cient n umber o f terms of th e series however , ,

the appro x imation can be made as close as desired ; that is ,

numerical val ues may be calculated as e x actly as necessary ,

so that for engineering p u rposes the in finite series (2 ) or ( 4)


gives just as e x act num erical val u es as calcula tion by a finite
e x pression ( 1 ) or provided a su fficient n umber of terms
are used I n most en gineering calc u lations an e x actness o f
.
,

pe r cent is s u fficient ; rarely is an e x actness o f per cent


or even greater req u ired as the u navoidable variations in the ,

nature o f the materials used in engineering stru ctures and the ,

acc u racy o f the meas u ring instruments impose a limit on the


e x actness o f the result .

For the val u e = o5


the e x pression ( 1 ) gives
z ,
.
y
w hile it s representation by the series ( 2 ) gives
= 1
y (5 )
and the s u ccessive appro xi mations of the numerical val ues o f
y then are :
i
u s ng o ne te rm : y= 1 1, e rror : 1
t w t e rm
o s: y= l+0 5 .

three te rm s .
y=
fo r te rm
u s . 1 87 5 ,
.

fi ve t rm e s: y
= 1+0 . 06 2 5 =

It is se en that the successive appro x imations come closer and


closer to the correct val u e y= 2 b ut In this cas e al w ays remain , ,

belo w it ; that IS the series ( 2 ) approaches its limit f rom below


, ,

as shown I n Fig 2 4 in which the s u ccessive appro x imations


.
,

are marked by crosses .

For the value x = 0 5 the approach o f the successive .


,

appro x imations to the limit is rather slo w and to get an acc u rac y ,

of per cent that is bring the error dow n to less than


, ,

requires a considerable number of terms .


54 E NGIN E E RING M AT H E M A T I C S .

For x= 0 1 . the series ( 2 ) is


=1
y
and the successive appro x i mations th u s are
1 : y= 1 ;

and as by ,
the final or limiti ng val u e is
1 10
1 9

1
+ ” t
1 a ;

FI G 2 4. . C onve rge nt Se rie s w ith O e s


n -
id e d Appr oa ch .

the fourth appro x imation already brings the error well below
per cent and su fficient acc u racy th us is reached for most
,

engineering p u rposes b y u sm g fo u r terms of the series .

=
4 1 The
. e x pression ( )
3 gi ves for x 0 5 the val u e ,
.
, ,

1 2
_ “
0 6 666
1 3
Represented by series it gives
y 1

the successi ve appro x imations are ;


l st : y= 1 1, e rror : + 0 33 3 .

2d : = 1
y
3d = 1 0 5+ 0 25 = 0 75 ;
.
y . . .

4t h : 1 25
5
-

5th =l 0 5 + 0 25 0 6 2 5 = 0 6 87
.
y . . .
,

As seen , the successive appro x imations of this series come


closer and closer to the corre ct value y= 0 6666 but in this .

case are alternately above and below the correct or limiting


v al ue that is the series (4) approaches its limit f rom both sides
, , ,

as sho w n in Fig 2 5 while the series ( 2 ) approached the limit


.
,

f rom below and still other series may approach their limit
,

from above .
56 E NG I N E E RING M AT HE M A TI C S .

while the infinite series ( 4) gives


= 1 —
2 +4 —
32 +
y
and the s u ccessive appro x imation s o f the latter th u s are
1; —1 +3; —
5; + 11 —
21 ;
;
°

he nce while the s u ccessive val u es still are alternately above


,

and belo w the correct or limiting val ue they do not approach ,

it with increasing closeness b u t more and more diverge there ,

f rom .

S u ch a series in which the values derived by the calc u la


,

tion o f more and more terms do not approach a final val u e


closer and closer is called di vergent while a series is called
, ,

c onve rgent i f the successive appro x imations approach a final


val u e with increasing closeness .

4 3 W hi le a finite e x pression as ( 1 ) or
. holds good for ,

all val u es of x and numerical val u es o f it can be calc u lated


,

w hatever may be the val u e o f the independent variable x an ,

infinite series as ( 2 ) and


, f requently does not give a finite
res u lt f or every value of x b ut only f or val u es withi n a certain
,

range For instance in the a bo ve series for — 1


,
.
,

the series is convergent ; while f or val ues of 2: o u tside of thi s


range the series is divergent and th us u seless .

When representing an e x pression by an infinite series ,

it th u s is necessary to d etermine that t h e series is convergent ;


that is approaches with increasing n u mber o f terms a finite
,

limiting val u e otherwise the series cannot be u sed Where


,
.

the series is convergent within a certain range of x diver ,

gent o u tside o f this range it can be u sed only in the rang e of


,

co nve rge nc y b u t o u tside of this ra nge it cannot be used for


,

deriving n u merical values b u t some other form of representa ,

tion has to be f o u nd which is convergent .

This can f req u ently be done and the e x pression thus re pre ,

sented by o ne series in one range and by another series in


1
y
0

another range For I n stance the e x press i on


O

.

1 + :c
,

l
the f orm
C

can be tten
O O

s u b st I t u t I ng , e In

=
y
P O TE N TI AL S E RIE S AND E X P ONE N TI AL F UN C TI ON . 57

and then developed into a series by dividing the n u merator


by the denominator which gives ,

= u — u2 + u3 — u4 +
y .

or ,
re s ub st it ut ing x ,

1 1 1 1
2 3
a: x x

whi ch is convergent f or x =2
,
and for x =2 it gives
y + 01 2 5

With the su ccessive appro x imations :

which approach the final limiting value ,

=
y 0 33 3 . .

44 . An infinite series can be used only i f it is convergent .

Mat he m e t ic al methods e x ist f or determining a series w hether


is convergent or not For engineering p u rposes however
.
, ,

these methods usually are unnecessary ; for practical use it


is not sufficient that a series be convergent but it m u st con ,


verge so rapidly that is the s u ccessive terms of the series
,


must decrease at such a great rate that acc u rate numerical
results are derived by the calc u lation o f only a very f ew terms ;
t w o or three or perhaps three or f our
,
This f or instance .
, ,

is the case with the series ( 2 ) and ( 4) for x = 0 1 or less For . .

x=0 5 the series ( 2 ) and (4) are still convergent as seen in


.
, ,

5 ) and b ut are u seless for most engineering p u rposes as


, ,

the successive terms decrease so slo w ly that a large number


o f terms have to be calculated to get accurate res u lts and f or ,

such le ngthy calc u lations there is no time in en gineering w ork .

If however the successive terms of a series decrease at such


, ,

a rapid rate that all b ut the first few terms can be neglected ,

the series is certain to b e convergent .

I n a series therefore in which there is a q u estion w hether


,

it is convergent or divergent as f or instance t he series ,

1 1 1 1 1
— +
+2
1 5 6 (d 1 ve rge nt ) ,
58 E NGIN E E RING M A TH E M A TI C S .

1 1 1 1 1
y
= 1 +
3 4
( convergent ) ,

the matter o f conve rge ncv is o f little importance f or engineer


ing calculation as the ser1 es is u seless in any case ; that is does
, ,

no t g ve acc u rate n u merical res u lts w ith a reasonably moderate


i
amo u nt of calc u lation .

A series to b e usable for en gineeri ng work m u st have


, ,
.

the successive terms decreasing at a very rapid rate and i f ,

this is the case the series is convergent and the mathematical


, ,

investi gations o f convergency th u s u s u ally becomes unnecessary


in engineering work .

4 5 I t wo u ld rarely be advantageous to develop s u ch simple


.

C x pressions as ( 1 ) and ( 3 ) into infinite series s u ch as ( 2 ) and ,

since the calc ulation o f n u meri cal val u es from ( 1 ) and (3 )


is simpler than from the series ( 2 ) and even tho u gh very
fe w terms o f the series need to be u sed .

The u se o f the series (2 ) or ( 4) instead of the e x pressions


( 1 ) and (3 ) therefore is advantageo u s only i f these series con
verge so rapidly that only the first two terms are required
for n u merical calc u lation and the third term is negli gible ;
,

t hat is f or very small val u es o f x


,
Thus f or x = 0 0 1 accord.
,
.
,

ing to
=1 1
y

as the ne x t term ,
is already less than per cent o f
the val u e o f the total e x pression .

For very small val u es o f 13 there f ore by ( 1 ) and


, ,

and by (3 ) and
1
_
1 +x

a nd t h ese e x pressions ( 1 0 ) and ( 1 1 ) are use f ul and very com


mo u ly u sed in en gineering calc ulation f or simpli f ying work .

For instance i f 1 pl u s or min u s a very small q u antity appears


,

as f actor in the denomi nator of an e x pression it can be replaced ,

by 1 min u s or pl us the same small q u antity as factor in the


numerator of the e x pression and inversely , .
P O TE N TI AL S E RIE S AN D E X P ONE N T I A L F UN C TI ON . 59

For e x ample i f a direct current receiving circ uit o f resist


,
-
,

ance r is fed by a s u pply voltage e o over a line o f lo w


,

resistance ro w hat is the voltage e at the receiving circuit ?


,

The total resistance is r + r0 ; hence the current


, ,
i

and the voltage at the receiving circ u it is

e = ri = e0

If now ro is small compared w ith r, it is

c = eo

As the ne x t term of the series would be the err0 1

made by the simpler e x pression ( 1 3 ) is less than Thus ,

i f ro is 3 per cent of r which is a f air average in interior light


,

ing circ u its ,


or less than pe r cent ;
hence is us uallynegligible
,
.

4 6 I f an e x pression in its finite form is more complicated


.

and thereby less convenient for n u merical calc u lation as f or ,

instance i f it contains roots development into an infinite series


,

freq u ently simplifies t h e calc ulation .

Very convenient for development into an infinite series


o f po w ers or roots is the b i nomi al theorem
, ,

” n—
( Ufn — Q
)

where

Thus for instance in an alternating c urrent circuit o f


, ,
-

resistance r reactance x and s u pply voltage e t he c urrent is


, , , ,
60 E NGIN E E RING III A TH E M A TI CS .

I f t his
circuit is practically no n ind u ctive as an incandescent
-
,

lighting circ u it ; that is i f x is small compared with r ( 1 5 )


, ,

can be w ritten in the form ,

and the sq uare root can be developed by the binomial th u s ,

1
I
n :
and g1ves
2

2 4
3 5
1 +
8 16

In this series if x = 0 1 r or less ; that is the reactance


.
,

is not more than 10 per cent o f the resistance the third te rm , ,

is less than per cent ; hence negligible and the


, ,
8
series is appro x imated w ith s u ffi cient e x actness by the fi rst
two terms ,

and equation ( 1 6) o f the c u rrent then gives

Thi s e x pression is simpler for n u merical calculations than


the e x pression as it contains no sq uare root .

4 7 D evelopment into a series may be come necessary


. if ,

further operations have to be carried o ut with an e x pression


for whi ch the e x pression is not suited or at least not well s u ited
,
.

T his is often the case w here the e x pression has to be integrated ,

since very few e x pressions can be integrated .

E x pressions under an integral si gn there f ore very commo nly


have to be developed into an infinite series to carry out the
integration .
POTE N T IAL S E R IE S AN D E X PON E N TI AL F UN C TI ON . 61

E X AMP LE 1 .

Of the eq uilateral hyperbola ( Fi g .

2
lly
fi (1 ) (2 0)
the len gth L o f the arc be t w een x l = 2 a and x g = 1 0a is to be
calc u lated .

An element dl o f the arc is the hypothenuse o f a right triangle


with dx and dy as cat he t e s I t there f ore is
.
, , ,

FI G 2 6
. . E qu ilate ral Hype rbola .

and from

S u bstit u ting (22 ) in (2 1 ) gives ,

hence the length


,
L of the are , from ml to 1 2 is ,
62 E NGINE E RING M ATH E M A TI CS .

S u bstit u ting
2 = v
; that is ,
(i i : a d v, also s u bstituting

2 and v2

1
1 db .

The e x pression under the integral is inconvenient for integra


tion ; it is pre f erably d eveloped into an infini te series by the ,

binomial theorem

u and n= l )
2

1 1 1 5
8 12 16
+
2v 4
8v 1 6v 1 28v

1 1 1 5
+ 12 16
2v 4
8i 8
1 65 1 2 811
1 1 1
— av 1

f 1 1 1 l 1
w — ”
M E
C

56

"
l — b

17 6 1
11

and s u bstit u ti ng the n u merical val u es ,

— O I 25
(10 ( .

0) — — o 000 1
0078 (
1 76
.

a {8 8 o2 06a
. .

As seen , in this series , o nly the first two terms are appreciabl e
in val u e the third term less than
,
per cent of the total ,

and hence negli gible there f ore the series converges very
,

rapidly and n u meri cal values can easily be calculated by it


,
.
64 E NGIN E E R I NG M ATH E IV
I ATI CS .

instead of the nat u ral logarithm as used above the


If , ,

decimal logarithm is req uired the follo w ing relation may be ,

a pph e d :

lo gl o e = lo
g l os log e
. a= 0 43 43
. log . a,

lo g l ois e x pressed by log.


a a, and th u s ( 2 0 ) (2 1 ) ass u me
the form ,

10 g 1 0 ( 1 + 13) x

or appro x imately
, ,

=
10 g 1 0 ( 1 + l ) 0 43 43 x ,

or more acc u rately


, ,

lo g o (1 .
)
v 0 43 43x.

B . DI FF E RE NTI AL E QU ATI O NS .

49 The representation by an infinite series is of special


.

val u e in those cases in which no finite e x pression of the f unc


,

tion is known as f or instance i f the relation between a: and y


, ,

is given by a di fferential equ ation .

Di fferential equations are solved by separating the variables ,

that is bringing the terms containing the one variable y on


, , ,

one side of the equ ation the terms with the other variable a: ,

on the other side of the eq u ation and then separately inte grat ,

ing both sides of the eq u ation Very rarely ho w ever is it .


, ,

possible to separate the variables in this manner and where ,

it cannot be done u s u ally no systematic method of solving the


,

di fferential eq uation e x ists b u t this has to be done b y trying ,

di fferent f u nctions until one is f ound which satisfies the


,

e q uation .

I n electrical engineering c u rrents and voltages are dealt ,

with as f u nctions o f time The current and e m f gi ving the . . . .

power lost in resistance are related to each other by Ohm s ’

law . Current also prod u ces a magnetic field and this magnetic ,


field by its changes generates an e m f the e m f o f self . . . . . .

ind u ctance I n this case e m f is related to the change of


.
,
. . .

c u rrent ; that is the di fferential coe ffi cie nt o f the c u rrent and
, ,

th us also to the di fferential coe fficient o f since the e m f . . .


POTE N TI AL S E R I E S AN D EX P ONE N TI AL F UN C TI ON . 65

is related to the c u rrent b y Ohm s law I n a condenser the ’


.
,

c u rrent and there f ore by Ohm s law the


,

depends u pon ,

and is proportional to the rate o f change o f the e m f impressed . . .

u pon the condenser ; that is it is proportional to the di fferential


,

coe ffi cient o f e m f


. . .

There f ore in circ u its havi ng resistance and ind u ctance


, ,

or resistance and capacity a relation e x ists bet w een c u rrents


,

and and their di ff erent ial coe fficients and in circ u its ,

having resistance ind u ctance and capacity a do u ble relation


, ,

o f this kind e x ists ; that is a relation be tw een c urrent or e m f


, . . .

and their first and second di ff erential coe fficients .

The most common di fferential eq u ations o f electrical engineer


ing thus are the relations between the f u nction and its di fferential
coe fficient which in its simplest f orm is
, ,

dy
y
dr

(2 7 )

and where t he circ u it has capacity as w el l as ind u ctance the ,

second di ff erential coe fficient also enters and the relati on in ,

its simplest form is ,

(2 9)

and the most general f orm of this most common di ff erential


e qu ation o f electrical engineering then is ,

d y 2
dy
dx 2
+ 2 c
d
-
( 30 )
The di fferential equ ations (26 ) and (2 7 ) c an easily be inte
gr ated by separating the variables b u t not so w ith eq u ations ,

(2 9 ) and the latter are pre f erably solved by trial .

5 0 The general method of sol u tion may be ill u strated with


.

the equation (2 6)
66 E NGINE E RING M A TH E M ATI CS .

To determine whether this eq u ation can b e integrated by an


infinite series choose s u ch an infinite series and then by su b
, , ,

st it u t ing it into eq u ation ascertain w hether it satisfies


the eq u ation that is makes the le f t side eq u al to t he right,

side for every val u e o f r . .

Le t ,

be an infinite series of w hich the coe fficients (t o a l a 2 a3


, , , ,
.

are still unknown and by substit u ting ( 3 1 ) into the di fferential


,

e q u ation determine whether s u ch val u es of these coe fficients


can be f o u nd which make the series (3 1 ) satis f y the e quation
,

Di ffere ntiating (3 1 ) gives ,

dy 2 3
a1 + 2 a g x + 3 a 3 4: + 4a 4r
dx

The di fferential equ ation ( 2 6 ) transposed gives ,

Substituting (3 1 ) and (3 2 ) into and arrangi ng the terms


in the order of 23 gives , ,

(a l ao ) (2a 2 al )m (3 a 3 a2 x ) 2

( a4
4 a3 r) 3
( 5as )
a4 x
4
0 .
(3 4)
then the above series (3 1 ) is a sol ution of the di fferential
If
equation the e x pression (3 4) m u st be an identity ; that is ,

must hold for every val u e o f x .

I f ho w ever it holds f or every val u e of x it does so also


, , ,

for m= 0 and in thi s case all the terms e x cept the first vani sh
, , ,

and ( 3 4) becomes ,

— ao = 0 ; a1 = a0
al (3 5 )

OI ,
.

To make (3 1 ) a sol u tion of the di ff ere ntial eq u atio n (a l a o)


m u st there f ore eq ual 0 This being the case the term ( l a o)
. c —
,

can be dropped in w hi ch then becomes ,

a1 ) x “
f“ "

1 (4 0 4 (1 9 25 3

x {(2 a g -
a1 ) ( —
3a3 a 2 x ) + = 0 .
P OTE N TIAL S E RI E S AND E X P ON E N TI AL F UN C TI ON . 67

this eq uation m u st h old f or every valu e o f x the second


Since ,

factor of the equati on must be z ero sinc e th e first factor x is , , ,

not necessarily z ero Thi s gives .


,

( 20 2 0 1 ) ‘

f (3 0 3
'
l
G a l? )
(I 3 :IZ
2
0 .

As this eq uation holds f or ever y val ue o f i t holds also f or x,

In t hi s case however all terms e x cept the firs t vanish


, , ,

and ,

2a2 a1 0

a nd from

Contin u ing the sa me reasoning ,

3 a3 ( 12 0, 4a 4 a3 0, etc .

"
There f ore ,
if a n ex pres s i on o f s u cce ssi ve o w
p e rs o f 33 ,
su ch a s

is an i de nti ty, tha t i s , holds f or e ve ry valu e of x ,


the n all

the coe fi ci e nts of all the p owe rs of 2 m u st se


parately be zero .

Hence ,
ao O;

2a2 — a1 =0
;

3 a3 — a2 =0
;

40 4 G3 = O,

Th e re ad r
t r ali e t h diff r nce b tw n n xp r ion
e m us e z e e e e ee a e e ss as
e q atio n in x
u s b s tit tio n p rod u ct of a f n ctio n ;
, a nd a s u that i s n u u s, a a
id ntity
e .

R e gardl of t he val e of t h e coe ffi cie nt an e xp r ion ( 34) as e q ation


e ss u s s, e ss u

g iv a n m b r of e parate val e s of t h root of t h q atio n w hi ch


es u e s u x, e s e e u ,

m ak t h l e ft id of ( 3 4 ) q al
e e s ro that i olve t h q ation I f ho w e ve r
e e u ze , s, s e e u .
, ,

t h e i nfi nite ri e ( 3 1 ) i s a ol tio n of t h di ff e r ntial e q atio n


se s sthe n u e e u
th e ex pre ion ss w hich i t h r lt of s b tit ti ng ( 3 1 ) i nto s m t e e su u s u us
b e co e ct n t o nly for a li m it d n m be r of val s of
rr o w hich e th roote u ue x, ar e s
of t he e qu ation b t for all val ,
of 2 that i n m atte r w hat val i
u ue s ,
s, o ue s
chose n for x t he le ft sid e of ( 3 4) m t alw ay give t he sa m e re lt 0 th t
,
us s su , ,
a

i it m ust n t b change d b y a cha nge of


s, o e or in oth e r w ord s it m u t n t 2 , , s o
conta in z he nce all t he coe fficie nts of t he po w rs of m st b e e ro
,
e 2: u z .
68 E NG I NE E R I NG M ATH E M A TI C S .

There f ore i f the coe fficients of the series (3 1 ) a re chosen


,

by eq u ation this series satisfies the di fferential e q u ation


that is ,
2 3 4
313 x x
1 — + +

is the solution of the di ff erential equation ,

dv
da:

51 . In the same manner ,


the di ff erential e quation
dz
dx
is solved b y an infinite series ,

(40 )
and the coe fficients of this series determined b y substituting
( 40) into in the same manner as done above This gi ves .
,

(0 1 ad o) (3 613 aa 2 x ) 2

( 4
4 a — a a3 x ) 3 = 0, (41 )
and as this e qu ati on must be an identity all its coe fficients
, ,

must be z ero ; that is ,

al — aao = 0; or a 1 = aa0 ;
2
a £1

2 a 2 aa l = 0; or a2 = a1
— =
ao
2 2
3
a a

3 a3 aa 2 = 0; or a3 = a2 — = a0
3
| 3
a G4

4a 4 d a g =0 or a 4 = a3
- =
; a0
4
L
4
etc .
,
etc .

and the sol ution o f di fferential eq u ation (3 9 ) is ,

2 2 3
3 4 4
a x a x a r
z = ao 1 + ax +
2
H
These soluti ons (38) and
52 . o f the di fferential eq ua
,

tions (2 6 ) and are not single solutions b ut each contains ,

an infinite number of sol utions as it contains an arbitrary ,


P OTE N TIAL S E R I E S AN D E X P ON E N TI AL F UN C TI ON . 69

constant d o ; that is a constant w hich may h ave any d esired


,

numerical val ue .

This can easily be seen since i f z is a sol ution of the di f


, ,

fe re ntial eq u ation ,

then any m ultiple or fraction o f


, ,
2, be , also is a sol ution of the
di fferential eq u ation ;

si nce the b cancels .

Such a cons tant d o which is not determined by the c oe ffi


, ,

cie nt s of the m athemati cal problem b u t is left arbitrary and , ,

requires for its determinations some further condition in


addition t o the di ffere ntial equation is called an i ntegrati on ,

cons tant .It usually is determined by some additional req u ire


ments of the physi cal problem which the di ffere ntial eq uation,

represe nts ; that is b y a se called termi nal condi ti on as f or


,
-
, ,

instance by having the val u e of y given f or some particular


,

value o f at us u ally for x = 0 or x = oc


, ,
.

The di fferential equation ,

th us is solved by the f unction


, ,

=a
y og o,

2 4
x 33
+

and the di fferential eq u ation ,

is solved by the function ,

2 0 02 0 ,

3 x3 4 4
d a st
1 ax +
I
3 lit
70 E NGINE E RING M A TH E M A TI CS .

yo and 2 0 thus are the simples t f orms o f the sol u tions 3] and z
of the di ff erential eq u atio ns (2 6) and
5 3 I t is interesti ng ne w to determine the value of
. To
raise the infinite series which represents yo to the nt h ,

power wo uld obvio usly b e a very complicated operati on


,
.

However ,

and since from (44)

by substituting ( 5 1 )

hence the same equation as


,
but with y instead o f ”
z .

Hence if yis the solution of the differential e quatio n


, ,

dy
31,
da:

then z= is the solution of the di fferential equation


n
y
dz
dr

However , the solutio n of t his di ffere ntial e quatio n from


and ( 49 ) is

= 1 + nx + +
20 0 0 0
;

t hat is i f .

2 3
x x
o

therefore the series y is raised to the nt h power by multipl y


ing the variable x by n .
72 E NGI NE E RING IlI A TH E MA TI CS .

and most f req u ently u sed methods o f solving e ngineering


problems .

E X AMP LE 1 .

a 4 pole
In 5 0 0—volt 5 0—kw direct c u rrent shunt motor
54 .
-
.
-
,

the resistance o f the field circ u it incl usive of field rheostat is , ,

2 5 0 ohms E ach fie ld pole contains 4000 turns and produces


.
,

at 5 00 volts impressed upon the field circ u it 8 m e galines o f ,

ma gnetic flux per pole .

What is the e q u ation of the field current and how m u ch ,

time a f ter closing the field s w itch is req uired f or the field cur
rent to reach 90 per cent o f its final val u e ?
Le t r be the resistance o f the field circ u it L the inductance ,

of the field circ u it and i the field current then the voltage , ,

cons u med in resistance is ,

In
general in an e l ectric circ u it the current produces a
, ,

magnetic field ; that is lines o f magnetic fl u x s u rro u nding the ,

conductor of the current ; or it is usually e x pressed i nte rli nked , ,

wi th the cu r rent This magnetic field changes with a change of


.

the current and us ually is proportion al thereto A change


,
.

of the magnetic field s u rro u nding a cond u ctor however gen , ,

e rat es an e m f in the cond u ctor and this e m f is proportional


. . .
,
. . .

to the rate of change of the magnetic field ; hence is pro ,

portional to the rate o f change o f the current or to ,

di

with a proportionality factor L which is ca lled the ,
i ndu ct
dt
a nce of the circ u it . This counter generate d e m f is in oppo
-
. . .

sit ion t o the current —L and thus co nsumes an


,

di
L which is called the e mf consu med b y self-i nductance,
II?
. . .

or i nd uc ta nce e . mf . .

There f ore by the ind u ctance L of the field circ uit a voltage
, , , ,

is consumed which is proportional to the rate o f change o f the


field current thus , ,
P O TE N TI AL S E R IE S AN D E X P ON E N T I AL F UN C TI ON . 73

Since the s u pply voltage and th u s the total voltage cons u med
,

in the field circuit is e = 500 volts


, ,

or rearranged
, ,

Substit u ting herein ,

= — ri
u e ;

This is the same di ff erential eq uation as with a

and there f ore is integrated by the f unction ,

i t
u = a0 € L

therefore ,
re s ub st it ut i ng from
-
t
6 — rt = a0 £ L

This solution still contains the un kno w n q uantity ao ;


or the integration constant and this is determined by kno w
, ,

i ng the c u rrent i f or some partic u lar val u e o f the time t .

Be f ore closing the field s w itch and thereby impressing the


voltage on the field the field c u rrent o bvio u sly is z ero I n the
,
.

moment of closing the field switch the current th us is sti ll ,

z ero ; that is ,

i = 0 for t = 0 .
74 E NG I N E E R I NG AI A TH E MA TI CS .

Su bstit u ting these val ues in ( 65) gives ,

or ao + e,

is the final solution of the di fferential eq u atio n t ha t is ,

it is the val u e of the field c u rrent i as fu nction of the time t , , , ,

a fter closing the field sw itch .

After infinite time t = oo the c u rrent i ass u mes the final


, ,

val u e i o whi ch is given by s u bstituting t = oo into equation


,

th u s ,

2 amperes ;

hence by substit u ting ( 68 ) into


,
this e quation can also be
written ,

(69 )

where i o = 2 is the final value ass u med by the field current .

The time tl a f ter whi ch the field c u rrent i h as reached 90


,

per cent o f its final val u e i o is given by s u bstituting


,
0 9i o .

into thus ,

0 9i 0 = i 0 1
.
( a

e L

Taking the logarithm o f both sides ,

T
— l1 l0 g — l
;
z
P O TE N TI AL S E RIE S AND E X PONE N TIAL F UN C TI ON . 75

55 The ind u ctance L is calc ulat ed f rom the data given


.

in the problem I nd u ctance is meas u red by the n um ber o f


.

inte rlinkage s o f the electric ci rc u it with the magnetic flux ,

prod u ced by one absol ute unit o f c u rrent in the circ u it ; that
is it equ als the prod u ct o f magneti c flu x and n u mber of turns
,

di vided by the absolute c u rrent .

A c u rrent of i 0 = 2 amperes represents absol ute units ,

since the absol ute unit o f c u rrent is 1 0 amperes The n u mber .

o f field t u rns per pole is 4000 ; hence the total n u mber of t u rns ,

The magnetic fl u x at f u ll e x citation ,

or i 0 absol ute units of c u rrent is given as 4) = 8 lines ,

of magnetic force The inductance of the field thus is


.

MD 1 6000
6 0 X 10
4 9
absolute units 64011,
02
\

the practical u nit o f ind u ctance or the henry (h) being 1 0 ,


9

absolute u nits .

Su bstit u ting L = 6 4O r = 2 5 0 a nd e = 5 00 into e q u ation ( 6 7 ) ,

and ( 70) gives


= — "
i 2( 1

6 40
SCC '
2 5 0 >< o 43 43.

Therefore i t takes abo u t 6 se e be f ore the motor field has .

reached 90 per cent of its final val u e .

The reader is advised to calc u late and plot the n u merical


values of i f rom equ ation for
t= o , 3 4 5 6 8 1 0 sec , , , , ,
.

This c alculation is best made in the form o f a table th us ; ,

log e

0 3 9i log = 0 1 69 4t;


. .
,

5
03
N —0 1 694t . .
76 E NGIN E E RING III A TH E M A TI C S .

“° 3 9‘
The val ues o f e can also be taken directly from the
tables o f the e xponential f unction at the end o f the book ,
.

— 0 3 9t
5
.

_ — O 39t
0 1 694 t
.
— 0 1 69 4t
. 1 e
.

= N — 0 1 694t .

E X AM PL E 2 .

56 . A condenser of 2 0 m f capacity , is charged to a potential


.

of e0 = 1 0 000
volts and then discharges thro u gh a resistance
, ,

of 2 megohms W hat is the eq u ation o f the discharge c u rrent


.
,

and a fter how long a time h as


the voltage at the condenser
dropped to its initial valu e ?
A condenser acts as a reser
voir of electric energy similar ,

to a tank as water reservoir .

I f in a w ater tank Fig 2 7 A ,


.
,

is the sectional area of the tan k ,

e the height of water or water , ,

press u re and water flo w s o ut ,

of the tank then the height e ,

decreases by the flow o f water ;


FI G 2 7 W at e r R e se rvoir
. . .

that is the tan k empties and ,

the c u rre nt o f water i is proportional to the change of the


, ,

d
water level or height of w ater ,
3 5 ,
and to t he are a A of the
tank ; that is it is , ,

( 72 )

The min u s sign stands on the right hand side as for positive -
,

i; that is o ut flow the height of the water decreases ; that is


,
-
, ,

de is negative .
P O TE N T IAL S E RIE S AN D E X P ONE N TI AL F UN C TI ON . 77 .

an electric reservoir the electric press u re or voltage 6


In ,

corresponds to the water press u re or height o f the w ater and ,

to the storage capacity or sectional area A o f the water tank


corresponds the electric storage capacit v of the condenser ,

called capacity C The current or flow o u t of an electric


.

condenser thus is , ,

i = — C

The capacity of c ondenser is ,

C = 20 mf 20 x 1 0 ‘
6
farads .

The resistance o f the discharge path is ,

r= oh m s ;
6
2 >< 1 0

hence the c u rrent taken by the resistance


, ,
r, is

and th u s

There f ore f rom ( 60 )


,

and for t= 0, e = e0 olts ; hence


v

e =e s
O

-o2 st
volts ;
of the initial val u e :
e = 01
%,
is reached at
92 sec .
78 E NG I NE E R I NG M A TH E M A T I C S .

The reader is advised to calc u late and plot the n u merical


val u es o f e from e q u ation
,
f or
t = 0 ; 2 ; 4 ; 6 ; 8 ; 1 0 ; 1 5 ; 2 0 ; 30 ; 40 ; 60 ; 80 ; 1 00 ; 1 5 0 ; 2 00 sec .

57 Wherever in an electric circ uit in addition to resistanc e


.
, ,

i nd u cta nce and capacity both occ u r the relations between ,

c u rrents and voltages lead to an eq u ation containing the second


di fferential coe fficient as disc u ssed above,
.

The simplest form o f s u ch eq u ation is :

t
I

To integrate this by the m ethod o f indeterminate coe fficients ,

we ass um e as sol u tion o f the eq uation ( 7 6 ) the infinite series ,

4
y
z
ao + a 4x + .

in which the coe fficients ao a l (1 2 , , ,


a3 , a 4 . are indeterminate .

D i fferentiating ( 7 7 ) t w ice gives ,

and s u bstit u ting ( 7 7 ) and ( 7 8 ) into ( 7 6 ) gives the identit y ,

2 3
2 a 2 + 2 X 3 a3 x + 3 >< 4a 4 3r + 4 >< 5 a s r
3
(
a a0 + a 1 x + a gx + a 3 3t + 2

or arranged in order of
,
at,

(2a 2 )+
a ao x ( 2 >< 3 a 3 aa 1 )+ x
2
( 3 X 4a 4 aa z )
— aa
g ) 0 .
( 7 9)

Sincethis eq uation ( 7 9 ) is an identity the coe fficients of ,

all po w ers of it must individ u ally equal zero This gives for .

the determination o f these hitherto indeterminate coe fficients


the eq u ations ,


2a 2 ae o = 0;
2 X 3 a3 — aa 1 = 0;
3 X 4a 4 — aa 2 = 0;
4 X 5a5 a a3 = 0, etc .
80 E NG I NE E R I NG MA TH E M A TI C S .

In this case instead o f the integration constants ac and a l


, ,

the t w o ne w integration constants A and B can be introd u ced


by the e q uations
a0 =A+B and

A and B
2 2
and substit u ting these into equatio n
,

b 2
x
2
5
3
13
3 b 4x 4
y=A 1 + b ll + —
3
| 4
2
2 2
53
1
5
b 4
x
4

1 b2 + +
+
2 . lg If
The first series how ever from f or
, ,
n= b is and
= — b
the second series from for n b is s ”
.

There f ore the infinite series (83 ) is


, ,

— bx
y Ae
= + bx
+ Be ;

that is it is the sum o f two e x ponential functions


, ,

a positive the other w ith a negative e x ponent


,
.

Hence the di ff erential eq u ation


, ,

2
d y
da: 2

is integrated by the f u nction ,

t bz —bx
4s + Bs
--

y
:

,

b \/a .

Ho w ever if is a negative q u antity


,
a ,
b = \/ a imaginary ,

and can be represented by

where

In this case eq u ation (86 ) ass u mes the f orm


, ,

y Ae
= + 5€ f
P OTE N TI AL S E RI E S AND E X PONE N T IAL F UNC TI ON . 81

that is if in the di fferential eq uation ( 7 6 ) a is a positive q uantity


, ,

+b 2
this di fferential eq uation is integrated by the su m o f
,

the t wo e x ponential f u nctions i f ho w ever a is a negative , ,

quantity — 0
2
the sol ution (86 ) appears in the f orm o f e x ponen
, ,

tial f u nctions w ith imaginary e x ponents


5 8 I n the latter case a form of the solution of di fferential
.
,

equation ( 76) can be derived which does not contain the


imaginary appearance by t u rning bac k to eq uation ,
and
s ubstituting therein a = —
c
2
which gives , ,

2
d y
?
d z

6
6
0 15
—a 1
y
H
o
2
I9 .

or wri ting A = a 0 and


,
B

2 2 4 4
0 33 0 23

3 3 5 15
c x 6 1
+8 cx

E I
5

The sol ution then is given by the su m of two infinite se ri es ,

thus ,

4 4
0 x 0 33 c a

5 x5
c
v(cx ) = cx
| 3
I5

y Au (c1: ) B v(cx ) .

the u series a change o f the sign


In -
,

the val u e of u ,

=u
u ( ex ) ( + cx) .

Such a f unction is called an even function .


82 E NG I NE E RING M A T H E M AT I C S .

In the v scries a change o f


-
, t he sign o f x reverses the sign
of v ,
as seen f rom

v( — c:c) — v
( + cx ) .

Su ch a f u nction is called an odd f u nction .

I t can be shown that

n ( er ) = cos ca: and v( cx )

y A
= cos +B sin
cx car, ( 98 )

where A and B are the integration constant s w hich are t o b e -


,

dete rmined by the terminal conditions o f the physical problem .

Theref o re the sol ution o f the di fferential eq u ation


,

has t wo di fferent forms an e xponential and a trigonometric


,
.

If a is p ositive ,

y Ae
= + bz bx
+ Bs

If a is negative ,

y A cos
= cx +B sin cr .

In the latter case the sol ution ( 1 0 1 ) w o u ld appear as


, ex

o ne nt ia l f u nction w ith imaginary e xponents ;


p

—j oz
( 1 0 4)

As ( 1 04) obvio usly m u st be the same f u nction as it


follo ws that e xponential f u nctions w ith imaginary e x ponents
m u st be e xpressible by trigonometric f u nctions .
84 E NG I N E E RING MA TH E III ATI C S ’
.

Adding and s u btracting gives respectively ,


5
cos jv 9

0 0 6
sm v
y .

23

these equations ( 1 0 6) to
By ,
e x ponential functions
with imagi nary e xponents can be trans f ormed into trigono
metric f unctions w ith real angles and e x ponential f u nctio ns ,

w ith real e x ponents into t rigno m e t ric f u nctions with imaginary


angles and inversely
,
.

Mathematically the trigonometric f uncti ons thus do not


,

constitute a separate cl as s o f f u nctions b u t may be cons idered ,

as e xponential f u nctions w ith imaginary angles and it can b e ,

said broadly that the sol u tion of the above di fferential eq u a


tio ns is given by the e x ponential f unction b u t that i n thi s ,

f unction the e x ponent may be real or may be i maginary and , ,

in the latter case the e xpression is put i nto real f orm by intro
,

d u cing the trigonometric f u nctions .

E X AMPLE 1 .

60 . A conden s er (as an
ndergro u nd high pote ntial cable )
u -

o f 2 0 m f capacity and of a voltage o f eo = 1 0 000 discharges


.
, , ,

t hro u gh an ind u ctance o f 5 0 m h and o f negli gible resistance


.
,

What is the eq uation o f the discharge c u rrent ?


The c u rrent cons u med by a condenser o f capacity C and
potential di fference 6 is proportional to the rate of change
o f the potential di ff erence and to the capacity ; hence it is
, ,

d
f
,

C —

and the c urrent f rom the condens er ; or its discharge
at

c urrent is ,

i = C (11 1)

The voltage cons umed b v an ind u ctance L is proportional


to the rate o f change o f the c u rrent in the inductance and to t he ,

ind u ctance ; hence ,


POTE N TI AL S E R IE S AN D E X P ONE N TI AL F UN C TI ON . 85

Di ff erenti ating ( 1 1 2 ) gives ,

de d i
2
L z’
dé dt


and s u bstit u ting this into ( 1 1 1 ) gives ,
'

d z
t d% 1
1 CL
E
'
dt CL 2

as the di ff erential eq uation o f the problem .

This equ ation ( 1 1 3 ) is the same as for


is solved by the e xpression ,

' t t
t = A cos +B si n 1 14
x/ L C x/ L C

and the potential di fference at the condenser or at the ind u ctance


is, by substit u ting ( 1 1 4 ) into
t i

x/ L C x/ L C

These equations ( 1 1 4) and ( 1 1 5 ) still contain t w o u nknown


constants A and B which have to be determined by the terminal
, ,

conditions that is by the kno w n conditions o f current and


, ,

voltage at some particular time .

At the moment o f starting the discharge ; or at the time ,

t = 0 the c u rrent is z ero and the voltage is that to which the


, ,

condenser is charged that is i = 0 anl e = eo


, , ,
.

Substituting these values in eq u ations ( 1 14) and ( 1 1 5 )


gives ,

hence
B 60
L y

and s u bstit uting f or A and


,
B the val u es in ( 1 1 4) and ( 1 1 5 )
gives
86 E NGINE E R I NG M A TH E M ATI C S .

Su bstit u ting the n u merical val u es , 6 0 volts C = 2 0


,

mf 20 X 1 0
.
6
farads , L 50 m h 0 05h . . . gives ,

and VEZ = 1 0 —3
;

i = 2 00 sin 1 000 t and e= 1 0 , 000 cos 1 000 t .

The discharge th us is alternating I n reality due to


61 . .
,

the unavoidable resistance in the discharge path the alterna ,

tions gra d u ally die o u t that is the discharge is oscillating


, , .

The time of one complete period is given by ,

2 7:

Hence the fre nque ncy,


1 1 000
f =_ -
1 59 cycles per second .

t r

As the circuit in addition


the ind u ctance necessarily to
contains resistance r besides the voltage cons u me d by t he
,

ind u ctance by e q uation voltage is cons u med b v t he


resistance thus ,

6 M (1 17) ,

and the total voltage consumed by resistance r and i nductance


L thus is
,

e = ri +L

D i ff erentiating ( 1 1 8 ) gives ,

dc
z
r
dt d

dt i2

and s u bstit u ting this into e q u ation


,
gives .

dz d i z

1 + CT —

f

CL 2
dt
as the di ff erential eq u ation o f the problem .

This di fferential eq u ation is o f the more general


6 2 The more genera l di ff e rential e q u ation
.

2
d y dy
+-
2 c
dxz
d
88 E NGINE E RING M ATH E M A TI C S .

From

or s u bstituting
, ,

V0 ? a
p,

into the eq u ation becomes ,

b at

Hence , two val u es o f b e x ist ,

b1 = —
c +p and b2 =

and there f ore the di ff erential eq u ation


, , ,

2
d z dz
1}
Z ?
20
2 i:
az 0, ( 1 30 )

is solved by As ; or by Ae or by any combination of ,
b”
, ,

these two sol u tions The most general sol ution is



.
,

2 Aus b c

that is ,

— —
y A
=1 5 ( +A 2€
( c p ) x

—c
n

Mm + AQ 8
-
e l e
px
}

As roots of a q u adratic eq u ation b 1 and b 2 may both b e ,

real q uantities or may be comple x imaginary and in the


, ,

latter case the sol ution ( 1 3 1 ) appears in imaginary form and


, ,

has to be red u ced or modified f or u se so as to eliminate the ,

imaginary appearance by the relations ( 1 06 ) and ,

E X AMPLE 2 .

63 . Assume , in the e x ample in paragrap h 60, the disc harge


circ u it o f the condenser o f C = 2 0 m f capacity to contain .
, ,

b esides the ind u ctance L = 0 05 h the resistance r = 1 2 5 ohms


,
.
, ,
.

The general e q u ati on o f the problem dividing by ,

C L becomes
, ,

di 2

( It
?
r di

Ea CL
+
i
“ L
P OTE N TI AL S E R IE S AN D E X PON E N TI AL F UN C TI ON . 89

This is the eq uation for :

x t, 20 2 5 00 :

I
I

and writing
,

and since
10 and

s= 75 and p
= 7 50 .

The eq u ation o f the c u rrent f rom 13 1 ) then is ,

Thi s eq u ation still contains t w o u nkno w n q uantities the inte ,

grat io n constants AI and A2 w hich are determined by the


,

terminal condition : The val ues o f c u rrent and o f voltage at the


beginning o f the discharge or ,

This req u ires the determination of the eq uation o f the


voltage at the condenser terminals This obvio usly is the voltage
.

cons u med by resistance and ind u ctance and is e x pressed by ,

equation

e = ri +L
90 E NG I NE E R I NG AI A TH E M A TI C S .

di
hence s u b s t i t u t l ng h e re 1 n the val u e o f t and from e qu at lo n
,
7
d

gives
' 1

—L 8
J —
gi
f s
1 t r s t T+ 3
e— r A1 c
2L
+ 42 5 + 11 {
3
A1 : 42 c
2L 2L
T

—3 —
t r s t
A1 8 2L
+ — — Ag € 2L
Q

+ i t r _
__
s
i

and s ubstit uting the n um erical val u es ( 1 3 3 ) and ( 1 3 6 ) into


,

equations ( 1 3 7 ) and gives


— 2 oooz
i = Al e 5 00 1
+ A2 6

_ 20 0 0 1
+ 2 5 A2 €

50 0‘
e 5

At the moment o f the beginning o f the discharge ,


t = 0,
the c u rrent is z ero and the voltage is that is ,

t = 0 ; t = o; e = 1 0 , 000

Su bstit u ting ( 1 40 ) into ( 1 3 9 ) gives ,

0 = A1 + A2 , 1 0 000 = 1 00 A 1 + 2 5 A
.

There f ore the c u rrent and voltage are


, ,

t = 1 33 3 {5 .

-
2 0 00 ‘
33 33 5

The reader is advised to calc u late and plot the n u merical


val ues o f i and e and o f their t w o components f or
, , ,

3
t = o,
-
1, 2, 3, e x 1o sec .
92 E N GINE E RING M A THE M ATI CS .

Since for t = 0 , t = o and e = 1 0 000


,
volts substit u ting
into ( 1 45 ) and
and 1 o ooo = 4o ,
B 1 + 30 3 2 .

There f ore , B 1 = 0 and and by ( 1 45 ) and


,

sin 600 t ;

( cos 600 t + 1 33 . sin 60 0 i ) .

As seen , in this case t he c u rrent i is larger , and c u rrent


and e m f are the prod u ct of an e x ponential term ( grad ually
. . .

decreasing val u e ) and a trigonometric term ( alternating value ) ;


that is they consist of s u ccessive alternations of grad u ally
,

decreasing amplitude Su ch f u nctions are called osci llati ng .

fu nc ti ons Practically all dist u rbances in electric circuits


.

consist o f such oscillating c u rrents and voltages .

600 t = 2 7r gives as the time o f one c omplete period


, ,

sec .
;

and the f req u ency is


1
f T = cycles per se e .

In thi s partic u lar case as the resistance is relatively high , ,

the oscillatio ns die o ut rather rapidly .

The reader is advised to calculate and plot the n umerical


val u es o f t and e and of their e x ponential terms f or every 30

, ,

T T T
degrees that is , for t= 0 2 3 — -
etc for the first two

.
, , , ’ , ,
12 12
peri ods and also to derive the eq u ations and calc u late and plot
, ,

the numerical val u es for the same capacity C = 20 ni f and , ,


.
,

same ind u ctance L = 0 0 5h but for the much lower resistance

“ "
, , ,
.

r = 2 0 ohms .

6 5 Tables o f e
. and e f or 5 decimals and tables of , ,

log a and log for 6 decimals are given at the end of ,

the book and also a table of


,
for 3 decimals For most .

en gineering p u rposes the latter is su fficient ; w here a higher


accuracy is required the 5 decimal table may be u sed and f or
, ,
POTE N TI AL S E RIE S AN D EX PON E N TI AL F UN C TI ON . 93

highest accuracy interpolation by the logarithmic table mav be


employed For instance
.

_
6

From the logarithmic table ,

0 -6

0 -0 8

0 a mt —
9 9 9 7 95 9 ,
i nterpolated
i
,

betw een log


— 0 00 5
added 1 and log 5 -

1 3 5 84 7
6 .

From com m b n logarithmic tables ,

s
X lo

5 .

NO TE mathematics f or the base of the natural log a


. In ,

ri th m s
,
is usu ally chosen the symbol 6 Since .
,

ho w ever in engineering the symbol 6 is u niversally used to


,

represent voltage f or the base o f nat u ral logarithms has been


,

chosen the symbol 5 as the Greek letter c orresp onding to e


, ,

and e is generally used in electrical engineering c alc u lations in


this meaning .
CHAPTE R I I I .

TR I GO NO ME TR I C SE RI E S .

A TRI GONOME TRI C FUNC TI ONS


. .

66 .For the engineer and especially the electrical engineer


, .

a perfect familiarity with the trigonometric f u nctions and


trigonometric formulas is almost as essential as familiarity with
the m u ltiplication table To use trigon ometric methods
.

e fficiently it is not s u fficient to understand trigonometri c


,

form u las en o u gh to be able to loo k the m up when required ,

b ut they m u st be learned by heart and in both directions ; that ,

is an e x pression similar to the left side of a trigonometric for


,

m u la must immediately recall the right side and an e x pression ,

similar to the right side m u st immediately recall the left side


o f the formula .

Trigonometric f u nctions are defined on the circle and on ,

the right triangle .

Le t in the circle Fig 2 8 the direction to the right and


,
.
,

u pward be considered as positive to the le f t and dow n w ard as ,

negative and the angle a be co u nted from the positive hori


,

z o nt al OA counterclockwise as positive
,
cloc kwise as negative , .

The projector s of the angle a divided by the radi u s is , ,

called sin a ; the projecti on c of the angle a divided b y the ,

radi u s is called cos a


,
.

The intercept t on the vertical tangent at t h e origi n A ,

divided by the radi u s is called tan a ; the int ercept ct on the


,

hori zont al tangent at B or 90 d e g behind A divided by the


, , ,

radius is called cot a


,
.

Thus in Fig 2 8
,
.
,

CO S a

tan a = c ot a :
96 E NGI N E E RING M A TH E M A TI CS .

negative : 8 being do w nw ard 0 to w ard the le f t ; but tan ,


a and
cot a are again positive as seen f rom t and ct i n Fi g 3 1
,
. .

FI G 30 . . S e co nd Q u adrant . FI G 3 1
. . T hird Q uadrant .

Inthe fourth quadrant Fig 32 sin a is negative as s is ,


.
, ,

downward but cos a is again positive as c is toward the right ;


, ,

tan a and cot a are both


negative as seen from t and ,

c t in Fig 3 2 . .

I n the fi f th quadrant all


the trigonometric functions
again have the same val u es
as in the first quadrant Fig ,
.

2 8 t hat is 3 60 d e g or 2 a , , , ,

or a m u ltiple thereof can be ,

added to or s u btracte d from ,

the angle a witho u t changing ,

the trigonometric f u nctions ,

Fm 3 2 Fo u rth Q u adra nt
. .
b ut these functions repeat
.

a f ter every 3 60 de g or 2 71 ; ,

that is have 2 7: or 3 60 deg as their period


,
. .

S I G NS O F FU NC TI O NS
TRI GON OM E TRI C S E RI E S . 97

68 . Re l a t i o ns b e tw e e n s in a a nd c o s a . Between sin a and


cos a the relation ,

si n
2
a
+cos 2
a 1,
e x ists ; hence ,

cos a
— sin 2
a .

E quation (4) is one of thos e which is f requ ently u sed in


both directions For instance 1 may be s u bstit uted f or the
.
,

s u m o f the sq u ares o f sin a and cos a while in other cases ,

sin 2 a
+ cos a may be s u bstit u ted f or 1 For instance
' 2
,
.

1 sin 2
a cos 2
a

cos 2
a cos ~
a

Re lat i o ns b e tw e e n S i ne s a nd Ta nge nts .

S in a
tan a
00 8 a

c ot (r

hence
cot a

tan a

As tan and c ot a are f ar less convenient for t rigono m e t ric


a

calc u lations than sin a and cos a and therefore are less fre ,

quently applied in trigonometric calc u lations it is not ne ce s ,

sary to memori z e the trigonometric form u las pertaining to


tan a and cot a b ut where these f u nctions occ u r sin a and
, ,

and cos a are s u bstit uted for them by e q u ations and the
calc u lations carried o ut with the lat ter f u nctions and tan a ,

or cot a re su b s t it u t e d in the final result i f th e latter contains ,

S in a
or 1t s re c 1 pro c al .

cos a

In
electrical engineering tan a or cot a freq u ently appears

as the starting—point o f calc u lation of the phase o f alternating


currents For instance i f a is the phase angle of a vector
.
,
98 E NGINE E RING M A TH E M ATI C S .

quantity tan a is given as the ratio of the vertical component


,

over the hori z ontal component or o f the reactive component ,

over the power component .

I n this case if ,

tan a :

and cos a

cot a
'

c
and cos a =
V0 2
+d
2

The secant f u nctions and versed sine f u nctions are so


,

little used l n engineering that t hey are of interest onl y as


,

c u riosities They are defined by the follow ing e qu ations :


.

1
se e a
cos a

0 0 86 0 a

s in vers a 1— sin '


a,

cos vers a 1— cos a .

69 . Ne gative Angl e s
From the circle diagram o f the
.

trigonometric functions follows as shown i n Fig 33 that when ,


.
,

changing from a positive angle that is co u nterclockwise , ,

rotation to a negative angle that is cloc kwise rotation s t


, , , , , ,

and ct reverse their direction b ut c remains the same ; that is


, ,

S in a
~

— a
) = + cos a ,

tan — a
) — tan a ,

cot — a
) —
cot a
,

cos a th u s is an even f u nction ,


while the three others are
odd f unctions .

1 00 E NGI NE E R I NG MA TH E M A TI C S .

70 . Angl e ( a Adding or s u btracting


to an angle a ,
7: ,

gives the same numerical val u es o f the trigonometric f u nctions

FI G 3 5
. . ctio n of C
Fu n s o m p le m e n FI G . Fu n t c ion of An gl
s es P l us
tary An gl es . or Min u s 7r .

as a as seen in Fig 3 6 b ut the direction o f s and


,
.
,
c is reversed ,

while t and ct remain i n the same direction th u s , ,

Sin ( i = —
a n ) sin a ,

— C OS
l
C OS ( Od i n ) a
'
,

tan ( a = + tan
i n ) a
'
,

cot ( a i n ) + 0 0 1; a .

Fun ctio ns of Angle s + 75 . FI G 3 8


. . Fu n ctions of An gl es Mi nu s 9

1
3: Adding g ,
or 90 d og . to an angle a , inter
changes the functions ,
8 and c, and t and ct, and also reverses
TRIG ON OM E TR I C S E R I E S . 1 01

the direction of th e cosine tangent and cotangent b ut leaves


, , ,

the sine in the same direction since the sine is positive in the
,

second qu adrant as seen in Fig 3 7


,
. .

g
'

S u btracting ,
or 90 deg f rom angle
. a ,
interchanges the
f u nctions 8 and c and t and ct and also
, , ,
reverses the direction ,

e x cept that o f the cosine which remains,


in the sa me dire ction ;
that is o f the same si gn as the cosine
, ,
is positive in the first
and f o u rth qu adrant as seen in Fig 38
,
. . There f ore ,

+ cos a ,

— cot a ,


tan a


cos a ,
'

+ 8 111 a ,

tan a — cot a ,

cot a tan a .

Nu m e ri ca l Valu e s From the circl e diagram Fig


.
,
2 8, etc .
,

f ollows the n u merical val ues :


cot 0 °
= co

cot
cot
—1 cot 1 3 5 ° — 1
( 1 2)
e tc .
1 02 E NG I NE E R I NG M ATH E M ATI C S .

7I . R e lat i o ns b e tw e e n Tw o An g l e s
The f ollo w ing relations .

are developed in te x t books o f trigonometry -

sin ( a + 6) = sin or cos 8 + cos


,
a sin fl ,

sin ( a — = sin cos fl cos sin (1


fl) a
— a ,

cos ( a + 6) = cos a cos fl— sin a sin 5 ,

cos ( a — fl) = cos a cos fl + sin a sin (9 ,

Herefrom f ollows by combining these e quations ( 1 3 ) in


,

pairs :
( ( er —
cos cos cos + cos fl)

a
{ a

sin a sin ( a -
fl) -
cos
sin a ( a + fl) + sin ( at — ml ,

cos a s in fl = l
s in
{

sin ( er — ml .

By substituting for and £1 f or (


d l
a —
fi)
equations gives the e qu ations

sin a 1 + sin ,8 1 = 2 sin

S in d l
— sin 81 =

cos cos = cos


a 1 + fi1 2

cos a l — cos fi1 = — 2 sin

These three sets of equ ations are the most important trigo
no m e t ric form u las Their memori z ing can be f acilitated by
.

noting that cosine f un cti ons lead to prod u cts o f equ al f unc
tions sine f u nctions to prod u cts of u nequal f u nctions and
, ,

inve rs ely prod u cts of equ al f u nctions resolve into cosine


, ,

products of u neq u al f unctions into sine f unctions Also cosine .

functions show a reversal o f the sign th us : the cosine of a ,

su m is given by a di fference o f prod u cts the c osine o f a di ff er ,

ence by a sum for the reason that with increasin g angle


,

the cosine function decre ases and the cosine o f a su m o f angles ,

thus wo u ld be less tha n the cosine of the single angle .


1 04 E NGIN EE RING IlI ATH E rlI A TI C S .

sin a da =0
;

cos a d a = 0;

sin a d a = 1;

cos a do +1 .

73 . B in o m i al . One
of the most frequent trigonometric
operations in electrical engineering is the trans f ormation of the
binomial a cos a + b sin a i nto a single trigonometric function
, , ,

by the substit ution a = c cos p and b = c sin p ; hence


, ,

p) ,
a c os a +b sin a =c cos ( oz —

whe re

or , by the trans f ormation ,


d = c sin q and b=c cos (
1,

a cos a +b sin a = c sin


where

74 . Polypha se Re lat i on s .

11

X
i
1
t cos a + a zt

z 1
i sin ,
a +a i

where m and n are integer n u mbers .

Proof The points on the circle which defines the trigo


.

no m e t ri c f u nction by Fig
,
. 2 8, of the angles + a +
TR I G ON OM E TR I C S E R I E S . 1 05

are corners of a reg ular polygon inscribed in the circle and ,

therefore having the center o f the circle as center of gravit y .

For instance f or n = 7 m = 2 they are sho w n as P I P 2


, , ,
P7 , , ,

in Fig 3 9 The cosines o f these angles are the proj ections on


. .

the vertical the sines the proj ections on the hori z ontal diameter
, , ,

and as the sum of the proj ections of the corners of any polygon ,

FI G 3 9
. . P ypol ha se Re lations . FI G 40
. . T i ng r a le .

on any line going thro u gh its center of gravity is z ero both , ,

sums o f e qu ation ( 2 3 ) are z ero .

a +a+ cos a +bi cos ( a -


t) ,

a + a 3t sin a +b+ cos (a — b ) , ( 2 4)

a +ai cos a +bi sin (a — b ) .

These e quations are proven by s u bstit uting for the produc t s -

the single fun ctions by e qu ations and substit uti ng them


in e q u ations
7 5 T r i a ngle
. I f in a triangle a fl and r are the angles
.
, , ,

opposite respectively to the sides a b c Fig 40 then , , ,


.
, ,
1 06 E NGIN E E RING M ATH E M A TI CS .

2= 2 2 —
0 a + b 2 ab cos 7 .

2
c sin a sin 6
2 S in 7

B . TRI GONOME TRI C SE RI E S .

76 . phenomena usually are either constant


E ngineering ,

transient or periodic Constant for instance is the terminal


,
.
, ,

voltage of a storage battery and the current taken f rom it


-

through a co nstant resistance Trans ient phenomena occur .

during a change in the condition of an electric circuit as a ,

change o f load ; or disturbances entering the circui t from the


,

o u tside or originating in it etc Peri o d ic phenomena are the


,
.

alternating currents and voltages pulsating c u rrents as those ,

produced by rectifiers t he distribution o f the magnetic flu x


,

in the air gap of a machine or the distri b u tion of volta ge


-
,

around the comm u tator o f the direct current machine the -


,

motion of the piston in the steam —engine cylinder the variation ,

o f the mean daily temperature w ith the seasons of the year etc ,
.

The characteristic o f a periodic f unction is that , ,

at cons tant intervals o f the independent variable x call ed ,

cycles or pe riods the same val ues of the dependent variable 3]


,

occur .

Most periodic f unctions of engineering are f unctions of time


or o f space and as such have the charact eristic of univalence ;
,

t hat is to anv val u e o f the independent variable a: can corre


,

spond o nly one value of the dependent variable y I n other .

words at any given time a nd given point of space any physical


, ,

phenomenon c an have one numeric al value only and there f ore ,

mu st be rep resented by a u nivalent f u nction o f t ime and spac e .

Any u ni v al ent p e ri odic f unc tion ,

y
1 08 E N GIN EE RING M ATH E MATI CS .

Theoretically obvi o u sly this condition can never be perfectly


,

attained an d freq u ently the deviation from sine shape is su ffi


,

cient to require practical consideration especially in those cases , ,

where the electric circ uit contains electrostatic capacity as is ,

f or ins tance the case with long distance transmission lines


,
-
,

u ndergro u nd cable systems high potential trans f ormers etc , ,


.

H o w ever no matter how m u ch the alternating or other


,

periodic wave di ffers f rom simple sine shape that is however —


,

much the wave is distorted it can always be represented ,

by the trigonometric series


As ill ustration the follo w ing applications o f the trigo
no m e t ri c series to engineering problems may be considered :

( A) The determination of the eq uation o f the periodic


f u nction ; that is th e evolution of the constants a and b of
, n n

the trigonometric series i f the numerical val u es of the periodic


,

function are given Th u s for instance the w ave of an


.
, ,

alternator may be taken by oscillograph or w ave meter and -


,

by measuring from the oscillogr aph the n u merical val u es o f ,

the periodic f u nction are derived for every 1 0 degrees or every ,

5 degrees or every degree dependi ng on the accuracy required


, , .

The problem then is from the n u merical val ues of the wave, ,

to determine its e qu ation While the oscillograph sho w s the .

shape of the wave it obvio u sly is not possible there f rom to


,

calc ulate other q u antities as f rom the voltage the c u rrent ,

u nder given circ u it conditions i f the wave shape is not first ,

represented by a mathematical e x pression I t therefore is of .

importance in engine eri ng to t ranslate the pict u re or the table


o f n u merical val ues o f a periodic f u nction into a mathematical
e x pression thereof .

( B ) I f one o f the engineering quanti ties as the e m f o f ,


. . .

an alternator or the magnetic fl ux in the air gap of an electric -

machine is given as a general periodi c f u nction in the form


,

of a trigonometric series to determine there f rom other engineer ,

ing quantities as the c u rrent the generated


,
etc , .

A E valu at i o n o f th e C o nstan t s o f t h e Tri go no m e tri c S e r i e s f ro m


.

t h e I nstanta ne o u s Va l u e s o f th e P e ri o di c Fu nc ti o n .

7 8 Ass u ming that the n u merical val u es of a univalent


.

periodic f unction y=fo ( 0) are given ; that is f or every val u e ,

o f d the corresponding val u e o f y is kno w n either by graphical


, ,

re pre s e nt at lo n Fig 41 ; or in tab u lated f orm Table I but


,
.
, , ,
TRIG ON OM E TRI C S E R I E S . 1 09

the equ ation o f the periodic f unction is not kno w n . It can be


represented in the f orm ,

= a +a cos 0 + a2 cos 2 0 + a3 cos 30 + + an cos n0 +


y o 1 .

+ b1 sin 0 + b2 sin 2 0 + b3 sin 30 sin n0 + .


(7)
and the problem no w is to determine the coe fficients
,
d o, or

(1 2 01 , Oz

FI G 4 1
. . Pe riodic Fu n ction s .

T AB L E I .

I nte grate the e q uati on ( 7 ) between the limits 0 and 2 7:

d 0 =o d0 + a 1 0d 0 + cos 2 0d 0 +
g

cos n 0d 0 + . . + b1 sin 0d 0 +

sin 2 0d 0 + . sin n 0d 0 +
1 10 E NGIN E E RING M A TH E M A TI C S .

Allthe integrals containing trigonometric f unctions vanish ,

as the trigonometric f u nction has t he same val ue at the upper


limit 2 7: as at the lo w er limit 0 that is , ,

— co s cos 0 ) = 0 ;
7L
( 2 n7:

( s in 2 77 7: sin 0 ) 0,

and the res ult is

yd fi (1 0

hence

yd 0 is an element of the area o f the c urve y, Fig . 41 , and

yd 0 th u s is the area o f the periodic f u nction y, for one


period ; that is ,

where A = area of the periodic f u nction f or one period ;


that is f rom ,
to
A
2 7: 1s the hori z ontal Wi dth of th 1 s area A and th u s 1s
5: ,

th e area divided by the w idth o f it ; that is it is the average ,

height o f the area A o f the periodic f u nction y; or in other ,

words it is the average val u e o f y Therefore


,
.
,

de = av e ( y) .
.

The first coe fficient ao th u s is the average val u e o f the


, , ,

instant aneo u s val u es o f the periodic f unction y bet w een ,

and 0 = 2 7r .

There f ore averagi ng the val ues of y in Table I gives the


, ,

first constant a o .

7 9 To. determine the coe f


f icient a n m u ltiply e q u ation (7) ,

b v cos n 0 and then integrate f rom 0 to


,
for the p u rpose o f
making the trigonometric f u nctions vanish This gives .
1 12 E NGIN E E RING M A TH E M A TI C S .

the instantaneous val ues of y are m u ltipli ed w ith cos n0


If .

and the prod u ct g = ycos n0 plott ed as a curve ycos n 0d 0 is


u ,

an element o f the area o f t his c u rve sho w n f or n = 3 in Fig 42


,
.
,

and th u s ycos n 0d 0 is the area o f this curve ; that is ,

FI G 42
. . Cu rve of y c o s 30 .

where An is the area of the curve ycos n0, bet w een and
0 = 2m
A
As 2 7: 1s t h e W i dth o f th 1 s area An ) t
) ”
is the average h e l gh t

1
of this area ; that is is the average val ue o f
, y cos 71 0, and :I A
,,
»

th u s is twice the average value o f y cos n0 ; that is ,

z”
ou = 2 avg ( ycos
. n0) o

FI G 4 3
. . Curve of y sin 30 .

The coe fficient a n of cos n0 is derived by m u ltiplying all


the instantane o u s val u es of y by cos n0 and taking t w ice t he ,

average o f the instantaneous val ues o f this prod u ct ycos n 0 .


TR I G ON OM E TR I C S E R I E S . 1 13

80 . is determined in the analogo us manner by m u ltiply


bn
ing y by sin n 0 and integrating f rom 0 to by the area o f the
c u rve y sin 71 0 shown in Fig 43 f or n = 3
,
.
, ,

y sin n 0d 0 a sin n 0d 0 + a sin n0 cos 0d 0

sin n0 cos 2 0d 0 + an sin n0 cos n 0d 0

sin n0 s in 0d0 bz sin n0 sin 2 0d 0

2
i
s n n 0d 0

sin n 0d 0 + 77 1
"m + sin ( n
513 ( 77
.

[sin (n + 2 ) e + sin (
% n — 2 ) 0]d 0

%sin 2 n0d 0

+b (n —
cos

+b } [oos
v
(n —
cos

cos 2 n0]d 0

= b n7r
5
=d 0 ;

hence ,

y sin n 0d 0

where An ’
is the area o f the c u rve y n

ysin n0 . H ence ,
2 7:
bn = 2 avg ( ys1 n n0)
.
o
1 14 E NGINE E RING M ATH E M ATI CS .

and the coe ffic i ent o f sin n 0 th u s is derived by multiplying the


instantaneo u s val u es of y w ith sin n 0 and then averaging as ,
,

twice the average of ysi n n0 .

8 1 Any u nivalent periodic f u nction of w hich the n u merical


.
,

val ues y are kno w n can th u s be e x pressed n u merically by the


e q uation ,

= a + a 1 cos 0 + a cos cos n 0 +


y o 2 .

+ b1 sin 0 + b2 sin sin n0 + .


( 1 7)

where the coe fficients d o, (1 1 , o z, bl, bg are calc ulated


as the averages :

a = av
o s

d l =2 avg (ycos 0)
. bl = 2 avg ( ysin 0) .

a2 = 2 avg ( y cos
. 2 0) b2 = 2 avg ( ysin . 2 0)

an =2 avg (ycos
. n0 ) bn = 2 avg ( ysin . n0)

Hereby any individual h armonic can be calc u lated , witho u t


calc ulating the preceding harmonics .

For instance let the generator e m f w ave Fig 44 Table


, . . .
,
.
,

I I col u mn 2 be impressed u pon an underground cable system


, ,

Ge ne rator e m . . f w . av e .

of s u ch constants (capacity and inductance ) that the natura l ,

f req u ency o f the system is 6 7 0 cycles per second w hile the ,

generator f requency is 60 cycles The natural frequency of the .


1 16 E NGINE E R I NG M A TH E M ATI CS .

hence its e ff ective val u e is


.

while the e ffective val u e of the total generator wave that ,

is the sq uare root o f the mean sq uares of the inst a nt a


,

neo us val u es y is ,

e= 30 5 ,
.

thus the 1 l t h harmonic is per cent o f the total voltage ,

and w hether su ch a harmonic is sa f e or not can now be deter ,

mined f rom the circuit constants m ore partic u larl y its resist ,

a nce .

8 2 I n general the successive harmonics decrease ; that is


.
, ,

with increasing n the val u es o f a n and b become smaller and


, n ,

when calculatin g a n and b n by eq u ation f or higher val ues


of 11 they are deri ved as the small averages of a num b er of
large q u antities an d the calc ulation then becomes incon
,

ve nie nt and less correct .

Where the entire series of coe fficients a n and b is to be n

calculated it th u s is preferable not to use the comp lete peri odic


,

function y b ut only the resid u al le ft a fter s u btracti ng the


,

harmonics which have alre ad v been calc u lated ; that is after ,

a o has been calc u late d it 1 s s u btracted f rom y and t h e di ffer


, ,

e nce 1111 = y— a o is u sed for the calc u lation of a l and b l


, ,
.

Then 0 1 cos 0 + b 1 sin 0 is s u btracted f rom y] and the ,

di fference ,

y a
=
y i

( a i cos 0 + b 1 sin 0)

y (a 0 + a 1 cos 0 + b1 sin

is u sed for the c alculation of G 2 and b 2 .

Then G 2 cos 2 0 + b 2 sin 2 0 is s u btracted from yz and the rest , ,

y g u,
sed f or the calc u lation of and b 3 etc ,
.

I n this manner a higher acc u racy is derived and the cale u ,

lation simplified by having t h e instantaneous val u es of the


f unction o f the same magnitude as the coe fficients a n and
As ill ustration is given in Table I I I the calc u lation o f t he
,

first three harmonics o f t h e pulsating current Fig 4 1 Table I : , .


,
TR I G ONOM E TR I C S E RI E S . 1 17

83 . In
electrical engineering the most important periodic ,

functions are the alternating c u rrents and voltages D ue to .

the constru ctive feat u res o f alternating current generators -


,

alternating voltages and c u rrents are almost al w ays sym m e t


rica l w aves ; that is the periodic f u nction consists o f alternate
,

half w aves which are the same in shape b ut opposite in dirce


-
, ,

tion or in other w ords the instantaneous val u es f rom 1 80 deg


, , .

to 3 60 deg are the same numerically b ut opposite in sign


.
, ,

f rom the instantaneo us val ues betw een 0 to 1 80 d e g and each ,

cycle or period th u s consists of two eq ual b ut opposite half


cycles as shown in Fig 44 I n the earlier days o f electrical
,
. .

en gineering the frequ ency has f or this reason frequently been


,

e x pressed by the n u mber of hal f w aves or alternations -


.

I n a svm m e trical wave those harmonics which prod uce a ,

di fference in the shape of the positive and the negative hal f


wave cannot e x ist ; that is their coe fficients a and b m u st be
, ,

z ero . Only those harmonics can e xist in which an increase o f


the angle 0 b v 1 80 d e g o r reverses the sign o f the f u nction
, .

This is the case with cos n0 and sin n0 i f n is an odd number , .

If h o w ever n is an even n umber an increase o f 0 by 7: increases


, , ,

the angle n0 by 2 7: or a m u ltiple thereo f th us leaves cos n0 ,

and sin n0 with the same sign The same applies to ( t o There . .

f ore symmetrical alternating w aves comprise o nlv the odd


,

harmonics b ut do not contain even harmonics or a constant


,

term and th u s are represented by


,

= al cos 0 + a3 cos 3 0 + a 5 cos


y
~

+ b1 sin 0 + b3 sin 3 0 + b 5 sin

When calc u lating the coe fficients a and b of a symmetrical ,, 1,

wave b y the e xpression it is s u fficient to average from 0


to 7: that is over one hal f wave only I n the second hal f w ave
,
-
.
-
,

cos n0 and sin n0 have the opposite sign as in the first hal f—w ave ,

i f n is an odd n u mber and since y also has the opposite sign


,

in the second hal f wave yco s n0 and ysin n0 in the second


-
,

hal f wave traverses again the same val u es with the same sign
-
, ,

as in the first hal f w ave and their average th u s is given by


-
,

averaging over one hal f w ave only -


.

There f ore a symmetrical u nivalent periodic f unction as an


,
1 18 E NG I N E E RING MA TH E MATI C S .

T AB L E

y1
= _
y 00 11 , CO "
"0 11 1 5 111 0
Ci =0 1 00 8 0
0
=
Uz ll n
“ ci

— 111

D ivi de d
1 20 E NGI NE E R I NG M A T H E M A TI CS .

alternating voltage and c u rrent u u s ally is can be represent e d


,

by the e x pression ,

= a cos 0 + a3 cos 3 0 + a5 cos 5 0 + a7 cos


y r

+ b1 S lll 0 + bg S in 3 0 + O5 S in 5 0 + b 7 sin 7 0 ‘

l

a1 =2 avg ( y cos 0)
. bl = 2 a .
(ysin
a3 =2 avg ( y cos
. 3 0) b3 =
avg ( y sin 3 0)
2 .

a5 =2 avg ( y cos
. 5 0) b 5 = 2 avg ( y sin 5 0)
.

a7 = 2 avg (y cos
. 7 0) b7 = 2 avg ( y sin
.

84 .From 1 80 deg to 3 60 d e g the even harmonics have


.
.

the same b ut the odd harmonics the opposite si gn as from 0


,

to 1 80 deg Therefore adding the numerical values in the


.

range from 1 80 deg to 3 60 deg to those in the range f rom 0


. .

to 1 80 d e g the odd harmonics cancel and only the eve n har


, ,

m oni cs remain I nversel y by subtracting the even harmo ni cs


.
, ,

cancel and the odd ones remain


,
.

Hereby the odd and the even harmonics can be separated .

I f y= y( 0) are the numerical val u es of a periodic function



from 0 to 1 80 d e g and y = y( 0 the numerical values of
,

the same function f rom 1 80 deg to 3 60 d e g .


,

atW ) + 9 09 + 7r) (22 )


is a periodic f u nction containing only the even harmonics and ,

9 1 0? ( 23 )
is a periodic f unction containing only the odd harmonics ; that i s

y1 ( )
0 = a1 cos 0 + a3 cos 30 + as cos 5 0 +
+ 51 sin 0 + b3 sin 3 0 + b5 sin 5 0 +

yw
= ao + cos 2 0 + a4 cos
e) 61 2

+ 62 sin 2 0 + b4 sin
and the complete f u nction is
M) =
y (
i 0> + y )
z ( 0
TR I G ON OM E TR I C S E R I E S . 12 1

By this method it is convenient to determine w hether even


harmonics are present and if they are present to separate
, ,

them f rom the odd harmonics .

B e f ore separating the even harmonics and the odd har


m oni cs it is us u ally convenient to separate the constant term
,

a o from the periodic f unction y by averaging the instantaneo u s ,

val u es o f y f rom 0 to 3 60 deg The average then gives a o .


,

and subtracted from the instantaneo u s val ues o f y gives ,


fi (2 7 )
yof ) y( )
= 9 Go

as the instantaneous valu es o f the alternating component o f the


periodic f unction ; that is the componen t yo contains o nlv the ,

trigonometric f u nctions b ut not the constant te rm yo is ,


.

t hen resolved into the odd series yl and the e ven series yg ,
.

8 5 The alternating w ave yo consists o f the cosine components


.

u ( )
0 = a1 cos 0 + a2 cos 2 0 + a3 cos 3 0 + a4 cos ( 2 8)

and the sine components :


= bl sin 0 + bz sin 2 0 4- O3 S 111 3 0 + 04 sin 40 + (2 9)
that is ,

0 0 ( )
0 11 0 (30 ) =u

The cosine f u nctions retain the same sign f or negative


a ngles as for positive angles ( w hile the sine f u nctions
reverse their sign ; that is ,

( u 0) + u ( 0) (3 1 )
and v( 0) -
v( 0)

There fore i f the val u es o f yo for positive and f or negative


,

angles 0 are averaged the sine f u nctions cancel and only the
, ,

cosine f u nctions remain while by s u btracting the val ues of ,

yo f or positive and f or negative angles only the sine f u nctions ,

remain ; that is ,

— =
yo fi) + 3/o f
f 0 ) 2 u f9) ;

y ( 0) y(o 0) o

hence the cosine terms and the sine terms can be separated
,

f rom each other by combining the instantaneo us val u es of yo


for positive angle 0 and f or negative angle th us :
u ( 0) + 90 (

— — 0
yo ( ) f
1 22 E NG I N E E RING M A TH E MA TI C S .

Us u ally be f ore separating the cosine and the sine terms


, ,

a and first the constant term 0 0 is separated as disc u ssed ,

above ; that is the alternating f uncti o n y0 y ao u sed I f


,
= —
.

the general periodic f u nction y is used in e q u ation t he


constant term a o of this periodic f u nction appears in the cosine
term a th us : ,

u ( 0) + y( = a + a1
o cos 0 + a2 cos 2 0 + a s cos
w hile v( 0) remains the same as when using yo .

8 6 B e f ore separating the alternating f u nction yo into the


.

cosine f u nction u and the sine f u nction it us u ally is more


convenient to resolve the alternating f unction yo i nto the odd
series yl and the even series 312 as disc u ssed i n the preceding
, ,

paragraph and then to separate yl and yg each into the cosi ne


,

and the sine terms :

y1 ( ) y1 (
0 — sin 3 0+ b 5 sin 5 0+

+ y2 ( = a
2 cos 2 0 + a4 cos

v3 0)
( —
y2 (
= b2 sin 20 + b4 sin

the odd f u ncti ons u l and 171 a change from the negative
In ,

angle 0) to the s u pplementary angle ( 7: 0) changes the angle


o f the trigonometric f u ncti on by an odd m u ltiple of 7: or 1 80
d e g that is by a m u ltiple o f 2 7: or 3 6 0 d e g plus 1 80 de g
, , , ,

w hich signifies a reversal o f the f u nction th us


°

109 i ll/d w yd —


fi) l
,

v1 ( 0) — mi
How ever , in the even functions and 172 a change from the uz

negative angle 0) t o the s u pplementary angle (7: chan g es


the angles o f the trigo nometric f unction by an even m ultiple
of that is by a m u ltiple o f 27: or 3 60 d e g ; hence leaves
,

the sign o f the trigonometric f u nction u nchanged th u s : ,


1 24 E NGINE E R I NG IlI ATH E M A TI CS .

analogously 11 4 is the cosine term o f,


t he even f uncti on a f ter
,

subtracting the term cos 2 0

u4 = u2— ag cos 2 0,

and in the same manner ,

u5 = u3 — a 3 C OS 3 0,
ue = u4

a 4 cos 4 0,

and so forth v1 v2 vs 17 4 etc are the corresponding sine


, , , ,
.
,

terms .

When calc u lating the coe fficients a and b by averagin g over n n

90 d e g or ove r 1 80 deg or 3 60 d e g it m u st be kept in mind


, .
,

that the terminal val u es o f y respectively o f u or that is ,

the val ues for and deg ( or deg or 3 60 . .

deg respecti vely ) are to be taken as one —half only since t hey
.
,

"
are the ends o f the meas u red area o f the c u rves a cos n0 and n

"
b sin n 0 w h ich area gives as t w ice its average height the val u es
,
.

a n and b as disc u ssed in the preceding


,
.

I n resolvin g an empirical periodic f u ncti on into a t rigono


metric series j ust as i n most engineering calc u lations the
, ,

most impo rtant part is to arrange the work so as to derive the


res u lts e x peditio u sly and rapidly and at the same time ,

acc u ratel y By proceedi ng f or instance imm e d iat e lv by the


.
, ,

general metho d eq u ations ( 1 7 ) and ,


the w ork become s so
e x tensive as to be a serio u s was te of time w hile b v the system ,

atic resol utio n into simpler f u nctions the work can be gre at lv
red u ced .

88 I n resolving a general periodic function y( 0) into a


.

trigonometric series the most convenient arrangement is :


,

1 To separate the constant term a o by averaging all the


.
,

instantaneo u s val ues o f y( 0) from 0 to 3 60 deg ( co u nting t h e .

end val u es at and at deg one hal f as discussed .


,

above )
a
o
avg .

and then subtracting a f rom y( ,


0 ) gives the alternating f u nc


o
tion ,

ye w) W?) 0 °
T RI GON OM E TRI C S E R I E S . 1 25

2 .resolve the general alternating f unction


To yo )
f0 int o
the odd f unction y1 ( 0) and the even f unction y2 ( ,

(2 3 )
9M ) (22 )
3 . resolve To y1 ( 0) gnd y2 ( 0) ) into the cosine terms u and
the sine terms 17 ,

9) l
'

y1 (

u2 ( 0) + yz (

vz lfi) i f9 2 W) —
yz ( 9) l
4 . To
calc u late the constants al, (lg, a3 . b l , Oz, b 3 .

by the averages ,

an 2 avg (u n cos n 0)
.

n
=2 avg . sin n 0) o

Ifthe periodic f u nction is known to c ontain no even har


monies that is is a symmetrical alternating wave steps 1 and
, , ,

2 are omi tted .

M arc h

FI G 4 5 . . Me a n D ail y Te m pe rat u re at S che ne ctad y .

89 . As ill ustration o f the resol ution of a general periodic

"
wave mav be sho w n the resol ution of the observed mean daily
temperat u res o f Schenectady thro u gho u t the year as shown ,

in fi g 45 up to the 7 th harmonic
.
,
.

1
Th e rical value s of t e m pe rature c annot clai m any gre at ab sol u t e
nu m e

a cc ur ac y as th y a e ave rage d ove r a r lative l y m ll nu m b e r of ye ars o nl y


,
e r e s a ,

and ob rve d b y in tr m e nts of onl y m ode rate accuracy For t he p u r p ose


se s u .

of illu trati ng t he re solution of th e e m p irical curve i nto a tri go no m e tri c


s

se ri e s thi s i s no t e sse nti al ho we ve r


, , .
1 26 E NGI NE E RING M ATH E M A TI CS .

T A B LE I V

Sep t . 2 40

D ivid e d b y 3 6
128 E NGIN E E RING M ATH E M A TI C S .

a m m w
m S m m m
m o o o m
d d
. .

c c d
l l + + r
r

« w m m
8
8

n 8 m a
8
h
2 8 8 8 o ~
.
m
d d
.

8 8 8 o 8 8 o
1 1 1 + 1 + + + l

m m .

la w fi ‘ h m
m

Uf ls 0C
TRI GONO ME TRI C S E RI E S . 1 29
E NGINE E RING M ATH E M A TI C S .

T A B LE V I II .

C O SI NE SE R I E S u, .

(3) (4) (5 ) (6) (7 ) (8 ) a s c os 60


na c o s 20 a2 co s 20 m cos 40 (1 4 co s 40 14 6

o + 9 15 + 9 31
19 — 9 10 —9 10
29 — 9 03

30
40 + 9 30 + 9 05 + 9 30 — 9 29
— 9 21
50 + 9 22 + 9 22

60 + 9 92 + 9 02
79 + 9 05 — 9 04 + 9 05 + 9 91
80 — 9 10 — 9 08
+ 9 09
90

— 9 002

T AB L E I X .

S I NE S E RI E S

(3 ) (4) (6 ) (7)
172 s i n 20 b e s in 2 0 0 4 sm 40 bl s in 40

10 + 9 20 + 9 07
20

+ O 30 .

-
9 37

— o s4 .

-
9 95

T otal
D ivid d
e by 9
D ivid e d by 2
1 32 E NGINE E RING M A TH E M A TI CS .

Since to an angle any m u ltiple o f 27 : or 3 60 deg may .

3 60
be added any m u ltiple of
,
n
may be added to the angle
2 7

and thus the angle


F
i
fi may be made positive etc ,

91 . The eq u ation o f the temperat u re wave th u s becomes


= cos ( 0 cos
y 8 75 .

cos cos
cos cos
cos

or transformed to sine f u nctions by the substitution


, ,

cos w = — sin
=8 7 5 + 1 3 69 sin ( 0 + 0 6 02 sin
y .

+ 0 35 9 sin sin
sin sin
sin

The cosine form is more c onve m e nt for some purposes ,

the sine form for other purposes .

Substit u tin g fi= 0 or these tw o


,

e q uations ( 0 ) and (b) can be trans fo rme d into the form ,

y c os 0— 0 6 2 cos 2 ( 0 ,
cos 3 ( 0
cos cos
cos c os
and

y sin a+ 9 002 sin 2 01 + + 9 359 sin


+ 9 41 0 sin + 9 91 sin
+ 9 1 84 sin + 0 0 85 sin (d )
The periodic variation of the temperat u re y as e x pressed ,

by these eq uations is a res u lt of the periodic variation of the


,

thermomotive force ; that is t he solar radiation This latte r


, .
TRI GONOM E TRI C S E R I E S . 1 33

is a minim u m on De c 22 d that is 9 ti m e degrees be f ore the


.
, ,
-

z ero o f 0 hence may be e x pressed appro x imately by


c = c —h cos
or s u bstit u ting 0 respectively 0 for 0:

z = c — h cos


e + h sin

This means : t he ma x im um of y occ u rs deg after t he .

nza x im u m of z ; in other words the temperat u re lags de g , ,

or abou t +3 peri od behind t he thermomotive f orce


1
, .

Near all the sine f u nctions in (d) are increasing ; that


is the temperat u re wave rises steeply in spring
,
.

Near d e g the sine f u nctions of the odd angles are


,

decreasing o f the even angles increasing and the decrease of


, ,

the temperature wave in f all thus is smaller than the increase


in spring .

The fu ndamental wave gr eatly preponderates with ampli ,

tude c l = 1 3 69 . .

I n spring f or 5 = ,
de g all the higher harmonics ,

rise in the same direction and give the su m ,


or
per cent o f the f undamental I n fall f or 0 = the
.
,

even harmonics decrease the odd harmonics increas e the ,

ste epn e ss and give the sum


,
or per cent .

There f ore in spring the temperat u re rises


, ,
per cent
faster and in aut u mn it falls
,
per cent slo w er than c orre
s po nds to a sine wave and the di ff erence in the rate o f tempera
,

ture rise in spring and temperat u re fall in a ut u mn th us is


,

per cent .

The ma x im u m rate of temperat u re rise is 90


deg behind the temperat u re minim u m and
.
,

deg behind the minim u m o f the thermomotive force


. .

As most periodic f u nctions met by t h e electrical engineer


are symmetrical alte rn ating f u nctio ns that is contain only , ,

the odd harmonics in ge neral the w ork o f resol ution into a


,

trigonometric series is very m u ch less t han in above e x ample .

Where su ch red u ction h as to be carried o ut freq u ently it is ,

advisable t o memori ze the trigonometric f u nctions f rom 1 0 ,

to 1 0 d e g u p to 3 decimals : that is w ithin the acc u racy o f


, ,

the slide rule as t h e re b v the necessity o f looking u p tables i s


,
1 34 E NGIN E E R I NG M A TH E M ATI C S .

eliminated and the work therefore done m u ch more e x pe


d it io u sly I n general the slide r u le can be used for the calc u la
.
,

tions .

As an e x ample of the simpler red u ction of a symmetrical


alternating wave the reader may resolve into its harmonics
, ,

u p to the 7 t h the e x citing c u rrent o f the trans f ormer of which


, ,

t he numerical val ues are given from 1 0 to 1 0 deg in Table X ,


. .

C . RE D UCTI ON OF TRI GONOME TRI C SE RIE S BY POLY


PHASE RELATI ON .

92 . In
some cases the reduction of a general periodic func
tion as a comple x wave into harmonics can be carried o ut
, ,

in a m u ch q u icker manner by the use of the polyphase equation ,

C hapter I I I Part A E specially is this true if the com


,

ple t e e q u at ion of the trigonometric series which represents the ,

periodic f u nction is not req u ired but the e x istence and the
, ,

amo u nt o f certain harmonics are to be determined as for ,

instance whether the periodic f u nctio n co ntain even harmonics


or third harmonics and how large they may be ,
.

This method does not give the coe fficients a n b o f the . n

individ u al harmon ics b ut derives from the numerical val u es ,

of the general wave the n u merical val ues of any desired


harmonic This h armonic ho w ever is given toget h er with all
.
, ,

its m u ltiples ; that is when separating the third harmonic , ,

in it appears als o the 6th 9t h 1 2t h etc , , ,


.

I n separating the even harmonics yg f rom the general


wave y in paragraph 84 by taking the average of the values
, ,

o f y f or angle 0 and the val u es of y for angles ( 0


,
this
method has already been u sed .

Ass u me that to an angle 0 there is successively added a


constant quantity a thus : 0 ; 0 + a ; 0 + 2 a ; 0 + 3a ; 0 + 4a
, ,

etc u ntil the same angle 0 pl us a multiple of 2 7 is reached ;


.
,
:

that is ,
a or in other w ords
, ,
71

l /n of a multiple of Then the su m of the cosine as


as the sine f unctions o f all these angles is z ero :
cos 0 + cos ( 0 + 2 a ) + cos ( 0 + 3 a )
+
ce s ( 0 + [n
1 36 E NGINEE RING M A TH E M ATI C S .

gives the t hird harmonic together with its m u ltiples the 6t h


, , ,

9 t h etc
,
.

This method does not give the mathematical e x pression


o f the harmonics b ut their n u merical val u es Thus if the
, .
,

mathematical e x pressio ns are req u ired each of the component ,

harmonics has to be red u ced from its n u merical values to


the mathe m ati cal e qu ation and the method then u s u ally off ers
,

no advanta g e .

I t is especially s u itable ho w ever w here certain classes of


, ,

harmoni cs are desired as the third together w ith its multiples


,
.

I n this case from the n u merical val u es t h e e ffective val u e ,

that is the e qu ivalent sine wave may be calc u lated


,
.

9 3 As ill u stration may be investigated the separation o f


.

the third harmonics f rom the e x citing c u rrent of a transforme r .

TA B L E X

In table X A are given in columns 1 3 5 t h e angles 0


, , , , , ,

from 1 0 deg to 1 0 d e g and in col u mns 2 4 6 the correspond


.
, , , ,

ing val u es o f the e x citi ng c u rrent i as derived by calc u lation


,

from the hysteresis cycle o f the iron or by meas u ring f rom the .
TR I GON OAI E TR I C S E R I E S . 1 37

photographi c film of the oscillograph Col u mn 7 then gives .

one third the su m o f col u mns 2 4 and 6 that is the third har
-
, , , ,

monic with its overtones ,

To find the 9t h harmonic and its overtones i g the same ,

method is now applied to i 3 for angle 3 0 This is recorded


,
.

in Table X B .

I n Fig 46 are plotted the total e x citing current i its third


.
,

harmonic i 3 and the 9th harmonic i g


,
.

This method has the advantage of showing the limitation


of the e x a ctness of the res u lts resulting f rom the limited num

ber of n u merical valu e s of i on which the calc u lation is based


, .

Th us in the e x ample Table X in w hich the valu es o f i are


, , ,

given for every 1 0 d e g val ues o f the third harmonic are derived
,

for every 30 d e g and f or the 9 th harmonic for every 90 de g ;


,

that is for the latter only t w o points per hal f wave are deter
, ,

minable from the n u merical data and as the two points pe r hal f,

wave are j u st s u fficient to locate a sine wave it f ollows that ,

within the accuracy o f the given n u merical val u es o f i the ,

9t h harmonic is a sine wave or in other w ords to determine


, ,

wheth er sti ll higher harmonics than the 9 th e x ist requ ires f or ,

i more n u merical val u es than f or every 1 0 deg .

As f urther practice the reader may separate from the gen


,
1 38 E N GINE E RING M ATH E MATI C S .

eral wave o f c u rrent ,


io in Table XI, the even harmoni cs i z,
by above method ,

i a= i 0 ( 0) + i o ( 0 + 1 80 deg ) .

and als o the su m o f the odd harmonics as the resid u e , ,

ii Co

then from the odd har moni c s il may be separa ted the third
harmonic and its m u ltiples ,

and i n the same manner from i 3 may be separ ated its third
harmoni c ; that is i g ,
.

Fu rthermore in the su m o f even harmoni c s f rom i z may


, ,

again b e separated its sec ond harmoni c i 4 and its m u ltiples , , ,

and there f rom i s and its third harmo nic i s and its multiples
, , , , ,

th u s gi ving all the harmonics u p to the 9 th with the e x ception ,

o f the 5th and the 7 t h These latter two would req u ire plotting
.

the curve and taking numerical val u es at di ff erent intervals ,

so as to have a n u mber of n u merical val u es divisible by 5 or 7 .

I t is f u rther recommended to resolve this u ns ymmetrical


e x citing c u rrent o f Table X I i nto the trigonometric series by
calc u lating the coe ffi cients a and b u p to the 7 th in the man
n n, ,

ner discussed i n paragraphs 6 to 8 .

T AB LE X I
1 40 E NG I NE E R I NG M A T HE M ATI C S .

value and the ma x im u m val ue of the f u ndamental voltage


,

wave thus is : volts or 3 6 kv ; that is ,


.
,

and
e = 3 6 sin 9 — 9 1 2 sin — sin ( sea
§ ( se 9 23
+ 9 13 sin ( 7 7) (2 )
would be the voltage s u pplied to the transmission line at the
high potential term i nals of the step —u p trans f ormers .

From the wire tables the resist ance per mile of No 0 B


,
S -
. . .

copper line wire is ro ohm .

The inductance per mile o f wire is given by the f orm u la


ls
LO log 0 0805m h ,
7;
.

where l is the distance bet w een the wires and l the radius o f
, , ,

the wire .

I n the present case this gives 1 5 ft 60 in


,
l = 0 1 62 5 in 8 . .
,
.

m h and herefrom it follows that the reactance at


.
, , ,

f 60 cycles is
= 2 7: L 0 ohms per mile
$0 f .

The c a pa c it v per mile o f w ire is given by the f orm ula

hence in the present case C0


, ,
ru f .
,
and the c o nd e ns ive

reactance is derived h erefrom as :

1 6 6 000 ohms ;

60 miles o f line then give the c o nd e ns i ve reactance ,

2 7 70 ohms ;

30 miles or hal f the line ( f rom the generating station to the


,

middle o f the line w here the line capacity is represented by a


,

sh u nted condenser ) give : the resistance r = 30 r0 = ohms ; ,


TRI GONOM E TRI C S E RIE S . 1 41

the inductive reactance x = 30 xo = 22 5 ohms and the e quiva


,
.
,

lent circ uit o f the line now consists o f the resistance r ind u cti ve ,

reactance x and c o nd e nsive reactan ce w in series wi t h each e,

other in the circ u it o f the s u pply voltage 6 .

05 I f.i = c u rrent in the line ( charging c u rrent ) the voltage


cons umed by the line resistance r is r i .

The voltage consumed by the ind uctive reactance a: is a


:

the voltage co ns u med by the con d e ns i ve reactance at , is r e i d 0,

and therefore ,

di
e =x— + ri + rc i d0 .

d0

Di fferentiating this eq uation f or ,


t he p u rpose o f eliminating
the integral gives ,

de
( 10
“S
d i
EE
2

H
di


fl fl »

i di
+2 7 7 9
F
.

The voltage 6 is given by w hich eq uation by resolving ,

the trigonometric f unct i ons g 1ve s ,

e = 36 sin 0 sin 30 sin 50 sin .


70
cos 30 cos 50 cos 70 ; (9)

hence di fferentiating
, ,

36 cos 0 cos 30 cos 50 cos 70

sin 30 sin sin 70 .


( 1 0)

Ass u ming now for the c u rrent i a t 1 igo no m e t ric series with
indeterminate coe fficie nt-s ,

i = al cos 0 + a3 cos 3 0 + a 5 cos 5 0 + a 7 cos 7 0


b l sin 0 + b 3 sin 3 0 + b 5 sin 5 0 + b 7 sin 7 0,
1 42 E NGINE E RIN G M ATH E M ATI CS .

substit u tion o f ( 1 0 ) and ( 1 1 ) into e q uation ( 8) must give an


identity f rom which e q u ations for the determination o f a and
, n

b n are derived ; that is since the prod u ct o f s u bstit u tio n m u st


,

be an identity all the factors of cos 0 sin 0 cos 3 0 s in 3 0


, , , , ,

etc must vanish and this g1ve s the eight e q u ations :


.
, ,


36 2 7 70 a 1 + 1 5 6b l
. 22 5 a 1 ;
.

0 2 7 70b 1 22 5 b i ;
.

46 8 b 3
. 2 02 5 a 3 ;
.

0 5 4 = 2 7 7 0b 3
.

= 2 7 7 Oa + 78b 5 5 62 5 a 5
5 .

= 2 7 7 0b 7 8a 5 5 6 2 5 b5 ;
5 .

= 2 770 0 + 1 09 2 b — 1 1 0 2 5 a
7 7 .
7; .

= 2 7 70b 7 — 1 1 02 5 07 .

Resolved , these equations give

I
I

I
I

i = 13 12 . 00s 0 - 5 03 . cos 3 0
sin 0 sin 30 sin 50 sin 70

cos ( 0 cos ( 3 0
cos ( 5 0 cos ( 7 0
1 44 E NGIN E E RIN G M ATH E M ATI C S .

kv .amp to 1 62 0 kv amp or
-
. times and while with a sine
.
-
.
,

wave o f voltage one of the 7 5 0 kw generators w ould easily be


,
-
.

able to supply the charging current o f the line due to the ,

z =1 5 0 -j ( 22 5n
. .

wave shape distortion more than two generators are required


,
.

I t would there f ore not be e conomical t o u se these generators


, ,

on the transmission line if they can be used for any other


,

purposes as short—distance distrib u tion


,
.

InFigs 47 and 48 are plotted the voltage wave and t h


.
e

c u rrent w ave f rom e qu ations ( 9 ) and ( 14) respectively and


, ,
TRIG ON OM E TRI C S E R I E S . 1 45

the n u merical val ues from 1 0 deg to 1 0 d e g recorded in


,
.
,

Table X I I .

I n Figs 47 and 48 the fundamental sine wave of voltage


.

and c u rrent are also shown As seen the distortion of c u rrent .


,

is enormo us and the higher harmonics predominate over the


,

f undamental Su ch waves are occasionally observed as charg


.

ing currents o f transmission lines or cable systems .

9 7 Assuming now that a reactive coil is inse rted in series


.

with the tra nsmission line betw een the step —u p trans f ormers ,

and the line what w i ll be the voltage at the terminals o f this


,

reactive coil with the distorted wave of charging current


,

traversing the reactive coil and how does it compare with the ,

voltage e x isting with a sine wave of charging c u rrent ?


L et L = ind uct ance th u s sc = 2 7rfL = re act ance o f the coil
, ,

and neglecting its resistance the voltage at the terminals of ,

the reactive coil is given by

Substituting herein the e qu at ion o f c u rrent ,


gives
'=
e x al{ sin 0 + 30 3 sin 3 0 + 50 5 sin 5 0 + 70 ; sm 7 0

—b cos 0 — 3b3 cos 3 0 — 5 b 5 cos 5 0 — 7 b 7 cos


1

hence s u bstit u ting the n u merical val u es


,

=x 70 l

e { sin 0 sin 30 sin 5 0 + 1 35 1 sin
cos 0 cos 3 0 cos 5 0 cos 7 0 }
sin ( 0— 9 3) 0
sin ( se
sin ( 5 0 + 1 39 1 sin ( 7 0
This voltage gives the e ffective val u e

while the e ffective val u e with a sine wave w o u ld b e f ro m

hence the voltage across the reactance x has been increased


,

times by the wave distortion .


1 46 E NG I N E E R I NG MA TH E AI A TI C S .

'
The i ns tant aneous val ues of the voltage e are given in t h e
last column of Table X I I and plotted in Fig 49 for
,
.
,

As seen f rom Fig 49 the fu ndamental wave has practically


.
,

vanished and the voltage wave is the seventh harmonic modi


, ,

fie d by the fifth harmonic .

TA B L E X I I

0 -
0 -1 0 17 2 7 -4 1 — 200
10 — 1 00
+ 2 23 40 3 1 -7 7 + 29 1 9
20 3 + 1 19

30 + 131 + 1 82
40 1 7 -3 5 + 3 9 39 — 1 10 — 3 8 -7 9 93
50 30 90

+ 1 07 — 1 38
+ 1 59 31
3 2 -87 54 54

4 -1 5 — 2 00 1 80 17
1 48 E NGIN E E RING M A THE M ATI CS .

a: and y that is t h e f unction


, ,
is un known or if the ,

f unction is so comp licated as to make the mathematical ,

calculation o f the e xtrema impracticable As e x amples of .

this method the following may be chosen :

FI G 5 1
. . Ma gne ti z ation Cu rve .

E x am pl e I . D etermine
that magnetic de nsity B at which ,

the permeability 71 of a sample of iron is a ma x imum The .

re lation between ma gnetic field inte nsity H magnetic dens ity ,

B and permeability 77 cannot be e x pressed in a mathematical


e quation and is therefore us u ally given in the form of an
,

FI G 52
. . r
Pe m e ability C rve u .

empiri cal c u rve relating B and H as sho w n in Fig 5 1 From


, , . .

this c u rve corresponding valu es o f B and H are taken and their


, ,

ratio that is the perm eability 71


, ,
g , plotted against B as abscissa .

This is done in Fig . 52 . Fig . 52 then shows that a ma x imum


M AX I M A AND M I NIM A . 1 49

occurs at point pm for B = “


kilo li ne s p = 1 3 40 and m i n 1ma
,

at the starting point P 2 f or B = 0 p = 370 and als o for B = oo


-
, ,

, , ,

where by e x trapo lation n= 1 .

E xam pl e 2 Find that out put o f an ind u ction motor


.

which gives the highest po wer f actor While theoretically -


.

an e q uation can be f o u nd relating o utp ut and po w er f actor -

of an ind uction motor the e quation is too complicated f or u se


,
.

The most convenient way of calc ulating ind u ction motors is


to calculate in tab u lar f orm f or di fferent values of S lip 8 the ,

torque o u tput current pow er and volt ampere input e fficiency


, , ,
-
, ,

power f actor etc as is e x plained in


-
, .
,
Theoretical E lements
of Electrical E ngineering third edition p 3 63 From this ,

,
. .

FI G 53
. . P weo r factor Max i m um of I nd uction Motor
- .

table corre s ponding val ues o f p ow er o u tput P and power


factor cos 0 are taken and plotted in a c u rve Fig 5 3 and the ,
.
,

ma x im u m derived from this curve is P = 4 1 2 0 cos 04 , - .

For the p u rpose of determining the ma x im u m obviously ,

not the entire curve needs to be calc u lated but only a short ,

range near the maxim u m This is located b y trial Th us . .

in the present instance P and cos 0 are calc u lated for s = 0 1


,
.

and s As the latter gives lo w er po w er f actor the ma x im u m -


,

po w er f actor is be low 8
-
Then 8 is calculated and gives
a higher val ue o f cos 0 than s = 0 1 ; that is the ma x im u m is .
,
.

below s = o l Then s = 0 02 is calc u lat ed and gives a lo wer


. . .
,

val u e of cos 0 than s = 0 05 The ma x im u m val u e of cos 0 . .

th us lies between s = o o2 and s = 0 1 and only the part of the


. .
,

c u rve between s = o 02 and s = 0 1 needs to be calc u late d for


. .

the determination of th e ma x imu m of cos 0 as is done in Fig 53 , . .

99
. When de t ermining an e x tremum o f a f u nction =
y f (x ) .

by plotting it as a c u rve the valu e o f x at w hich t he e x treme


, ,
15 0 E NG I NE E RING M A TH E M ATI C S .

occurs is more or less inaccurate since at the e x treme the


, ,

c u rve is hori z ontal For instance in Fig 5 3 the ma x im u m


.
,
.
,

of the c u rve is so flat that the val u e o f po w er P for which ,

cos 0 became a ma x imum may be anyw here between P = 4OOG ,

and P = 4300 within the acc u racy o f the c u rve


,
.

I n such a ca s e a higher acc u racy can frequ ently be reached


,

by not attempting to locate the e x act e x treme b ut two points ,

o f t h e same ordinate on each side of the e x t reme Thus in , .

Fig 5 3 the p o w er P 0 at w hich the ma x im u m power factor


.
,

c os 0 = 0 904 is reached is somewhat u ncertain The val u e o f


.
, .

power factor somew hat below the ma x im u m cos


-
, ,

is reached be f ore the ma x im u m at P 1 = 3 400 and a fter the , ,

ma x imum at P 2 4840 The ma x imum then may be calculated


,
.

as half way between P 1 and P 2 that is at


-
, ,

4 1 2 0 watts .

This method gives us u ally more acc u rate results b ut is ,

based on the ass u mption that the curve is symmetri cal on


both sides of the e x treme that is falls o ff from the e xt reme , ,

val u e at the same rate for lo w er as for higher val u es of the


abscissas Where this is not the case this method of inter
.
,

p o la t i o n does not give the e x act ma x im u m .

E xam pl e 3 The e fficiency of a steam t u rbine no zz le


.
,

that is the ratio of the kinetic energy o f the steam j et to the


,

energy of the steam available betw een the two press u res between
which the no z z le operates 1 8 g1ve n m Fig 5 4 as determined b y ,
.
,

e x periment As abscissas are u sed the no z zle mouth opening


.
,

that is the widest part of the no z zle at the e x haust en d as


, ,

fracti on of that corresponding to the e x ha ust press ure while ,

the no z zle throat that is the narrowest part o f the no zz le is


, , ,

ass u med as constant As ordinates a re plott e d the e fficiencies


. .

This c u rve is not svmm e t rical b ut f alls o ff f rom the ma x im u m , ,

on t h e sides of large r no z z le mo u th far more rapidly than on ,

the side of smaller no z z le mo u th The reason is that with .

too large a no z zle m o u th the e x pansion in the no zz le is carried


below the e x ha u st pre ssiI re pg and steam eddies are prod u ced ,

b v this overe x pansion .

The ma x im u m e fficiency of per ce nt is fo und at the point


P 0 at w hich the no z z le mo u th corresponds to the e x haust
,

press u re If however the ma x im u m is determined as mid


.
, ,

wav between two points P I and P 2 on each side o f the ma x im u m , ,


1 52 E NG I N E E R I NG M ATH E MATI CS .

Thus in apparat u s design when determining e xtrema of ,

a f u nction to select them as operating condition ,

co ns ideration m ust be given to the shape of the curve and ,

where t he c u rve is unsymmetrical the most e ffi cient operating ,

point may not lie at the e x treme but on that sid e o f it at which ,

the curve falls o ff slo w er the more so the greater the range of
,

variation is which may occ u r d u ring operation T his is not


, .

always reali zed .

100 I f the f u nction


. is plotted as a c u rve Fig ,
.

50 at the e x tremes o f the f u nction the points P I P 2 P3 P 4


, , , , ,

of c u rve Fig 50 the tangent on the c u rve is hori z ontal since


.
, ,

FI G 5 5. . S t ame Tu rbine No le E ffici ncy; De te rm i nation of Max im


zz e um .

at the e x treme t h e fu nction changes f rom rising t o decreas ing


( ma x im u m P 1 and P 2)
,
or f rom decreasing to increasin g ( min
,

im u m P3 and P 4) and there fore for a moment passes thro u gh


, ,

the hori z ontal direction .

I n general the tangent of a c u rve as that i n Fig 5 0 is the


, , .
,

line which connects t wo points P and P of the c u rve which ’
,

are infinitely close together and as seen in Fig 5 0 the angle , , .


,

0 which this tangent P P makes with the hori z ontal or X —


” ’
,
a xis ,

t hus is given b v

tan 0
M AX I M A AN D M I NI M A . 1 53

At the e xtreme , the tangent on the c u rve is hori z ontal ,

that is A 0 = O and there f ore it f ollo w s that at an e xtreme


, , , ,

o f the f u nction ,

y
dy
_ 0
dx

The reverse ho w ever is not , ,


ne ce ss arlly the case ; that is ,

d
if at a p0 1nt x , y: th 1s po i nt may not be an e x treme ;
that is a ma x im u m or minim u m but m a y be a hori z ontal
, ,

in fl ection point as points P5 and P 6 are i n Fig 5 0


,
. .

With increasing as when passing a ma x im u m ( P I and P 2 , ,

Fig .
y rises then stops rising and then decreases again
, , .

When passing a minim u m ( P3 and P 4) y decreases then stops ,

decreasing and then increases again When pas sing a hori z ontal
,
.

in fl ection point y rises then stops rising and the n starts rising
, , ,

again at P5 or y decreases then stops decreasing but t hen


, , , ,

sta rt s decreasing again (at P 6 ) .

The points o f the f unctio n y= f (x ) determined by the con


d
dition ,
d Z =0
,
thu s re qu lre further 1 nve st 1 gat 1 o n, whether t he v

represent a ma x im u m or a minim u m or merely a h e li zo nt a l , ,

in fl ection point .

This can be done mathematically : f or increasing at when ,

passing a ma x im u m tan 0 changes f rom positive to negative ;


,

that is decreases , ,
or in other words (tan Since

13
=
Z
Z ,
it th u s f ollo w s that at a ma x im u m
33
1 0 . I nversely ,

at a minimum tan 0 changes f rom ne g t ive to positive hence


ig ,

increases that is , ,

dr
(tan 0) 0; or
ii

x
2
0 . When passing
a hori zontal in fl ection point tan 0 first decreases to zero at
the inflection point and then increases again ; or inversely
, ,

tan 0 first increases and then decreases again that is tan 0


, , ,

has a ma x im u m or a minim u m at the in fl ection point and ,

=0 at the inflection point .


1 54 E NG I N EE RING M A TH E AI A TI CS .

In engineering problems the investigation ,


whether the
solution of the condition of e x tremes ,
represents a
mini mum or a ma x imum or an in fl ection point is rare lv
, , ,

requ ired b ut 1t is almost always obvio u s f rom t h e nature o f


,

the problem whether a ma x im u m of a min im u m occ u rs or ,

neither .

For instance i f the problem is to determin e the S peed at


,

which the e ffi cien cy of a motor is a max imum the sol ution : ,

speed = 0 obviously is not a m a mm um b ut a mi mim urn as at


, ,

z ero speed the e fficiency is z ero If t he probl em is to find .


,

the current at w hi ch the o u tput o f an alternator is a ma x im u m ,

the solution i = 0 obvio u sl y is a mini mum and of the other ,

two solutio ns i 1 and i 2 the larger value i g again gives a


, , , ,

mini mu m z ero o u tp u t at S hort —circuit current while the inter


, ,

mediate value i 1 gives the ma x i m u m .

10 1 The e x tremes of a f u nction t h ere f ore are determined


.
, ,

by equ a ti ng its di fferential qu otient to z ero as is illu strated ,

by the following e x amples


E xam pl e 4 I n an imp u lse t u rbine the speed of the j et
.
,

( steam j et or water jet ) is S 1 At what peripheral speed S 3 is.

the outp u t a ma x im u m .

The impu l e force is proport ional to the relative speed o f


s

the j et and t he rotating imp u lse w heel ; that is t o ( 8 1 ,
7

The power is impulse force times spe ed S 2 ; hence ,

P —S
kS g (S 1 g ) ,

P
and 1s an e x treme f or the val u e o f S2, g i ven by
22 = 0; hence ,

_
S r

8 1 2S 2 = 0 and Sz ,
2

that is when the peripheral speed of the imp u lse wheel e quals
,

half the jet ve lo cit v .

E x am pl e 5 I n a tr ans f ormer of constant impressed


.

e m f e o = 2 300 volts ; the constant loss t hat is loss w hich


. . .
, ,

is indepe ndent o f the o utp u t (iron loss ) is Pi = 5 00 watts The ,


.

i nternal re u st ance ( primary and secondary combined ) is r = 2 0


1 56 E NGIN E E RING M ATHE M ATICS .

is shunted by a resistor of resistance r What m ust be the .

resistance of this shunting re 81 stor r to make the pow er con ,

s u med in r a ma x im u m ? ( Fig
,
.

Le t i be the current in the sh u n ting resisto r r The po w er .

consumed in r the n is ,

P ri
2
( 9)
The current in the resistor is i —i and there f or e the ro
0 ,

voltage cons umed by ro is r0 (i o — i ) and the voltage cons u med ,

by r is ri and as these two volt ages m u st be e q ual since both


,
-
,

FI G 5 6 . . S hunte d Re s i stor .

resistors are in shunt with each other thus receive the same ,

voltage ,

ri = ro i 0 - i
( ),
and here f rom it f ollows that
, , ,

Su bstituting this in eq u ation ( 9) gives ,

dB
and thi s power is an e trem e f or x — 2
0; hence
fir

( 2 ( 7 “i“To)
0

2 2 '
r 77 0 1 0

( T


7 To ;

that is the power consumed in r is a ma x im u m i f the resistor


, ,

r of the sh u nt eq u als the resistance ro .


IlI AX I M A AN D M I NI MA . 1 57

The c u rrent in r then is by e qu ation ,

and the power is ,

1 03. I f, after the function (the e quation ( 1 1 ) in


e xample has be en derived the di fferentiation ,
is
immediately carried out the calc u lation is very f re qu ently
,

much more complicated than necessar y I t is there f ore .


, ,

advi sable not to di fferentiate immediately b u t first to simpli f y ,

the function y=f (x) .

I f y is an e x treme any e x pression di ff ering t h e re fo rm by


,

a constant term or a constant factor etc also is an e x treme


, ,
.
, .

So also is the reciprocal of y or its squ are or sq u are root etc


, , ,
.

Thu s before di fferentiation constant terms and constant


, ,

factors can be d ropped fractions invert ed the e x pression


, ,

raised to any power or any root thereof taken etc ,


.

For instance in the preceding e x ample in equation


, ,

22
rr 0 i o

(r re )
the val u e o f r is to be found ,
w hich makes P a m ax im u m .

If P is an e x treme ,

(7

which di ffers I re m P by the omission of the constant factor


2
r0 i 0
,
2
also is an e x treme .

The reverse o f y, ,

is also an e x treme (yg is a minim u m where yl is a ma x imum


.
,

and inversely ) .

Therefore the equation ( 1 1 ) can be simpli fied to the form


.

r + r0 ) 2
r
2

r
1 58 E NG I NE E RING M ATH E M ATI C S .

and leaving
,
o ut the constant term 2 r o, gives the final form ,

=r+
y3

This di fferentiated gives ,

T = T0 .

1 0 4. E x am ple 7
From a source of constant alternating
.

e m f 0 power is trans mitted over a line of resistance r o and


. . .
,

reactance 350 into a non inductive load -


What m u st be the .

resistance r o f this le ad to give ma x im u m power ?


I f i = cu rre nt transmitted over the line the power deli vered ,

at the load of resistance r is


P ri
g
.

The total resistance o f the circuit is r + ro ; the reactance


is 2 0 ; hence the current is
e
2
2
(r $0

and by s u bstituting in eq u ation


,
the power is
2
re
2 ”
( r + r o) + 500

e x treme , omitting 6
2
and i nverti ng ,

(r ro ) z
+ x0 2

T+ 2 To

is also an e x treme and li kewise , ,

32
1
r

is an e x treme .
1 60 E NGIN E E RING IlI A TH E M A TI C S .

dropping the constant term ro gives


= r
y5 ;

that is the curre nt i is an e x treme when y5 = r is an e x treme


, ,
~

and this is the case for r = 0 and r = oo r = 0 gives ,

as the ma x im u m val u e o f the c u rrent and ,


r = oo gives

as the mi nimum val ue o f t h e c u rrent .

With some practice from the original e quation


,
imme
diat e ly or in very fe w steps the simpli fied final eq u ation can
, ,

be derived .

105
. I n the calculation of ma x ima and minima of engineer
ing qu antities x y by di fferentiation o f the f u nction
, ,

it m u st be kept in mind that this method gives the values of


x for w hich the qu antity yof the mathematical eq u ation y f ( 33)
,

becomes an e xt reme but w hether this e xtreme has a physical


,

mea ning in engi neering or not re quires fu rther investigation ;


t h at is the range of numerical values o f a: and y i s unlimited
,

in the mathematical e q u ation but may be li mited in its engineer


,

ing application For i nst ance i f a: is a resistance and the


.
-
, ,

di ff erentiation o f y=f (23) leads to negative val u es of it these ,

have no engineering meaning ; or i f the di fferentiation leads,

to val u es o f x whi ch s u bstit uted in y=f ( x ) gives ima ginary or


, , , ,

negative val u es of y the res u lt also mav have no engineering


,

application I n still other cas es the mathematical res u lt


.
,

may give values w hich are s o far b e vo nd the range of i nd u s


,

t rially practicable nu mer i cal val ues a s to be inapplicable .

For ins tance :


E x am pl e 9 I n e x ample
. to determine the resist ance
r,
which gives ma x im u m current transmitted ove r a trans
mission line the equation
,
M AX I M A AND M INIM A . 161

immediately di ff erentiated gives as condition of the e x tremes


,

dl:
dr

hence either
,
r ro = O;

+ r0 ) + =
(
2 2
r xo co

the latter equ ation gives r = oo hence t = 0 the m i n i mum val u e


,

o f c u rrent .

The f ormer e q u ation gives


To, (2 4)
as t ne val ue o f the resistance w hich gives ma x im u m c u rrent
, ,

and the c u rrent wo u ld then be by s u bstit uting ( 2 4) into ,

The sol u tion however has no engineering meaning , ,

as the resistance r cannot be negative .

Hence m at he m e t ically there e x ists no ma x imum val u e


, ,

o f 75 in the range o f r w hich can occ u r in engineering that is , ,

within t he range O< r < 00 ,


.

I n s u ch a case where the e x treme falls outside of the range


,

o f numerical values to which the engi neering quantity is


,

limited it follows that within the engineering range the q u an


,

tity contin u o u sly increases toward one limit and contin u ously
decreases toward the othe r limit and that there fore the two ,

limits of the engi neering ra nge of the q u antit y give e x tremes .

Th u s r = 0 gives the ma x im u m r = oo the minim u m of c u rrent


, .

1 06 E x a m pl e 1 0
. An alternati ng c u rrent generator
. of -
,

generated e m f e = 2 500 volts internal resistance ro = 0 2 5


. . .
, .

ohms and synchrono u s reactance xo = l 0 ohms is loaded by


, ,

a circ uit comprising a resistor of constant resistance r = 2 0


ohms and a reactor o f reactance x in series with the resistor
,

r. Wh at val ue o f reactance x gives ma x im u m outp ut ?


If f = ourre nt of the alternator its power output is ,

'

P n 2
2 01 ; 9
1 62 E NGINE E RING M ATH E M ATI CS .

the total resistance is r + ro = 2 0 2 5 ohms ; t h e total reactance .

is a: + sco = 1 0 + 1 ohms and t here f ore the c urrent is


,

and the power outp u t by s u bstituti ng ( 2 7 ) in ,

2
re 20
'

( 1 0 + x)
2
( r + r0 )
2
( a: + x o) 2

Simpli fied , this gives

( r + r0 ) ( a: + wo)
2 z
y1 ;

=
y2 ( x i -
23 0 2
5

+ 330 ) = 0 ;

o t '

that is a negative or c onde nsive reactance of


, ,

power output would then be by substituting ,

2 2
re 20 2 5 00
watts = 305
(r + r 0 ) 2

I f,
however a c ond e nsive reactance is e x cluded that is
, , ,

it is assumed that 3: > 0 no mathematical e xt reme e x ists in the


,

ra nge of the variable x w h ich is permissible and the e x treme


, ,

is at t h e end of the range m= 0 and gi ves , ,

2
T6 0
2 45 kw .

7
10 . E x am ple
a 500 kw alternator at volt a ge
1 1 . In -
.
,

e = 2 500 the friction and windage loss is


,
Pw = 6 kw the iron .
,

loss P¢ = 2 4 kw the field e x citation loss is Pf = 6 kw and the


.
,
.
,

armat u re resistance r = 0 1 ohm At what lo ad is the e ffi cie ncy


. .

a ma x im u m ?
1 64 E NGINE E RING M ATH E M ATI CS .

is a f u nction of two independent variables u and s u ch a


pair o f val u es of u and of v is to be found which makes y a ,

ma x imum or minim u m
,
.

Choosi ng a ny val u e no o f the independent variable u , ,

then a val u e o f 1) can be f o u nd which gives the ma x im u m ( or ,

minim u m ) valu e of y which can be reached f or u = u o T h is


,
.

is done b v di ff erentiating y= f over thus

df ( u o, v)
dv

From this e quation a val u e ,

v =fl ( u 0 ) )

is derived which gives the ma x im u m val u e o f y for the given


, ,

val u e of no and by s u bst ituting ( 3 9 ) into


,

y = f2

is obtained as t h e equ ati on which relates the di fferen t e x tremes


,

o f y that correspo nd to the di fferent valu es o f no with u o


, , .

Here f rom then that val u e o f u o is fo u nd which gives the


, ,

ma x im u m of the ma x ima b v di fferentiation : ,

df9 ( u 0 )
d UO

G eometrically mav be represented by a surf ace


,

in space w ith the coordinates y u v y f


, t h en represents , ,
.
,

the c u rve of intersection o f this s u rface with the plane u 0 =


constant and the diffe re nt at ion gives the ma x im u m point
,

o f this intersection curve y f 2 ( uo) then gi ves the curve


= .

in space which connects all the ma x ima of the various inter


,

sections with the u o planes and the second di fferentiation ,

gives the ma x imum o f this ma x im u m c u rve or the


m a x imum o f the m a x i ma ( or more c o rre ct lv the e x treme of ,

the e x tremes ) .

I nversely it is possible first to di fferentiate over u , thus


, ,
and thereby get
( 43 )
as the val u e o f u w hich makes y a ma x imum for the given
,

val ue o f v= vo and s u bstit uting ( 43 ) int o


,
-

y (44)
is obtained as the e quation o f the ma x ima ,
w hich di fferentiated
over v0 thus , ,

gives the ma x imum of the ma x ima .

G eometrically this represents the consideratio n of the


,

intersection curves o f the su rfac e with t he planes v= const ant .

H o w ever equ ation s (38 ) and (41 ) ( respectively (42 ) and


,

give an e x tremu m only if both equations repr e sent ,

max ima or both m ini ma I f one o f the equ ations rep resents
,
.

a max imum the other a mi ni m u m the point is not an e x t re


, ,

m u m b u t a saddle p oint so called f rom the shape of the su r


, ,

f ac e y= f (u v) near this point


,
.

The working of th is will b e plai n f rom the following e x ample


1 09 E x am ple 1 2
. The alternating voltage e = 30 000 is
.
,

impressed upon a transmission line of resistance r o = 20 ohms


and reactance x 0 = 50 ohms .

What sho u ld be the resistance r and the reactance x o f the


receivi ng circ u it to deliver ma x imum power ?
"
Le t i = cu rre nt delivered i nto the receiving circuit The .

total resistance is ( r + r o) ; the total reactance is (2: hence ,

the curren t is

+ ro z ) (1:
The po w e o u tp u t is r P r 2
i ;

henc e s u bstitu ting


,
gives
re z


.

( T l (x + x 0 )
” 2

(a) Fo r any given val u e o f r, the reac tanc e x, w hich


P
max i mum p o w e r ,
is d eri ved by
fi x
E NGIN E E RIN G M ATHE M ATI CS

P simp lifi ed , gives y1 henc e ,

dy1
+ x0 ) = 0 and a:
d z:

that is for any chosen resistance r the power is a max imum


, , ,

i f the reactance of the receiving circuit is chosen e qual to that


o f the line b ut of opposite si gn that is as co nde nsive reactance
, , ,
.

Substituting ( 49 ) into (48 ) gives the ma x imum po wer


available for a chosen value o f r a s : ,

or , simp li fied ,

and gg
= r+

and r To,

and by substituting ( 5 1 ) into the ma x im u m power is ,

2
0

( b ) Fo r any given value of x, the resistance r which ,

dP
ma x i mum power IS gi ven by
,
E '

P simp lified gives ,

( r + 0
33 2
.

31
r r

r /
x ro2
(x x0 ) 2
, ( 53 )
which is the value o f r that for au v given val u e o f 13 gives
, ,

ma x imum power and this ma x im u m power by s u bstit u ting


,

( 53 ) into (48) is ,
i

l2 2
V ro
z
(I I o €

[ x/ ro2 ( x + x o)
2
ro ( a:
2
6
1 68 E N GINE E RING M A TH E M A TI CS .

and the voltage at the receivi ng end o f the li ne would be

that is the voltage at the receiving end wo u ld be far hi gh er


,

than at the ge nerator end the current e x ce ssive and the e fficiency
, ,

of trans mission only 5 0 per cent This e x treme case th u s is .

hardly practicable and the concl u sion wo u ld be that by the


,

us e of negative reactance in t h e receiving circ ui t an amo u nt ,

of po w e r co u ld be delivered at a sacrifice of e fficiency far , ,

greater than economical trans mission wo u ld permit .

I n the case where capacity was e x cl u ded from the re ce iv


,

ing circuit the ma x imum power w as given by equ ation ( 5 6 ) as


,

6 1 0 0 kw .

I I I . o f engineering q u antities
E xt remes x, y, are u sually
determined by di ff erentiating the f u nction ,

y
and from the e quation ,

deriving the values o f x whi ch make y an e x treme ,


.

Occasionall y ho w ever the e quation ( 5 8) cannot be solved


, ,

f or as b ut is either o f higher order i n x or a transcendental


, ,

e qu ation I n this cas e e quation ( 58) may be solved by appro x


.
,

imation or pre fe ra b lv the function


, , ,

is plotted as a c u rve the val u es of 36 taken at which z = 0


, , ,

that is at w hich the c u rve inte rsects the X —axis For inst ance :
,
.

E x am pl e 1 3 The e m f w ave o f a three phase alternator


. . . .
-
,

as determined by oscillograph is represented by the e quation , ,

e= 3 6000 lsin 0 sin (23 —


) sin ( 5 0
0 —
0 2 3 23
sin ( 7 0
M AX I M A AND M I N I M A . 1 69

T his alternator connected to a long distance transmission line


,
-
,

gives the charging c u rrent to the line o f .

i = cos ( a c os (3 0 cos ( 5 0
— 99
cos ( ) (3 1 )
0
+ 1 9 59 7 0

(see Chapter I I I paragraph ,

What are the e x treme values of this current and at what ,

phas e angles (9 do they occ u r ?


The phase a ngle 0 at which the current 3 reaches an e x treme
,

value is given b y the e quation


,

FI G 5 7. .

Su bstituting ( 6 1 ) into ( 62) gi ves ,

d
— 1 sm ( ta sm si n
z ‘

da
+ 1 51 2 (3 0 + 93 3

(50 sin ( 7 0 = 0 .

This e quation cannot be solved f or 0 Therefore z is .

plotted as f u nction of 0 by the curve Fig 5 7 and from this , .


,

c u rve t h e valu es o f 0 taken at which the c u rve intersects the


z ero line T h ey are :
.

°
47
1 70 E NGINE E RING M ATH E M ATI CS .

For these angles 0 the corresponding values of


,
13 are calculated
by equation and are :
i

to = + 9 ; —1 9
+3 ; — 30 ; + 30 ; — 42 ; +4 amperes
; .

The current t hus has d u ring each period 1 4 e x trema o f ,

which the highest is 42 amperes .

1 12 I n those cases where the mathematical e x pression


.
,

of the f u nction y=f (x) is not known and the e xtreme val ues ,

therefore have to be determined graphically frequently a greater ,

accuracy can be reached by plotting as a curve the di ff erential


of y=f (x) and picking out the zero val ues instead of plotting
= x) and picking out the highest and the lowest points
y f( ,
.

I f the mathematical e x pression o f y= f ( x) is not known o b vi ,

o u sly the e qu ation of the di fferential curve z =—


Z
Z ( 64) is usually
not known either Appro x imately however it can fro
.
, ,

quently be plotted f rom the numerical val ues of y=f ( x ) as ,

follows
If x 1 , x2 , x 3 are s u ccessive numerical values of x,

and y1 , y2 , y3 the corresponding n u merical values of y ,

dy
a p p ro x i mat e po m t s o f the d i ff erent i al curve are gi ven
by the corresponding valu es :

as abscissas

— — —
yz yl 92 y4 t/3
as ordinates : . ,

232 — 331 233 — 1132 1134 — 1133

1 1 3. E x am ple
the problem 1 4, In
the ma x imum permea
b ilit y point of a sample of i ron of whi ch the B H curve is given , ,

as Fi g 5 1 was dete rmined by taking from Fi g 5 1 corresponding


.
,
.

val ues of and H and plotting n


B ,B in Fig 5 2 = 2 again s t
,
H
. .

A considerable inaccuracy e x ists in this method in locating ,

the value of B at w hich p is a ma x imum due to the flatness


, ,

of the cu rve Fig 5 2 , . .


1 72 E NGI NE ERING M ATHE M A TI CS .

E x am ple A storage
battery o f n = 80 ce lls is to be
15.

connected so as to give ma x im u m power in a constant resist


= 0 1 E ach battery cell has the e m f co = 2 l
ance r ohm . .
. . . .

volts and the internal resistance ro = 0 02 ohm How must . .

the cells be connected ?


Assu ming the cells are connected with r in parallel hence .
,

in se ri es . The internal resistance of the battery then is

71
ohms and the total resistance o f the circ uit i s
,

FI G 58
. . irst D i ffe re ntial Q u otie nt of
F B ,p Cu rve
n n
The e m f acting on the circ u it is 6 0’ since cells of e m f
5
. . . . . .

6 0 are in se ri es . There f ore the c u rrent d elivered by the battery


,

is ,

and t he po wer which this c u rrent prod u ces in the resistance


r, is )

P = ri 2
M AXI M A AND M I NI M A .
1 73

T hi s i s an e x treme if ,

is an e x t reme hence ,

that is x = 4 cells are conn ected in multiple and


, ,

cells in series .

51 1 . E x am pl e
an alternating current transformer the
1 6, In -

loss of po w er is li mited to 9 00 watts by the permissible temper


at u re rise The internal resistance of the trans f orme r w inding
.

( primary pl us secondary red u ced to the primary ) is 2 ohms


, ,

and the core loss at 2 000 volts impressed is 400 watts and , ,

varies with the ] 6t h power of the magnetic density and there


.

fore of the voltage At what impressed voltage is the outp ut


.

of the trans f ormer a ma x im u m ?


I f e is the impressed e m f and i is the c u rrent input the . . .
,

po w er inp ut into the trans f ormer ( appro x imately at non ,

ind u ctive load ) is P = e i .

I f the o utput is a ma x im u m at constant loss the input P , ,

also is a ma x im u m The loss of power in the windi ng is


.

ri = 2
2 i 2
.

The loss of power in the iron at 2 000 volts impressed is


400 watts and at impressed voltage e it therefore is
,

and the total loss in the transformer there f ore is , ,

PL = 2 1 2 + 4
0
1 74 E NGINE E RING M ATH E M ATI CS .

he re from , It f ollows that ,

'

and substit u ting into


, ,
P = e1 :

P= e 4 5 0 — 2 00

Simplified , this gives ,

and di ff erentiating
, ,

dy
de

Hence .

and e = 2 300 volts ,

which . s ubstit u ted gives ,

52 kw .

1 1 6. 7 E xam pl e 1 . I n a 5—kw
alternating current transformer .
-
,
2
6 i
at 1 000 volts impressed the core loss is 0 watts the r lo ss
, ,

1 5 0 watts How must the impressed voltage be changed


.
,

to give ma x imum e fficiency ( a) At full load o f 5—kw ; ( b ) at ,

half load ?
The core loss may be ass u med as varying with the l 6t h .

power of the impressed voltage I f e is the impressed voltage .


,

2 5 00
i =— is the current at f ull load and ,
fl is the current at
e 6

half load then at ,


1 000 volts impressed the f u ll load c u rrent is
,
-

M =5 amperes and since the


,
i 2
r loss is 1 50 watts this , gives
8
1 76 E NGIN E E RING M ATH E M ATI CS .

( b ) Anadditional resistor of resistance r = 2 l O ohms is ’

then inserted in series making the total resistance o f the con


,

denser charging circuit r = 2 50 ohms What is the ma x im u m


,
.

val u e o f the charging c u rrent ?


The e q u ation o f the charging current of a condense r through
a circuit of low resistance is ( Transient E lectric Phenomena
,
“ ,


and Oscillations p ,
.

where
-
=
q C

and the e quation of the charging c u rrent of a condenser th rough ,

a circ u it o f high resistance is Transient E lectric Phenomena


,

and Oscillations p ,
.

where

S u bstit u ting the n umerical val ues gi ves


( )
a i= sin 980 t ;
= 6 6 67 s
( b) i {
5 OO‘

Simplified and diff ere ntiated this gives ,

'

dz
d ; cos 980 t — s 1n 980t 0;

hence tan 980 t


°
980 i 68 5

+ n7c
'

90 8

2 00 0 z_ 500 :
4
~

€ 5
M A X I M A AND M I NI M A . 177

s
+ l 500 t =4
,


l og 4
1 5 00 t
lo g e

t = 0 00092 . sec .
,

s u bstituting these val u es o f t into equations o f the


gives the ma x imum val ues

( a) i = 10 e
1 20 +
. n 7r
64
- " ampe res ;
that is an infinite number of ma x ima of grad u ally decreasing
, ,

values : etc .

46 1 -34
b
( )
-
) amperes

e .

11 8 . E xam pl e 1 9 I n an i nd u ction generator , the fric


.

tion losses are P, = 1 00 kw ; the iron loss is 200 kw at the ter


. .

minal voltage of e = 4 kv , and may be assumed as proportional


.

to the 1 6th power o f the voltage ; the loss in the resistance


.

o f the cond u ctors is 1 00 kw at i amperes o u tp u t and mav .


,

be ass u med as proportional to the sq u are o f the current and ,

the losses resulting f rom stray fields due to magnetic saturation


are 1 00 kw at e = 4 kv and may i n the range considered be
. .
,

ass u med as appro xi mately proportional to the 3 2 th power .

o f the voltage Under what conditions of operation regard


.
,

ing output voltage and c u rrent is the e fficiency a ma ximum ?


,

The losses mav be s ummari zed as f ollows :


Friction loss ,
Pf :
1 00 kw :

I ron loss ,
P

loss ,
PC 1 00

Sat u ration loss ,


P3 1 00

hence the total loss is PL Pf P, PC P8


1 78 E NGINE E RIN G M A TH E M ATI CS .

The output is =
P e i ; h e nce ,
'
percentage o f los s is

The e fficiency is a ma x imum if the percentage loss I is a ,

minimum For any valu e of the voltage 6 this is the case


.
,

all

at the current i, given bv 0: hence simplifying and di ffer
di
,

e nt iat ing I,

i = 300

then substituting i in the e x pression of


,
11, gives

and 11 is an e xt reme i f the simplified e xpression


, ,

1 2 1
= + + 33
y 2 4 1 6 9 4
°

c 1
6

is an e x treme at ,

dy 2 .
2
3 1 -6 1 ;
de e 4 e

2+
m 6

and e= 5 5 0 kv
. .
,

and by substituti on the follo w ing val u es are obtained : 11


,

e fficiency per cent ; current i = 8000 amperes ; o u tput


P = 44, 000 kw .

1 1 all probability this o u tp u t is beyond the capacity


9 . In ,

o f the generator as limited by heating The foremost limita


,
.

tio n pro bably will be the 15 7 heating of the cond u ctors ; that is
2 :

,
1 80 E NGIN E E RING M A TH E M A TI CS .

is a f u nction having t h e cons tants a b , ,


c and the f orm of
the f u nction ( 1 ) is kno w n f or instance , ,

= a + b r + 0 132
y , (2 )
and the constants a b c are not kno w n b ut t he n u m erical , ,

valu es o f a n u mber o f corresponding val ues o f x and yare given ,

"
for instance by e xperiment ,
x3 x 4 and y1 y2 y3 y4 , ,
. .
, , ,
.

then f rom these corresponding n u merical valu es and y


the constants a b c can be calc ulated i f the numerical
, , ,

val ues that is the observed points o f the c u rve are s u fficiently
, , ,

n u mero u s .

I f less points x l y1 x 2 y2 are obs erved then the e qu a


, ,
.
,

tion ( 1 ) h as constants o b vio uslv these co nstants cannot be ,

calc u late d as not s u fficient data are available therefor


,
.

I f the n u mber o f observed poi nts e qu als the n u mber o f con


s t a nt s they are j u st s u fficient to calc u late the constants
,
For .

instance in eq u ation ,
i f three corresponding valu es x 1 y; ; ,

1 2 yo ; x 3 y; are observed by s u bstit u ting these i nto e quation


, , ,

three eq u ations are obtai ned :


= a + b x 1 + er 1 2
yl ;

y2
= a+ b x 2 + CI 2
2
;

y3 as b i: 0 x3
2
,

which are j u st s u fficient f or the calc u lation o f the three const ant s
a, b, 0 .

Thre e observations wo u ld there f ore be s u ffi cient f or deter


mining three c onstants i f the observations w ere absol utely ,

correct This ho w ever is not the case b ut the observations


.
, , ,

always contain e rrors of ob serva ti o n that is u navoidable inac , ,

c u racies and constants calc u lated by u sing onl y as many


,

o bservations as there are con s tants a re not very acc u rate , .

Th u s in e x perimental work al w ays more observations


, ,

are made than j ust n ecessary for the determination o f the

c onst ants f or the p u rpose of getting a higher acc u racy Th u s


, .
,

for instance i n astro nomy for the calc u lation of the orbit of
, ,

a comet less than f o u r observations are t h eoretically s u fficient


, ,

b u t i f possible h undreds are taken to get a greater accuracy ,

in the determination of the constants of the orbit .


M AXI M A AND M I NI M A . 181

If,then for the determination o f the constants a b c of


, , ,

equ ation s i x pai rs o f correspondi ng val u es o f x and y were

determined any three of these pai rs wo u ld be s u fficient to


,

give a b c as seen above but u sing di fferent sets o f three


, , , ,

obs ervations wo u ld not give the same val ues o f a b c ( as it


, , ,

sho u ld i f the observations were absol utely acc u rate ) b ut


, ,

di fferent values and none o f these val u es w ould ha ve as high


,

an accuracy as can be reached f rom the e x peri mental data ,

si nce none of the values uses al l observations .

12 1 . If (1)
is a f unction containing the constants a b c w hich are still , ,

u n kno w n and x 1 y1 ;
,
x2 y2 ; etc are c orresponding
, ,
.
,

e x perimental va l u es then i f these valu es w ere absol utely cor


, ,

rect and the correct val u es of the constants a b c


,
chosen , , ,

=
y1 f ( x 1 ) wou ld be true ; that is ,

f ( 171 )

f( )r2
.
y
o
= 0, etc .

Du eto the errors of observatio n this is not the case b ut , ,

even i f a b c are the c orrect val ues


, , ,

y ¢ 1 f( x1) etc .
;

that is a small di ff erence or error e x ists th u s


, , , ,

instead of the correct val ues of the constants a b c


If , , ,

other val ues w ere chosen di fferent errors 5 1 5 2 would , ,

obvio u s l y res u lt .

From probability c a lc u lation it f ollows that i f the correct , ,

val ues of the constants a b c are chosen the s um o f the , , ,

squ ares of the errors ,

is less than f or anv other value o f the constants a, b, c


is it is a minim u m
,
.
1 82 E NGINE E R I NG M A TH E M ATI CS .

The problem of determinin g the constants a b c


1 22 .
, ,
.

th us consists in finding a set o f constants which makes the ,

sum of the s qu ares of the errors 5 a mi ni mu m ; that is ,

3 25 2
minim u m , ( 9)
is the requ irement which gives the most acc u rate or most
,
'

probable set o f val ues o f the constants a b c , ,


.

Since by 3 =f ( x ) — y it f ollo w s f rom ( 9 ) as the condi


,

tion which gives the most probable val u e o f the constants


,

a, b, c

z 2 y}
?
minim u m ; ( 1 0)
least su m o f the squ ares o f the errors gives the most
t hat is , t h e
probable val u e of the constants a b c , ,
. .

To find the val u es o f a b c w hich f u lfill e qu ation


, ,

the di ff erential q u otients o f ( 1 0 ) are e qu ated to zero and give ,

2
l
jjf
d’
2 (fi x) yi

df
2
Z iao —
y;
5
( 0
0; etc .

This gives as many e q u ations as there are constants a b c , ,

and there f ore j u st s u ffices for their calc u lation and the val u es ,

so calculated are the most probable that is the most accurate , ,

values .

Where e xt remely high acc u racy is re quired as for ins tance ,

in astronomy when calculating f rom observations e xt endin g


over a fe w months only the orbit of a comet which possibly
,

lasts tho u sands o f years the method of least sq u ares m u st be


,

used and is freq u ently necessary also in engineering to get


, ,

f rom a limited n u mber of observations the highest accuracy


of the c onstants .

123 As instance the method of least squ ares may be applied


.
,

in separating from the observations of an ind u ction motor ,

w hen runni ng light the component losses as friction hysteresis


, , , ,

etc .
1 84 E N GIN E E RIN G M ATH E M A TI C S .

as indicated by the abrupt rise of curre nt and of power b ey ond


47 3 volts This obviously is d ue to beginning ma gneti c s atura
.

tion of the iron stru ct u re Since with be ginni ng saturation


.

a change o f t h e magnetic distrib u tion m u st be e x pected that ,

is an increase o f the ma gnetic stray field and thereby increase


,

of e dd v current losses it is probable that at this point the con


,

FI G 5 9
. . Ex citatio n Po we r of I nd ction Motor
u .

st ant sin e quatio n ( 1 3 ) change and no set of constants can be


,

e x pected to represent the entire range of observa tion For .

the calcula tion of the constants in th u s only the observa


tions below the range of magnetic saturation can safe lv be used ,

that is up to 47 3 volts
,
.

From e quation ( 1 3 ) follows as the error o f an indivi dual


observation of e and po z

po ;
III AX I M A AN D MI MI MA . 1 85

hence
= 1 6 2 — 2 =
z 25 2 =
E fa + b e °

+c e
p oi mi ni m um, ( 1 5)

+ be — =
p ol
l '6 2
E {a . + ce 0:

=
E fa + b e l 6
+ 2—
j 1 '6
0;
°

ce po e

Z {a + b e l '6
+ ce 2

and i f n is the n u mber o f observations used ( n = 6 i n this


,

instance from e = 1 48 to e = 47 3 ) this gives the f ollowing


, ,

e quations :
na +b2e 1 '6
Z
+c e 2
2 po = 0 ;

Ze l '6
+b2e 3 3
+ c Ee 3 '6
Ee 1 '6 = 0;
°

a po

a2e +b2e
2 3 '6
2
+0 e 4
2 e2 p0 = 0 .

Su bstituting in ( 1 7 ) the n u merical val u es f rom Table I gives ,

a + 11 7 b . c

a + 14 6 b . c 1 0 = 1 s30 ;
3

a + 15 1 b . c l o = 1 880 ;
3
j
a= 5 40 ;

= 2
po 5 40 6 .

The values of po calculated from e quation


, are given
i n the si x th col u m n of Table I and their di fferences from the
,

observed val u es in the last c ol u mn As seen the errors are in .


,

bo th directions f rom the calc u l ated val u es e x cept for the thr ee ,

highest voltages in w hich the observed val u es rapidly increase


,

beyond the calc u lated d ue probably to the appearance o f a


,
1 86 E NGIN E E RING M A TH E M ATI CS .

loss w hi ch does not exist at voltages the eddy c u rrentslo w er —

cau sed by the mag netic stray field o f sat u ration .

This rapid diverg ency o f the observ ed f rom the calc u lated
val u es at hi g h voltages sho ws that a calc u lation o f the constants ,

based on all observations w o u ld have led to wrong val u es , ,

and de m onstrates the necessity first to critically revie w the , ,

series o f observations be f ore u sing them f or deriv ing constants


, ,

so as to e x cl u de constant errors or u nidirectional deviation I t .

m u st be reali z ed that the method o f least sq u ares g ives the most


probable val u e that is the m ost accu rate resu lts derivable
, ,

f ro m a series o f observations only so f ar as the accidental ,

errors o f observations are co ncerned that is su ch errors which , ,

f ollo w the g eneral law o f pro b ability The method o f least .

s q u ares
,
ho w ever cannot eli m inate constant errors that is
, , ,

deviation o f the observations w hich have the tendency to be


in one direction as cau sed f or instance by an instru ment reading
, , ,

too high or too lo w or the appearance o f a ne w phenomenon


, ,

in a part o f the observation as an additional loss in above ,

instance etc Against su ch constant errors only a critical


,
.

revie w and stu dy o f the m ethod and the means o f observa


tion can g u ard that is j u dgment and not mathematical
, , ,

f ormalism .

The method o f least square s gives the highest accu racy


available w ith a g iven n u mber o f observations b u t is freq u ently ,

very laborio u s especially i f a n u mber o f constants are to be cal


,

c u la t e d
. I t there f ore is mainly employed w here the n u mber o f
, ,

observations is li m ited and cannot be increased at w ill ; b ut w here


it can be increased by taking some more observations as is —

g enerally the case w ith e x perimental engineering investigations


—the same acc u racy is u s u ally reached in a shorter time by

taking a fe w more observations and u sing a simpler method


o f calc u lation o f the constants as the EA method described in ,
-

parag ra phs 1 53 to 15 7 .

D i oph anti c E qu ati o ns .

I 23 A
.

The method o f least sq u ares deals with the case ,

w hen there are more eq u ations than u nkno w n q u antities In .

this case there exists no set o f val u es o f the u nkno w n q u antities


, ,

w hich e x actly satisfies the eq u ation s and the problem is to find , ,


1 86b E NGIN E E RING M ATH E M ATI CS .

hence
— — 2s
15 +

since y m u st be an integ er n u m ber ,


m u st also be an

i nte g er n u mber . C all this a, it is

l — u

since a: m u st be an integ er n u m b er m u st also be an integ er

n u m ber that is :
,

u 3 1)

h e nce , su bstit u ted into ( 4) and ( 2 )


x = l — 4v
y = 13 + 1 1v
z = 2 — v

( 7 ) th u s are the sol u tions o f the eq u ations ( 1 ) where v is any


integer n u mber .

As seen mathe m atically there are an infinite n u mber o f


, ,

sol u tions .

S u bstit u tin g no w f or v inte g er n u m bers

v = + 2 + 1 0 1 2
a: 7 — 3 + 1 + 5 + 9
y = + 35 + 24 + 13 + 2 — 9

z 0 + 1 + 2 + 3 + 4

As seen there are only t wo sol u tions f or v O and v


,
l , ,

w hi ch g ive f or x y and z t h ree inte g er positive val u es and w hich


, , , ,

th u s satis f y the physical restri ction .

v 0 ; a: 1 y 13 z 2 is e x cl u ded by engin e ering con


, ,

sideration as nobody w o u ld consider thirteen exciters w ith one


,

g enerator and th u s there remains only one a p p licable sol u tion :


,
M AX I M A AND M I NI M A . 1 86c

W e thus have here the case o f two equ ations w ith three u n
kno wn q u antities which have only one sin gle set o f these u n
,

kno w n qu antities satis f yin g the p roblem and th u s g ive a definite


,

sol u tion tho u g h mathe m atically indefinite


,
.

This ty p e o f equation has first b e e n st u died by D io ph ant e s


o f Ale x andria
.
1 88 E NGINE E RING M A TH E M ATI CS .

po wers of such quant ities may be dropped and the e x pression


thereby s implified ; or if the q u ant ities are not qu ite as small
,

as to permit the neglect of the ir squares or where a h igh ,

acc u racy is requ ired the fir st and second po w ers may be retained
,

and on ly the cube s and higher powers dropped .

The most common method of proced u re is to resolve the ,

e x pression into an infinite series of successive pow ers o f the


small quantity and then retain of this se ries only the first
,

term or only the first t w o or three terms etc depending on the


, , .
,

smallness of the qu ant ity and the req u ired acc u racy .

125 The fo rms most freq u ently used in the red u ct ion of
.

e x pressions containing small quant ities are m u ltiplication and


d ivision the b inomial series the e x po nent ial and the logarithm ic
, ,

series the sine and the cosine series e t c


, ,
.

D enoting a small q u ant ity by s and w here several occur , ,

by 3 1 3 2 3 3
,
the follow ing e x pression hold s
,

= 1 zt
$ 82 ) S i : i: 8 2 $ 81 82 ,

and since 3 1 3 2 is small compared w ith the small quantities


,

3 1 and 3 2 or as usually e x pressed


, ,
3 1 3 2 is a small q u antity of ,

higher order ( in th is case o f second order ) it may be neglect e d , ,

and the e x pression w ritten :

(1 i sl ) ( + 2) 1 + 8 1
1 3 =

This is one of the most u sef u l simplifications : the mult iplica


tion of terms containing small qu ant ities is replaced by the
s imple addit ion of the small qu ant ities .

I f the small qu antities 3 1 and 3 2 are not added ( or s u btracted )


to 1 but to other finite that is not small quant it ies a and b
, , , ,

a and b can be taken o u t as f actors th u s , ,

Ii = ab I i

81 82
where and m u st be small q u ant i t i es
71 5
.

As seen in th is case 3 1 and so need not necessar ily be a b s o


, ,

lu te ly small q u ant ities but may be qu ite large provided that


, ,

a and b are still larger in magnit u de ; that is 8 1 m u st be small ,

compared with a and 3 2 small compared w ith b For ins tance


, .
,
M E TH ODS OF APPROXI M ATI ON . 1 89

in astronomical calc u lations t h e mass o f the ea rth ( w hich


absol utely can certainly not be considered a small quanti ty )
is neglec ted as small q uantity compared wi th the mass o f the
sun Also in the e ff ect o f a lightning stroke on a primary
. .

distrib ution c irc u it the normal line voltage o f 2 2 00 may b e


,

neglected as small compared w ith the voltage impressed b v


lightning e t c , .

1 26 E xam ple
. I n a direct c u rrent shunt motor the im
.
-
,

pressed voltage is e 0 = 1 2 5 volts ; the armature resi st a nce is


r o = 0 0 2 ohm ; the field resistance is r 1 = 5 0 ohms ; the power
.

consumed by frict ion is pf = 300 watts and the pow er consumed ,

by iron loss is p i = 400 watts Wh at is the power o utput of .

the motor at to 50 1 00 and 1 50 amperes inp u t ?


,

The power prod u ced at the a rmature cond u ctors is the


product of the voltage 6 generated in the armat u re cond u ctors ,

and the current 17 through th e armat u re and the power output ,

at the motor p u lle yis ,

P ei —
pf 17.

The current in t h e motor field is and the armature current


there fore i s,

where is a small q u ant ity compared w ith to


7l , .

The voltage co n su med by the arma ture resistance is rot,

and t he voltage ge nerated in the motor armat u re thus


e = 60

where r ot
is a small qu antit y compared w ith e o .

Su bst it u ting herein for i the val u e ( 4) gives ,

c = e o — ro To

S ince the second term o f (6 ) is small compared w ith 60,

e9

C 0 O

and in th i s sec ond term the second term ,


-
is small com
r1

pare d w ith to, it can be neglected as a small term o f h igh e


1 90 E NGINEE RING M A TH E M A TI CS .

order ; that is as small compared w ith a small term and


, ,

e x pression ( 6 ) simp lified to

e eo r ota .

Substituting ( 4) and ( 7 ) into ( 3) gives ,

10 (6 0 r oi o ) Pi Pi

1 pj pi .

(8) conta ins a prod u ct of two terms w ith small


E xp ression
quant ities wh ich can be multiplied by e quat ion
,
and t he re b v
g i ves ,

e 0 50 1

P Pi Pi

To o
2
t pj pi

Subst itut ing the n u merical val u es gives ,

p 1 2510 300 400


1 2 60 appro x imately ;

thus for f0 = 5 0 1 00 and 1 5 0 amperes ; p = 4940


, , , and ,

watts respect ively .

1 27 E x pressions contain ing a small qu a n


. t ity in the denom
inat o r are freq u ent ly s implified b v bringing the small quant ity
in the n u merator by div ision as d isc u ssed in Chapter II para
,

graph 3 9 that is by the series


, , ,

1
2 3
1+ x + x i + r
42
F1 5
( 1 0)
l i q:

w hich series if ,
x is a small qu antity 3, can be appro x imated
by '

H
1 92 E NGIN E E RING M ATH E M ATI CS .

o f the primary c irc u it the generated e m f o the secondary


,
. . .
f

c ircuit is sE .

S ince 131 is the reactance of the secondary c irc u it at full


f requency at the f raction 3 of f u ll frequency the reactance
,

of the secondary c irc u it is 3 1131 and the impedance o f the sec


,

o nd ary c irc u it at slip 3 therefore is r 1 + jsx 1 ; hence the


, ,

secondary current is ,

r1 + js r 1.

the e x c iting current is neglected the primary current


If ,

e quals the secondary current (ass u m i ng the secondary o f the


same number o f t u rns as the primary or reduced to the same ,

n umber of turns ) ; hence the current input into the motor is


,

S E
71 + j8 131

The second term i n the denom inator is small compared


with the first term and the e x pression ( 1 6 ) thus can be
,

appro x imated b v

The voltage E generated i n the primary c ircu it equals t he


impressed voltage e o m inu s the vo ltage cons u med b v the
,

current I in the primary impedance ; r0 + j co thus is .

E = eo

Subst ituting ( 1 7 ) into ( 1 8) give s

In
e x pression the second term on the righ t hand side -
,

wh ich is the impedance drop in the primary c irc u it is small ,

compared w ith the first term c o, and in the factor

of t h is small term the small term j


, can thus be neglected
M E THODS OF APPROXI M ATI ON . 1 93

as a small term of higher orde r and e quat io n ( 1 9 ) abbreviated


,

to

From ( 20 ) it f ollo w s t h at

and by

E = 60 1

Su bst itut ing (2 1 ) int o ( 1 7 ) gives


88 °
i ( ro + xo)
1 j
and by
S

}
r O + 3: 1

.

— s
.

l ] s
rl

— '
then I
If , t is the e x c it ing current ,

current input i nto the motor is appro x imately , ,

I
o
= 1 + 100
86 0
1 +s 33 + 10 31 0 .

1 29. of the most important e x pressions used for t he


One
red uct ion of small terms i s the binom ial series
n(n —1 ) n( n — —
l ) ( n 2)
2

n(n — 1 ) ( n — 2 ) ( n — 3)

If r
. i s a small term 3, this gives the appro x ima t ion ,
1 94 E NGINE E RING M ATH E M ATI CS .

or us ing the second term also it gives


, ,

i ns +

In a more general f orm thi s e x pression gives,

( j; s)
a
"" a 1

the binom ial higher po w ers o f terms containing small


By ,

q uantities and ass u m ing n as a fract i on roots containing


, , ,

small q u antities can be eliminate d for instance


, ,

N ai s ( a i s )
1 1

an l i i
"V"
2
5 li
a na

One of the most common u ses of the binom ia l series is for


the e liminat ion o f sq uares and square ro ots and very fre ,

quently it can be conven iently applied in mere numerica l ca lcu


lat i ons ; as f or ins tance
, ,

1 1

1 —
7 9 00 1 = 900
1 50

_
_
1
V 10V 1

1 1 1
etc
m
.
1 96 E NGINE E RING M ATH E M ATI CS .

or i f for c onven ience the fract ion


, ,
is denoted by a, the n
S
a = and e = ae o,
So

where a is the ratio o f the actua l speed to that speed at which ,

the generated vo ltage is co .

I f r is the resistance of the alternator armature fro the ,

s vnc h rono u s reactance at speed S o the sy nchrono u s reactance ,

at speed S is x = axo and the c u rrent at short circuit then is


,

V 2
r + 33 2 \/ r2 + 2
a ro
z

Usually r and x 0 are of such magn itude that r cons u mes


at fu ll load abo u t 1 per cent or less o f the generate d vo ltage ,

whi le the reactance voltage of i s of the magn itude of from


20 to 50 per ce nt Thu s r is small compared w ith mo and if
.
,

a is not very sm a ll equat ion (2 9 ) can be appro x imated by


,

Then if xo = 2 0 r, the follow ing relat ions e x ist

That is the short c ircu it current of an alternator is pract i


,
-

cally const ant independent of the speed and begins to decrease ,

o nly at very low speeds .

1 31 E x ponential f u nct ions logarithms and trigonometric


.
, ,

f unctions are the ones freq u ently met in e lectri cal engineering .

The e x po nential function is defined by the series ,

? 3 5
at x x
M E THOD S OF AP PROXI M ATI ON . 97

and i f x is a small q u ant ity 3 the e x ponentia l f unct ion may


, , , ,

be appro x imated by the eq uation ,

4
5
3
1i s;

or by the more general eq uat ion



, ,

f
s 1i as ;

and i f a greater acc u racy is req u ired the second term may
, ,

be inc luded thus, ,

e
i 8 = 1 313 8 +

and the n
e
i as = 1i ds —
l

The logarithm is defined by log e x =

log ( 1 ie x ) 21
:

1
Reso lvm g i nto a seri es by ,
and the n i ntegrat i ng ,

2 3 4 5
x x x 1
" i f i i
3 4 s

This logarithmic se ries (36) leads to the appro x imation ,

logs i s;

i nc luding the second term it gives ,

lo g (1 i s) s— 3
2
. i ,

and the more general e xpression is l e S pe c t i ve ly,


+ log : t gai
a
, (3 9 )
1 98 E N GINE E RING M ATH E M A TI CS

and more accur ately


, ,

2
s 3
log ( a i s) = log a i
e

S i nce logl o N = log 1 0 e X lo ge N log; N equations ( 39 )


,

and (40) may be w ritten thus ,


lo g l o ( 1 i 0 4343 8 ,
.

9
= 3
lo g l o (a i ) s log 10 a

1 The trigonometric funct ions are represented by


32 . t he


infin ite series :
x4 x6
2
x
_1" —
FE T
M
9:
3 x5
x x7

if ié 1
S ‘W f ‘ ‘

7
d

w h ich when 3 i s a s m all qu ant ity may be appro x imated ,

cos s =1 and sin s = s;

or they may be represented in closer appro x imation by


,

cos 3 1

s in 3 s

or by the more general e x pressions


, ,

cos a s 1 and cos a s = 1

sin as = as a nd sin as as

1 33 . f unctions contai ning small terms may f reque ntly


Other
be appro x i mated by T aylor s series or its special case ’
, ,

Mac Lauri n s s eries ’


.

Mac Laurin s series is written thus


f ( r ) f (0) + xf f0) 0)

.
=
2 00 E NGIN E E RING M A TH E M ATI CS .

The first appro x imat io ns derived by neglecting all higher ,

terms b ut t he first power of the small quantity x = s in these


series are :
,

l
1 1 3

(1 -l f 718 ;

e
i 8 =1 + s;

10 8 6 i s ) :
i s

cos s = 1 ;

s in s = s ;

f f ) f (0 ) 3 =

tb ) ;

sf
f (5 i

in add ition hereto is to be remembered the multiplication

(1 1 8 00 + 82) 1 1 81
( 52 )
1 The acc u racy of the appro x imati o n can be est i mated
35 .

by calc u lating the ne x t ter m beyond that whi ch is used .

This term is given in brackets i n the above equati ons ( 50 )


and
Th u s when calc u lati ng a seri es of numeri cal val u es by
,

ap pro x imat ion for the one val ue for whi c h as seen by t h e
, , ,

nature o f the problem the appro x imation is least c lose the , ,

ne xt term is c alculated and if this is less th a n the perm iss ible


,

lim its of accurac y the appro x imat ion is sat isfactory


, .

For instance in E x ample 2 of paragraph 1 30 the appro x i


, ,

mate val ue o f the short c ircu it current was found i n as -


M E TH ODS OF APPROX I M ATI ON . 2 01

The ne xt term in the parenthesis of equation by the


n n — l
binom ial wou ld have been
,
s u bst i t u t i ng n

the ne x t be comes + The smaller the a the


8 ,

less e xa ct is the appro x imation .

The smallest valu e o f a considered in paragraph


,
1 30 ,
was
3
a= 02
. . For xo = 2 0r , this gives as the
val ue o f the first neglected term and i n the acc u racy of the,

2 2
9 2
res u lt th is is of the magn it ude o f o ut of X
3 0

the val ue g iven in paragraph 1 30 ; that is the appro x imat ion


,

0 0 1 46
2
.

gives the resu lt correctly W i th i n or W i th i n one


0 3680
si x th of one per cent which is su fficient ly c lose for all engineer
,

ing p u rposes and with large r a the values are still c loser
,

appro x imation s .

1 36.I t is interesting to note the di fferent e x pressions ,

which are appro x imated by and by ( 1 Some o f


them are given in the following :

1
202 E NGINE E RING M ATH E M ATI CS .

14 0 g
. (1
S
1 + 71 lo g
. 1 -


l n lo ge 1

l— sin s;

I— n s in
2 04 E NGINE E RING M ATH E M ATI CS .

3 /2
4a 1+

8
a 1

1 38 . As further e x ample may be considered the equations


of an alternating current electric c ircu it contain ing distribu ted
-
,

resistance i nductance capac ity and shunted conductance f or


, , , ,

instance a lo ng distance transm ission line or an underground


,
-

high pote nt ial cable


-
.

E qu ati o ns of th e Transm i ssi o n L i ne .

Le t I be the d istance along the line from some start i ng ,

point ; E , the vo ltage ; I ,the curre nt at po int I e x pressed as ,

vector quant ities or general n u mbers ; Z o = ro + jxo the line ,

impedance per u nit len gth ( for instance per m ile ) ; Y0 = g0 + jb 0 ,

line ad mittanc e shunted per uni t length ; that is r o is the


, , ,

ohm i c effect ive resistance ; fro the se l f i nduct ive reactance ; ,


-

b o t he co nde ns i ve susceptance that is w at t-less chargin g


, , ,

c u rrent divided b y volts and go = e ncrgy component o f adm it


,

tance that is e nergy compo nent of charging current divided


, , ,

by volts per un it lengt h as per m ile


, , , .

Considering a line e lement dl the vo ltage d E consumed , , ,

by the impe d ance is Z oI dl and t h e current ( II cons u med b y


'

, , ,

the admitta nce is YoE dl; hence the follo w i ng re lation s may be ,

wr itte n
M E TH ODS O F APP ROX I M A TI ON . 2 05

Di ff erent iat ing and s ubst it ut ing (2 ) there in gives


(1
2
13
ac

and from ( 1 ) it -
f ollows that ,

1 dE
'

Z0 dl

E qu atio n 3 ) is integrated bv

E 4

and ( 5 ) s u b s t ituted in (3 ) gives


= /
B i \ Z UYO ;

hence f rom ( 5 ) and


,
it f ollo w s

{4 18
+ vm 01
A2 e

Ne x t as s
u me
l = lc, entire len gth o f line ;
'

the

Z = loZo, the total line impe dance ;


Y = ZO Y0 , the total line adm ittance ;
then su bst ituting ( 9 ) into ( 7 ) and
,
the fo llow ing e x pressions
are obtai ned :

{A1 5
+ VZ Y Ag e

VZ
T’
}

as the voltage and cu rre nt at the generator end of the line .

1 39 .I f now E 0 and I 0 respectively are the c u rrent and


vo ltage at the step —do w n end of the line f or l = 0 by sub , ,
s

t i t ti n l = 0 into ( 7 ) and
s u g
4 +4 =
1 2 EM

i
{i
206 E NG I N E E RING M ATH E M ATI CS .

Substitut ing in f or the e x ponent ial f unct ion the series , ,

ZY Z YV Z Y Z2 Y2 Z2 Y2 V Z Y
s
i zy —
l i V ZY l -
j:
2 6
ZY Z Y 2 2 —
1 + + i x/Z Y
T

and arrangi ng by ( I + 4 2 ) and ( l A2 ) , and substit uting


A 1 A —
h e re fo r the e x pressi ons gives
ZY Z Y 2 2
ZY Z Y 2 2

E 1 = E° 1 +
i 9 24
+ ZI O 1 +
6
+
1 20

ZY Z Y 2 2 f ZY Z Y 2 2

l
1 1 = lo 1 + + YE ° l + +
0
24 6 1 20

When l = —
lo that is for E 0 and I 0 at the generator s ide and
, , ,

E 1 and I I at the step down side of the line the sign o f the-
,

second term of e quations ( 1 3) merely reverses .

1 40 From the foregoing it f ollo w s that if Z i s the total


.
, ,

impedance ; Y the t otal sh u nted adm ittance o f a transm ission


,
'
line E 0 and I o the voltage and current at one end ; E 1 and I 1
, , ,

the voltage and current at the other end of the tra nsm ission

" "
line ; then ,

ZY Z2 Y2 ZY Z 2
Y 2
EF E 1+ i zl 1 +
2 24 6 1 20

ZY 2 2 17 2 ZY Z2 Y2
i YE O 1 + +
2 24 0 1 20

where the pl us sign applies i f E 0 I 0 is the step down end ,


-
,

the m in us sign if E 0 I 0 is the step —up end o f the tra nsmission


, ,

li ne .

pract ically all cases the q u adrat i c term


In ,
ca n be neglected ,

and the eq uations simplifie d th us ,

ZI O 1 +

A = a, 1 + i n} , 1 +

and the error made hereby is of the magn it u de o f less than


2 08 ENGI NE E RI NG M A TH E M ATI CS .

The u ra cy o f the preceding appro x imat io n can b e est i


acc

m ate d by cons idering the physical meaning o f Z and Y : Z


is the line impedance ; hence ZI the i mpedance voltage and ,
°

21
u = the i mpedance voltage of the l i ne fract i on of total
-
as
E , ,

vo ltage : Y is the shunted adm ittance ; hence YE the charging


YE
c u rrent and v= — the chargi ng cu rrent of the kne as fract i o n
,
7 ,

of total c u rrent .

Multiplying gives u v= Z Y ; that is t he cons tant Z Y is the ,

produ ct of impedance voltage and charging current , e x pressed


as fractio ns of fu l l vo ltage and fu l l current respect ive ly I n , .

any economically feas ible power trans mission i rrespect ive of ,

its length both of these fractions and espec i ally the first
, , . ,

must be relat ively sma ll and their product t h erefore is a smal l


,

qu ant ity and its higher pow ers negligible


,
.

In any economically feasible constant potent ia l transm ission


line t he precedi ng appro x im at ions a re therefore perm issible .

Appr o x i m ati on b y C h ai n Fracti on .

A —
A convenient method o f approx imating n u merical
1 4 1 .

val u es is o f ten a ff orded by the chain f raction A chain f raction .

is a f raction in w hich the deno m inator contains a f raction


, ,
w hich
ag ain in its deno m inator contains a f raction etc Th u s : ,
.

O nly integ er chain f raction s that is chain f ractions in which


, ,

all n u merators are u nit y are o f interest ,


.

A com m on f raction is converted into a chain f raction th u sly


APP ROX I M A TI ON B Y CH AI N FRAC TI ON .
z08a

3 1

That is to convert a common f raction into a chain f racti o n


, ,

the n u merator is divided into the deno m inator the resid u e


,

divided into the divisor and so o n u nt il no resid u e remains


, ,
.

The s u ccessive qu otients then are the s u ccessive den omi nator s
o f the chain f raction
.

For instance
20 86 E NGI N E E RI NG M A TH E M A TI CS

511
1 1 52

2
1 0 22
1 30 /5 1 1 3
3 90
1
12 1
13
9
31
27
2
8

hence

I nversely , the chain f raction is converted into a common


-

f racti on by rolling it u p f rom the end


,

1 9
4 4

1 9
4
zo8d E NG I N E E RI NG M ATH E M ATI C S .

di ff erence

0 5 64

428 5 7

4 4 44 4 00 08 6

4 43 548 0 000 2 8 00 68

1 1
13

4 43 5 80 . 00 0004 0 00 9
2 1
3 1

443 57 6
APP ROX I M ATI ON B Y CH AI N FRAC TI ON . 2 o8e

Aseen s u cces s ive a pp roxi m ations are alternately above and


s ,

belo w the tr u e val u e and the ap p roach to the tru e val u e i s


,

extremely ra p id I t is the latter f eatu re w hich makes the chain


.

f raction val u able as w here it can be u sed it gives very ra p idly


, ,

convergin g a pp roxi m ations .

I 4I B C hain f raction re p resentin g irrational n u m bers as


.
-
,

7r 6,
etc may be endless Th u s
,
.
,
.

v
r = 3 + 1

1 1

3 + l

7 +

The first three a pp roxi m ations o f this chain f raction o f 1r are

di ff erence :


3

(2 ) 3 1
3 0 0 00 83 2 00 2 6 %

(3 ) 3 l 3 00000 0 3
7 1
15 1
1

As seen , the first a p pro ximation , 3 1 7 , is already s u fficiently


clo s e f or most practic a l p u r p oses and the third a p p roximati on
,

o f the chain f raction is correct to the 6 t h deci m al .

I 44 —Freq u ently irrational n u m bers


. s u ch as sq u are roots , ,

can be expressed by periodic chain f ractions and the chain ,


20 8 f E NG I NE E RI NG M ATH E M A TI CS .

f raction o ffers a convenient o f ex p ressin g n u merical val ues


way
contai ni n g sq u are roots and deriving their appro x i m ations
, .

For instan ce :
R esolve x/ E i nto a chain f raction .

As the chain f raction is 1 x/E has to be e x p ressed in the f or m


,

v6 = 2 + t¢ 6 —
m a)
and the latter term : ( x/ E w hi ch is < 1 ex p ressed as chai n ,

f racti o n


.

To rationali z e the n u merator w e m u lti ply n u merator and


,

denominator by ( x/é 2)

(v s — 2x ¢ o+ a 2 1
( fl —
m
v 6 + 2 v e+ z v6 + 2
2

is it is ag ain resolved into

v 6 = 2 + 1

contin ui ng in the same manner



x/ 0 2 ( x/ 6 2 1
2 fl v t + 2) mv t + 2> v6 + 2

hence
v %: 2 + 1
2 l

v6 + 2

— m
hence
C HAPT E R VI .

E MPI R I CAL C U RVE S .

A G e n e ra l
. .

1 42 The res u lts of observat ion or tests usually are p lotted


.

in a c u rve Su ch curves f or in s tance are given by the core


.
, ,

loss of an electric generator as f u nct ion o f the voltage ; or , ,

the c u rrent in a c irc u it as f unction of the t ime etc Wh en


, ,
.

plott ing from n u merical observations the c u rves are empirical , ,

and the first and most important problem which has to be


solved to make s u ch curves us e f ul is to find equ at ions for the
same that is find a f unction y= f (x ) which represents the
, , , ,

curve As long as the equ ation of the c u rve is not known its
.

u tility is very limited Wh ile n u merical valu es can be taken


.

from the plotted curve no gene ral concl u sions can be derived
,

from it no gene ral invest igat ions based on it regarding the


,

condit ions of e fficie ncy o utp u t etc An illustrat ion hereof is


, ,
.

a ff orded by the comparison of the e lectric and the magnet ic


c irc u it I n the electric c irc u it the relation between e m f and
.
,
. . .

e
t =
0 0

g i ven by Ohm s law


o o 0

current is ,
and calc u lat i o ns are uni
r

ve rsally and easily made I n the magnetic c ircuit however .


, ,

the t e rm corresponding to the resistance the relu ctance is not , ,

a constant and the re lation between m m f and magnet ic flu x


,
. . .

cannot be e x pressed by a general law b ut only by an emp irical ,

curve the magnetic characterist ic and as the res u lt calcula


, , ,

tions of magnetic c irc u its cannot be made as conveniently and


as general in nature as calc u lat io ns of electric c ircu its .

I f by obs ervation or test a number o f correspond ing va lues


o f the independent variable x and the dependent variable yare

"
determ ined the problem is to find an e quation y=f ( x) which
, , ,

represents these c o rresponding values : 1131 x , and ,

y1 y2 ya
, , yn appro x imatel y that is w ithin the e rro rs o f
, , ,

observation .
210 E NGI N E E RI NG AI A TH E M A TI CS .

The mathemat ica l e x pression which represents an empir ical


c u rve may be a rat ional equ ation or an emp irical eq u at ion .

I t is a rat ional equat ion if it can be derived t h e o re t i ca llv as a


concl usion from some general law o f nature or as an appro x im a ,

t ion thereof but i t is an e mpiric al equation i f no theoretical


,

reason can be seen f or the part icular form of the e qu ation .

For insta nce when represent ing the dy ing out of an elec trica l
,

c u rrent in an ind u ct ive c irc u it by an e xponential f u nct ion o f


t ime we have a rat ional equation : the ind u ced voltage and
, ,

there fore by Ohm s law the c u rrent varies pro po rt ionallyt o the
,

, ,

rate o f change o f the current that is its di ff erential qu ot ient , , ,

and as the e x ponent ial fu nction has the characteristic of being


propo rtional to its differential qu otient the e xp onential function ,

th us rat ionally represents the dying out o f the c u rrent in an


ind u ct ive c ircuit On the other hand the relat ion be tween the
.
,

loss by magnet ic hysteresis and the magnet ic d e ns it v : W 9B = W


,

is an empirical e qu at ion since no reason can be seen for this


law o f the l 6th power e x cept that it agrees with the observa
.
,

t ions .

A rational e qu at ion as a ded u ct ion from a general law of


,

nat ur e app lies un iversally w ithin t h e range o f the observa


, ,

t ions as w e ll as beyond it w hil e an emp irical e qu ation can w ith


,

certainty be re li ed upon only w ithin the range of observat ion


from which i t i derived and e xt rapolat ion beyond this range
s
,

becomes incre as ingly u ncerta in A rat ional e qu at ion there .

fore is f ar pre f erable to an empir ical one As regards the .

acc u racy of represent ing the observat ions no materia l di fference ,

e x ists be twe en a rat ional and an emp irical e qu ation An .

emp irical e qu at ion f requ ent ly represents the observat ions w ith
great accuracy wh ile inversely a rat iona l e qu ation usually
,

does not rigidly represent the o b servat ion for the reason that s
,

in nat u re the conditions on which the rat ional law is bas ed are
rarely per f ectly fulfilled For in s tance the representat ion o f a
.
,

decaying c ur rent by an e x ponential f unct ion i s bas ed on the


ass u mption that t h e res istance and the i nd u ctance of the c ircuit
are constant and capac ity absent and none of these cond it ions
, ,

can ever be perfect ly satisfied and th u s a deviat ion occurs from ,

the theoret ical condition by what is called second ary e ffects


, .

1 43 To derive an e quat ion w hich re presents an emp irical


.
,

c u rve care f u l cons iderat ion sho u ld first be given to the physica l
,
2 12 E NG I NEE RI NG M ATH E M ATI CS .

instead of decreasing as result of the friction le ad when the


, ,

voltage is so low that the mechan ical fri ct ion cons t itute s an
appreci able part o f the motor output ; Th us empirical curves ,

can be represented by a single e quat ion on ly when the physica l


conditions remain constant w ithin the range of the obs ervat ions .

From the shape of the curve then frequent ly w ith some ,

e x perience a g uess can be made on the probable f orm of the


,

e quat ion w hich may e x press it I n t h is connect ion therefore .


, ,

it is o f t he greatest assistance to be familiar w ith the shapes of


the more common forms of curves by plott ing and studying ,

various forms of e q u ations y=f


By changing the scale in which ob s ervat ions are plotted
the apparent shape of the curve mav be mod ified and it is ,

there f ore desirable in p lott ing to use such a scale that the
average slope of the curve is abo ut 45 deg A much greater or .

m u ch lesser slope sho u ld be avoided s ince it does not show the ,

character of the curve as well .

B . No n-P e ri o d i c C u rve s .

The most c ommon non periodic c u rves are the potent ial
1 44 . -

series the parabol ic and hyperbolic curves and the e x ponenti a l


, ,

and logarith m ic curves .

TH E PO T E N T I AL SE R I E S .

Theoret ically any set of observat ions can be represented


,

e x act ly by a potent ial series of any one o f the f ollow ing forms

y = a0


3
.r I

y +
6

a:
a2

x
2
+
a

x
3

if su ffic ient ly large number o f terms are chosen


a .

For in tance i f 71 correspond ing numerical val u es of x and y


s
,

are given 2 1 yl ; 2 2 yo ;
, , they can be represented
,
E M PI RI C AL C URVE S . 213

by the series when choosing as many terms as requ ired to


give 71 co nstants a
= a o + a 1 x + a2 2 2
y
By substitut ing the corresponding val ues x 1 y1 ; y2 , ,
.

into e quat ion there are obta ined 1 1 equations which de ,

term ine the n constants ao a l a 2 , , ,

Usually however s u ch representat ion is irrational and


, , ,

there fore meaningless and useless .

T AB L E I .

1
- —2x 5

2
6
5

L
98 2 89
76 6 13

Le t , f or instance the first col umn o f Table


,
I represent the
e
volt age
- =x in hundreds of volts and the second column
,
1 00
,

the core loss p y in kilowatts of an 1 25 volt 1 00 h p d irect


, , , ,
- -
.

c u rrent motor S ince seven sets of observations are given


.
,

they can be represented by a potential series w ith seve n con


st ant s thus
, ,

=a ( 6)
y o

and by s ubstit uting the observat ions in and calculat ing the
const ants a f rom the seven equat ions derived in this manner,
there is obtained as emp irical e x pression of the core loss o f
the motor the equation ,

— 2 :r5 —
(7 )
6
y :
+ 2 :z: 1 -
1 :c
-


This e x pression however while e x actly representing ,

the seven observat ions has no physical meaning as easily


, ,

seen by plott ing the individual terms I n Fig 60 y appears . .


,
214 E NGI NE ERI NG M A THE M ATI CS .

as the resultant of a number of large posit ive and negat ive


terms E i rt he rm o re if one of the observat ions is om itted
.
, ,

and the potential series calc u lated from the remaining s ix


values a series reaching up to
,
would be the result thus , ,

5
+ a sa ,

FI G 6 0 . . r
Te m s of E m piri cal E xpre ssion of E x cit ation Powe r .

but the constants a in (8) wou ld have ent irely di fferent numer
ical values from those in t hus showing that the equ ation
( 7 ) h as no rat ional meaning .

1 45 The potent ial series ( 1 ) to ( 4) th us can be u sed to


.

represent an emp irical curve only under the fo llow m g condi


tions :
1 I f the s u ccessive coe ffic ie nts a o a t (1 2
. decrease in
, , ,

val ue so rapidly that w ithin the range of observation the


h igher terms be come rap idly smaller and appear as mere
secondary terms .
21 6 E NGI NE E RI NG M ATH E M ATI CS .

The values of column 4 can now be represented by the same


form of equation nam e lv , ,

y1 00 + 02 10 + 0
?
3 17
3
, (12)
in which the co nstants b o Oz b 3 are calculated by the method, ,

of lea s t squares as describe d in paragraph 1 2 0 o f Chapter I V


, ,

and give
2
y1 0 0 93 27

( 1 3 ) added t o ( 1 1 ) gives the final appro x imate


E quation
equat ion of the torq ue as , ,

yo ( 1 4)
The e quat ion ( 1 4) probably is the appro x imat ion of a
rational e quat ion since the first term ,
represents the ,

bea ring f rict ion ; the second term (w hich is the largest ) , ,

represents the work d one by the fan in mo ving t he air a ,

resistance proportional to the square of the speed and the ,

third term appro x imates the decrease of the air res istance due
t o the churn ing mot ion o f the air created by the fan .

I n general the potential series is of lim ited usefu lness ; it


,

rarely has a rat ional meaning and is mainly used where the ,

curve appro x irnat e lv follows a simp le law as a straight line , ,

to represent by small terms the deviation from this simple law ,

that is the secondary e ffects etc I ts use th u s is often


, ,
.
, ,

temporary giv ing an empirica l appro x imat ion pending the


,

derivation of a more rat ional law .

Th e Pa ra b o li c a nd t h e H pe r y bo l ic C u rve s .

1 46 . o f the most u seful c lasses o f c u rves in engineering


One
ar e those represe nted b v the e q u at ion ,

or , the m ore general e q u ation ,

b —
=a x
y (
E quation ( 1 6 ) di ffers from ( 1 5 ) only by t he co nsta nt terms b
and c; that is it gives a di fferent loc ation to the coordinate
,
E M PI RI C AL C UR VE S . 217

center but the curve shape is the same so that in d is cuss m g


, ,

the general shapes only equ ation ( 1 5 ) need be co nsidered


, .

I f n is positive the curves y = a x n


are parab olic cu rves
, ,

pass ing through the origin and increasing w ith increasing 2 .

If n > 1 y incre ases w ith increasing rapidity if n < 1 yincreases


, , ,

w ith decreasing rapidity .

If the e x pone nt is nega tive , the c u rve s = ax "= n


y
hyperbo lic curves starting from y= 00 for x = 0 and de creasing
, ,

to y 0 = for x = 00 .

m
a 1 gives the straight line through the origin n = 0 and ,

n = 0 0 give respectively straight hori z ontal and vertica l lines


, , .

Fi gs 6 1 to 7 1 give various curve shapes correspo nding t o


. ,

di ff erent values of n .

Parab oli c Cu rves .

Fig 61 n= 2 = 232 the common parabola


. .
y ; .

Fig . 62 . n 4; the biq uadrat i c parabola .

Fig 63 n= 8: = rs
. .
y .

Fig . 64 . n y V5; again the common parabola .

Fig . 65 . n t he biquadrat ic parabola .

Fig . 66 . n

Hype rb oli c C u rve s .

Fig . 67 . n= 1; y the equila teral hyperbola .

Fig . 68 . n=

1
Fig . 69 . n= — 4; y 4
°

13

Fig . 70 . n

Fi g 7 1. . n=
21 8 E NGI NE E RI NG M A TH E M A TI CS .
220 E NG I NE E RI NG M ATH E M A TI CS .

FI G 6 6
. . Parabolic C rve u .

FI G 6 7
. . H yp e rbolic Cu rv e ( E q u ilate ral H ype rbola ) .
y
EIV
I PI RI C A L C URVE S . 22 1

FI G 6 8
. . H ype rbolic Cu rve .
=
y

FI G 6 9. . H yp e rbolic C rve
u .
y
2 22 E N GI N E E RI N G M A TH E M ATI CS .

FI G 7 0
. . H ype rbolic Cu rve .
y

FI G 7 1
. . H yp e rbolic Cu rve .
y
22 4 E NG I NE E RI NG MA TH E M ATI CS .

becomes more di ffic u lt ; that is vario u s val u es o f c and of


,
b
are to be tried to find one w hich gives t he proport ionality .

FI G 7 2
. . araboli c and
P H ype rbolic Cu rve s .
y
=2n
~
.

1 48 . Ta king the logarithm of eq u ation ( 1 5 ) gives


log g = log a + n log 2 ;
E M PI RI CAL C UR VE S . 225

that is a straight line ; hence a p arabolic or h y


,
perbolic cur ve can
,

be recogni zed by plott ing the logarit hm o f g aga ins t the l e gal
r it h m o f x If this gives a straight l ine the c u rve is paraboli c
.
,

or hyperbolic and the slope o f the logarithm ic c u rve tan B= n


, , ,

is the e x ponent .

FI G 7 3
. . arabolic Cu rve s
P .
= xn
y .

This again app lies on ly if the c u rve contain no cons ta nt


term If constant terms e x ist the logarithm ic line is c u rved
.
, .

There f ore by trying di ff erent constants c and b the c u rvat u re


, ,

of the logarithm ic line changes and by interpolat ion su ch ,

constants can be fo u nd which make the logarithmic line straight


, ,

and in this way the constants c and b may be eval uated I f


, .

only one const ant e x ist that is only b or only c the process is
, , ,

re lat ive ly simple b ut it becomes rather complicated w ith both


,
22 6 E N G I NE E RI N G M A TH E M ATI CS .

c ons tants This fac t makes it all the more d esirable to get
.

from the physic al natu re o f th e problem some id ea o n th e


e x i stenc e and the val u e of the c ons tant terms .

Differentiating equation ( 2 0 ) gives :

that is in a paraboli c o r hyperboli c c urve the p erc ent u al


, ,

change or vari ati on of y is n times the p erc entual c hange


, , ,

or variation o f x if n is th e e x ponent
,
.

Here f rom follows :

that is in a parabolic or hyperbolic c u rve the


, ,
rati o o f va ri ati o n,
dv

y
m= is a c ons tant and equ als th e e x p onen t n
E;
.
, ,

3:

Or , inversely
I f i n an empiric al c urve the ratio of variation is c o nstant
the cu rve is — within th e range i n which the ratio o f variation
,

is c onstant a parabo li c o r h yperboli c cur ve w hich has as



,

e x ponent the ratio of variation .

I n the range ho w eve r in which th e ratio of var iation is


, ,

not c ons tant it is not th e e x ponent and w hile the empiric al


, ,

cu rve mi ght be e x pressed as a parabo lic or hyperbolic c u rve


with changing e x p onent ( or c hanging c oefficient ) in this c ase ,

the e x ponent may be very di ff erent from the ratio of v ar ia


tion and the c hange of e x ponent freq u ently is very muc h
,
.

smaller than the change of the ratio of variation .

Thi s rati o o f vari ati on and e x ponent of th e p araboli c or


hyperbolic appro x imation of an empi ric al c u rve m u st not be
mi s taken f or each other as h as occ asionally b e e n d o ne in
,
'

r ed u cing hysteresis curves or radiation cu rves


,
They c oincid e .
2 28 EN GI N EE RIN G IVATH E fi/I A TI C S .

The c u rve y e = +
has t h e same shape e x cept that the
x
, , ,

posit ive and the negative side ( right and le ft ) are interchanged .

I nverted these e quat ions ( 2 1 ) to (2 4) m ay a lso be written


thus ,
nx log

n ( x — c) log

— 3/
nx = lo Co ,
a

;r = — log y;
that is as logarithmic c u rves
,
.

FI G 7 4 E xp one ntial Function = e —x


. . .
y .

1 50 characterist ic o f t he e x ponential f unct ion ( 2 1 ) is


. The ,

that an i ncrease of a: b y a constant term i ncreases ( or in ( 2 3 ) ,

and decreases ) y b y a cons tant f ac tor .

Th u s if an empirical curve y=f ( r ) has such c h aracterist ic


, ,

that
constant for constant,
(1 ,
E MP I RIC AL C UR VE S . 2 29

the curve is an e x ponent ial f unct ion , and the followm g


e q uation may be written

f ( r q)

.

f (x) as

Hereby the e x ponential f u nction can easily be recogni z ed ,


and dist ingu ishe d f rom t h e parabolic c u rve ; in the former a
constant term in the latter a constant factor of x causes a
,

change of y by a constant fac tor .

As res u lt hereof the e xponential curve with negat ive


,

e x ponent vanishes that is becomes negligibly small w i th f ar


, , ,

greater rapidit y than the hyperbolic curve and the e x ponent ia l ,

FI G 7 5
. . H yp e rbolic and E xp o ne ntial Cu rve Com pari son
s .

fu nct ion with positive e x ponent reaches practically infinite


values far more rap idly than the parabolic c u rve This is .

ill ustrated in Fig 7 5 in which are sho w n s uperimposed


"
.
,

the e x ponent ial curve y= e and t he h yperbolic curve


,
z
, ,

which co inc ides w ith the e xp onent ial c u rve


( so l
at x = and at x = 1
0 .

Taking the logarithm of e quatio n ( 2 1 ) gives log g = log a +


nx lo g e that is log y is a linear f u nction of x and plott ing
, , ,

lo g y against x gives a straight li ne This is characteris tic o f .


2 30 ENGINEERING M A TH E MATI CS .

the e x ponential functio ns and a convenient method of recog


,

ni z ing them
.

Ho w ever both of these characterist ics app ly only if x and y


,

contain no cons tant terms With a single e x ponent ial f unct ion
.
,

o nlv the c onstant term o f y needs considerat ion as the cons tant ,

term of a: may be e liminated E q u at ion (22 ) may be writte n


.

thus
_ = ( — )
y b ae
n z c

m n'
s
e

where A= a? “i a c onstant
s .

e x ponent ia l function w hich conta ins a constant term b


Am

wo u ld not give a straight l ine when plott ing log y agains t 33 ,

FI G 7 6
. . x one ntial
E p Fu n ction s .
ENGIN EERING M A T HE M A TI CS

2 32 .

Fig . 76 gi ves the followi ng combinations of e


x
and r

—2 1 »

0 5
—2
x
+0 z
;

35
y 5
:
>


_
f
O5s
.

J”

FI G ; 7 8 . Hype rbolic Functions .


EM PIRI C AL C URVE S . 2 33

—o
Fig . 77 gives the follow ing combi nat ion of e
x
and s
.

(1) y =

(2 ) y =

( 3) = —
01
-
10 “
y
7

3 3
5 6

“x
( 4) y
=

Fig . 78 gives the h vpe r b o lic f unctio ns as combinations


a” and th u s
x
e

= c os h “
y s

= s inh
y )

x
e .

C . E va l u ati o n of E m pi ri c al C u rve s .

1 52 . In
attempt ing to solve the problem of find ing a mathe
m at i ca l e qu at ion y= f (x ) for a series o f observations or tests
, , ,

the correspond ing values of x and y are first tabulate d and


plotted as a curve .

From the nature of the physical problem w hich is re pre ,

sented by the numerical values there are derive d as many ,

data as possible concerning the nat u re o f the c u rve and of t he


fu nct ion which represents it especially at the z ero values and ,

the values at infinity Freq u ently hereby t he e x istence or


.

absence of constant terms in the e quation is indicated .

The log 17 and log yare tab u lated and c u rves p lotted between
23 y log 50 log y and seen whether some of these c u rves is a
, , , , ,

straight line and thereby indicates the e x ponential function or ,

the parabolic or hyperbo lic f unction .

If cross section paper is ava i lable having both coord inates


-
,

d ivided in logarithm ic scale and also cross section paper having,


-

one coord inate divided in logarithmic the other in common ,

sc ale as and y c an be direc tly plotted on these tw o f orms of


,

loga rith m ic c ross— s ec tion paper Us u ally not mu ch is saved .

thereby as fo r the numeric al c al cu lation o f the c onstants the


,

logarithms still have to b e tab u lated .


2 34 ENGINEE RING IlI ATH E M ATI CS .

If ne ither of the fou r c u rves : x y; 33 log y; log x y; log x , , , ,

log y is a straight line and from the physical condition the


,

absence of a constant term is ass u red the fu nction is ne ither ,

an e x ponent ial nor a parabolic or hyperbolic I f a constant .

term is probable or possible curves are plotted between at , ,

b log log ( b ) for various val u es of b and if hereby


y ,
s y , ,

one of the c u rves straighte ns out then by interpolation , , ,

that value of b 1s found which makes one of the cur ves a straight
,

line an d thereby gives the c ur ve law A c onvenient way of


,
.

d oing thi s is : if the c urve with log y (c urve 0 ) is c u rved by angl e


a o ( a o being for instance the a ngl e b etween the tangents at the

two end points of the c u rve or the di ff erence of the s 10 pes at the
,

t w o end points ) use a val u e b , and pl ot the c ur ve with l og


, ,

— b
( y l ) ( c urve an d observe its c urvat ure a t Then inter .

po la t e a val u e b 2 between b 1 and 0 in proportion to the c urva


, ,

t ures a , an d (X0 and plot c urve with l og (y— b g ) ( c ur ve


,
and
again interpolate a val ue b 3 bet w een b 2 and either b 1 or O which ,

ever c urve is nearer in slope t o c urve 2 contin u e until either th e ,

c urve with l og (y— b ) b ecomes a straight line or an S c urve and ,

i n this latter case sho w s tha t the empir ical c urve cannot be
represented in this manner .

I n this w ork logarithmic paper is very useful as it permits


, ,

plotting the c ur ves witho u t first looking u p the logarithms the ,

latter being done only when the last appro x imation of b is


fo u n d I n the same manner i f a constant term is s u spected in
.
,

the x the val u e (ac — c) is used an d c urves plotted for various


,

values of c Freq u entl y the e x istenc e and the character o f a


.

constant term is indicated by the shape o f the cur ve ; for


instance i f one of the c urves plotted between x y log 33 l og y
, , , ,

approaches straightness for high or for l ow values of the a h ,

s ci ss as b u t c u rves considerabl y at the other end


,
a c onstant ,

term may be suspected w hich becomes l ess appreciable at one


,

end of the range For instance the e ff ect o f the c onstant c in


.
,

— c decreases with increase of x


( c
c ) .

Sometimes one of the curves may be a straight line at one


end but curve at the other end This may indicate the presence
,
.

of a term which vanishes for a part of the obs ervat ions I n


,
.

this case only the observat ions of the range which gives a
straight line are used for deriving the c u rve law the curve ,

calc ulated there from and then the d ifference between th e


,

calculated c u rve an d the observations further investigated .


E NGI NE E RING M A TH E M A TI CS
i

2 36 .

Fig 7 9 .I n the thir d and four th column o f Table I II are


.

g iven log 6 and log t I n Fig 7 9 then are pl otted log 6 i as


. .
, ,

curve H ; e log 13 as curve I I I ; l og e log 15 as cur ve I V


, , , , .

As seen fro m Fig 7 9 curve I V is a straight line that is


.
, ,

=lo
2 .

f ge
log 13

ve sti gation of Volt-am pe re Characte ri stic of te n a


"
FI G . 79 . In T ungs L mp
ilam e nt
F .

log or i i = A + n lo e = ae
g ; ,

which is a p ar abolic cu rve .

The constants a and 71 may now be calc ulated from .

the n u merical d ata of Tabl e I I I by th e me thod of least


sq u ares as d iscussed in Chapter I V p aragraph 1 2 0
, Whi le , .

this meth o d gives the most acc ur ate res ults i t is s o laborious ,

as to be seldom used in engineering ; generally val u es o f the ,

constants a and n su fficiently accurate for mo st practical


,
E M P I RI CAL C UR VE S . 237

pur poses are derived by the s o called E A method w hich


,
-
, ,

w ith proper tab u lar arrangement of the n umerical values gi ves ,

high acc u racy with a minim u m of work .

T AB L E I II .

VOLT-AMPE RE CH ARACTE RI STI C OF 1 1 8 VOLT TU NGS TE N L AMP - .

+ 0 8 10 g
. e

+ 0 -0 0 3

a vg .
00 3
4 7 per
. c e nt

4-
42 2

-
6


1 3 5 51 = f
£ 95 4

lo g 0 lo g 6 a nd

The fourteen sets of observations are divided into two


groups of seven each and t h e sums of log 6 and log i formed
, .

They are indicated as 2 7 in Tabl e I II .

Then subtracting the two groups 2 7 from each other ,

e liminates A and divi ding the two di fferences A gives the


, ,

e x ponent n = O 60 1 1 ; this is so near to


,
. that it is reasonable
to assume that n = o 6 and this value then is u sed
.
, .

No w the sum of all the val u es o f log 6 is formed given as ,

2 1 4 in Table I I I and multiplied with n


,
and t he p roduc t
2 38 E NGI N E E RI NG M ATH E M ATI CS

subtracted from the s um o f all the log 75 . The di ff erence A


then eq uals 1 4A and divided by 1 4 g ives , , ,

A = lo g a= 8 21 1 ;
.

hence a = 0 0 1 62 5 and the volt ampere characteristic of this


,
.
,
-

t ungsten lamp th u s f ollow s the eq u at ion ,

log i = 8 2 1 1 . +O6 . log 6 ;

From e and i can be derived the po w er inp u t = ei and the


p ,

e
re s 1 st a nce r
l
=
p

and , e li m inating e f ro m these two e xpressio ns gives ,

5 4
p 0 0 1 62 5 1 1 1 35 r
.
4

that is t he po w er inp u t varies w ith the fourth po wer of t he


,

resistance .

As suming t h e resistance r as proport ional to the absol ute

"
temperature T and considering that the power input int o the
,

lamp is radiated from it that is is the power of radiat ion P , , T

the e quation between p and r also is the e quation between P ,

and T thus , ,

P, = 1m

that is the radiat ion is propo rt ional to the fourth power of the
,

absolute temperat u re This is the law o f blac k body radiation .


,

and above e quat ion o f the volt —ampere characterist ic of the


tungsten lamp t h u s appears as a conc lusion from the radiat ion
law t h at is as a rational e qu ation
, ,
.

1 54 E x a m pl e 2
. I n a magnet ite a re at constant arc length
.
, ,

the voltage cons u med by the are e is observed for di fferent , ,



val u es o f c u rrent i To find the e q u ati on of the volt ampere
.
-

characteristi c o f the magnetite arc


240 E NGI NE E RI NG M ATH E M ATI CS .

Proceed ing in Table I V in the same manner w ith log i


and log (e — 30) as was do ne in Table III w ith log 6 and log t ,

gives
n= and a = 90 4 ; .

FI G 8 0
. . ve sti gation of Volt am pe re Charact e ri stic of Magne tite
In - Arc .

hence
log (e 30 ) log 5;

e = 30 +
E M PI RI C AL C URVE S . 2 41

which is the e qu ation o f the magnetite arc volt ampere charac -

t e rist ic .

1 55 . The change of current res u lt ing from a


E x am pl e 3 .

change o f the cond itions o f an e lectric c irc u it containing resist


ance ind uctance and capac ity is re corded by osc illograph and
, ,

gives the c ur ve reprod u ced as I in Fig From this c u rve .

FI G 8 1 ve sti gatio n of Curve of Curre nt Change


In in E le ctric Circ u it

"
. .
.

are taken the n u merical val ues tab u lated a s t and 13 in the first
two col umns o f Table V I n the third and fourth col umns are
.

given log t and log 0 and c u rves then plotted in the u sual
,

manner Of these curves only the one between t and log ?


.

is shown as I I in Fig 8 1 since it gives a straight line for the


, .
,

higher val ues o f t For the higher val ues of t there f ore
.
, ,

log i = A — nt ; or ,

that is it is an e x ponent ial f unct ion


, .
2 42 E NGI NE E R I NG M A TH E MATI CS .

T AB L E V .

TR ANS I E NT C URRE NT C H ARAC T E RI S T I C S .

t 1. l gt
o lo g z 2 A

0 0
+00 2

02 0

0-
8 0 -90 3 0 -
30 1 0 -8 8 -9 54

0 0

22 —o s .

A=
>< m lo g e - lo g a2

41 a2= 2 85
.

lo g i z = 0 4 00 — 3 0 4 t lo g
. . e

= — 3 -8 4t
a ,= 4 94. i2 2 055 .

lo g .
69 3 1 0 7 t lo g
. 0

21 =
°
—1 07 t
4 9 4e
~
.

i c = 4 94¢ — 3 8 4t
85
°

. . 6

To calc u late the const ants a and n the range o f val u es is ,

us ed in which the curve I I is straight ; that is from t = 1 2


, , .

to t = 3 As these are five observat ions they are grouped in two


.
,

pai rs the first 3 and the last 2 and then f or t and log one
, , ,

third o f the sum of the first 3 and one half o f the sum o f the ,
-

last 2 are taken Subtract ing this gives


.
, ,

At = 1 1 5 ; A log i = .

S ince , however the eq u at ion , ,


when logarit hm at e d ,
g1 ve s
2 44 E NGI NE E RI N G M A TH E M ATI CS .

1 56 . As further e x ample may be considered the resolutio n


of the core loss c urve o f an ele ctric motor which had been ,

e x pressed irrationally by a potential series in paragraph 1 44


and Table I .

T AB L E VI .

CORE LO SS CU R VE .

A _l g } )

P 1; k w . lo g P 1; lo g 6
f 1 6lg o 6

2 563 7 -2 3 6 0 -7 0
2 -8 4 5 7 2 89 1 -34 + 0 -0 2
a vg . + 0 -0 6
3 -20 0 7 277 7 2 82 3 -03 —0-
03
3 326 7 -2 6 8 4 -0 5 — 0 -1 2
3 4 34 7 -8 60 5 20 + 1 02
3 526 7 -40 8 84 3 + 2 -1 6

£ 3 P2= 1 in w a t ts

1 -0 7 1

A= O -2 785
0 44 5
1 59 0 2 1 0
0 -2 7 8 5

The first two columns o f Table VI give the observed val ues
of the voltage e and the core loss P in kilowatts The ne x t , .

two columns give log 0 and log P i Plotting the curves shows .

that log e log P is appro x imately a straight line as seen in


, , ,

Fig 82 with the e x ception o f the two highest points of the


.
,

curve .

E x cluding therefore the last two points the first five o b se r ,

vat io ns give a parabolic curve .

The e x ponent o f this c u rve is found by Table VI as


n = 1 598 ; that is with s u fficient appro ximation as n = 1 6
.
, , . .

To see how far the observations agree with the curve as ,

give n by the e quati o n ,

3
1 ,

in the fifth c olumn log 6 is recorded and in the si xt h column , ,

A = log a = lo g P , log e As seen t h e first and the last .


,

t wo val u es of A di ff er from the rest The first val u e corre .


E M PI RI CAL C UR VE S . 2 45

sp e nds to s u ch a lo w val u e of P as to lo w er the acc u racy o f


,

the observation Averaging then the fo ur m iddle val ues


.
,

A henc e ,

log P i = log e ,

Pi = 1 . 9 14e l 6'
in watts
°

FI G 82
. . In ve sti gation of Ou vre s .

This equation is calc u lated as P and plotted i n Fig 82


, C, . .

The observed val u es of P are marked by circles As seen


, .
,

the agreement is satis factory with the e x ception of the t w o


,

highest values at w hich apparently an additional loss appears


, ,

which does not e x ist at lo w er voltages This loss probably is .

d ue to eddy c u rrents ca u sed by the increas ing magnet ic stray


field res ulting f rom magnetic sat uration .
2 46 E N G I NE E RI NG MA TH E MA TI CS .

1 57 .
m
As a f u rther
may be c o n e x am
sid ered the resol u tion e

of t he m agnetic characteristic plotted as c ur ve I in Fig 83 ,


.
,

and given in the first two colum ns o f Table VI I as H and B .

T AB L E VI I .

MAGNE TI C CH AR ACTE R I S TI C .

B
kilo li nes
lo g H lo g B

“ 0 -1

=b

I B= 26 3 1 5 -0 2 0
2 63 X 0 33 4
-

4 = 1 -6 8 6
0 -2 1 1 = a

2 1 1 + 0 0 5 07
.
H and B=
0 2 1 1 + 0 0 50 7
. .
H

Plott ing H B log H log B against each other leads to no


, , ,

resu lts n e ither does the introduct ion of a constant term do


,

this Thus in the fi fth and s i xt h columns o f Table VI I are


.

calcu lated
5 and 1
5
and are plotted against H and against
, B .

H
Of these fo u r curves only the curve o f against H is shown
B ,

in Fig 83 as 1 1 This c urve is a straight line w ith the e x cept ion


.
,
.

of the lowest values ; that is

a +b H .
48 E N G I NE E RI N G J I A TH E MA TI CS .

Th e d ifferenc e between the observed valu es of


g ,
and the
value give n by above e qu ation which is appreciable u p to ,

H 6 c o u ld now be fu rther i nvestigated and w ould be fou nd


, ,

t o appro x i mately follow an e x ponential law .

As a final e x ample may be c onsid ere d the investigation of


a hysteresis c u rve of silic on steel of which the nu meric al valu es ,

are given i n c olumns 1 and 2 of Table V I II .

The first c ol u m n gives the magnetic d ens ity B i n li nes o f ,

magnetic forc e per e m ; the sec ond c olu mn the hysteresis loss
f

w i n ergs per cycle per kg ( specific d ensity


,
Th e third
.

c olu mn gives log B and the fo u rth c olumn log w


, .

Of the f our c u rves bet w een B w log B log w o nly the , , , ,

c u rve relating log w to logB appro x imates a straight li ne and ,

is given i n the upper part of Fig 84 Thi s cu rve is not a . .

straight li ne througho u t its enti re length bu t o nly two sections ,

of it are straight from B = 50 to B = 400 and f rom B = 1 600 to


, ,

B = 8000 but the cu rve bends between 500 and 1 2 00 and above
, ,

8000 .

Thu s tw o empirical formulas o f the form : w = aB n are , ,

c alculated i n the usual mann er in Table VI I I


,
The one , .

appli es f or lo w er d ensities the other fo r med i u m dens ities


,

-1 1
L ow d e nsity : B 400 : w = 0 003 4l
.

Medi u m d ensi ty : 1 600 s B s 8000 : w 60


-

In Table VI I I the val u es for the lower range are d enoted


by the i nd e x 1 for the higher range by the inde x 2
,
.

Neither of these empiri c al formulas appli es strictly to the


range : 400 < B < 1 600 and to the range B > 8000 They may
, .

be applied within these ranges by ass u ming either the c oe fficient,

a as varyi ng or the e x ponent n as v a rying that is applying a


, , ,

c orrection fac tor to a or to n ,


.

Thu s i n the range : 400 < B < 1 600 the loss may be repro
, ,

sented by :
( 1 ) An e x tension of the low d ens ity f ormula :
2 11
w a1B w = 0 0034 a
°

. 1.

( 2 ) An e x tension of the medi u m d ensity f orm u la


= l 6
w a qB '
or w = G 1 096 B
.
E M P I RI C A L C UR VE S . 2 49

by giving tables or c u rves of a respectively 22 S uch tables are


.

most c onveniently given as a perc entage c o rrec tion .

FI G 84
. .

The pe rcentage c orrec tion w hic h is to be applied to a l and


,

respectively to nl and n2 to make the formul as applic able


, ,
2 50 E N GI N E E RI N G MATH E M ATI CS .

T AB L E V I II .

H YS T E RE SI S O F SI L I C O N S TE E L .

4 4m
1 1

lo g B lo g w a2 m 1 22

1 5 44 + 0 4 0

1 50
245
3 94
571
1 02 5

1 01 0
2 320 — o s7.

40 30
0 1 50
0 0 00

1 60 0
2 00 0 2 1 00 0
2 500 3 03 00
3 0 00 4 0 500

4 00 0 0 3 4 00
50 00 9 0 0 00
60 0 0 1 2 0 00 0
8 0 00 1 941 0 0

1 0 0 00 2 02 5 00
1 200 0 3 9 7 500
1 4000 00 9 5 0 0
1 60 0 0 90 7 5 0 0 5 9 50
zs=
£ s = 1 1 9 02
.

4=

_
5
2 11

I .o= 2 1 44 1 .

4 lo g w = 7 5 3 3 + 2 1 1 10 0 8
. .

10 2.
w = 0 0 0 34 I B
.

a .= 0 0 03 41
.

4= 2 5 02
.

2 = 1 599
.

1 60 2
.

1 60 X 2 8 3 6 2
. .

' lo g m = e 0 40 + 1 00 lo g
. . B
c
t
1 on
o4 0 = lo g a: w = 0 1 09 6 3
.

H0 9 6
2 52 E N GI NE E RI NG M ATH E M ATI CS .

n log B = log w log a o;

n log B log w— log a o


no no lo g B no log B

log w— log ao —n log B


0

72 0 10 g B

by these equations ( 1 ) and ( 2 ) the c orrection factors in c olumns


5 to 8 o f Table V I I I are c alculated by using f or a o and no the
,

values o f the lo w e r range c u rve i n c olumns 5 and 7 and the


, ,

val u es of th e medi u m range cu rve i n c ol u mns 6 and 8


, .

Thus for instanc e at B =1 000 the loss c an be c alculated


, , ,

by the equation ,

w = alB

b y applying to al the c o rrection f ac tor

per cent at c onstant : n1 that is ,

a1 = 0 003 4 1 ( 1
.

or by applying to nl the c orrection factor

per c ent at c onstant : al that is ,

n1 = 2 . 1 1 (1

Or the loss c an be c alc u lated by the equation

"
,

w 5 CLQ B
' 2
,

by applying to the c o rrection factor :

per cent at c onstant : n2 = 1 . 60, that is ,

az
E M PI RI C AL C URVE S . 2 53

or by applying to nz the c orrection factor :


per cent at c onstant : az= 0 1 09 6 ,
. that is ,

nz 1 1

and the loss may th u s be given by either of the fou r ex

pressions
-08 6

As seen , the variation o f the e x ponent requ i red to e x tend 72,

the u se o f the parabolic eq u ation into the range f or which i t


d oes not stric tly apply any more is m u ch less than the varia ,

tion o f the c oe fficient a and a fa r greater acc u racy is thus ,

secured by c onsidering the e x ponent n as c ons tan t for


medi u m and high values o f B and making the correction in
c oe fficient a outside o f the range where the 1 6 th power law
,
.

holds rigidly .

I n the last c ol u mn of Table V I II is rec orded the rati o of


A log w
vari ation 772 ,
as the averages each of tw o successive
A log B
val u es As seen 772 agrees w ith the e x ponent n within the
.
,

tw o ranges where it is c onstant b u t di ff ers f rom it o u tside


, , ;

o f these ranges For instanc e if B changes f rom 1 600 down


.
,

wa rd the ratio o f variation m inc reases w hile the e xponent


, ,

nslightly d ec reases .

I n Fi g 84 are show n the pe rc entage c orrection of the


.

c oefficients a l and a 2 and also the tw o e x ponents nl and 722


, ,

t ogether with the ratio o f variation 772 .

The ratio o f va riation 772 is v ery use f u l i n c alc u lating the


change of loss resulting f rom a s mall change o f magnetic d ensity ,

as the perc entual change o f loss w is m times the percentual


( small ) change o f d ensi ty .

As f u rthe r e x ample the read e r may red u ce to empirical


,

e quations the series of obse rvations given i n Table I X This .

table gives
A The c andl e po w er L as f u nc tion of th e po w er i n
.

.
-
put p, ,

of a 40—watt tungsten filament incandesc ent lamp .

B The loss of powe r by c o rona ( discharge into the ai r) p


.
, ,

i n kw i n .
,
km of c ond u ctor as f u nction o f the voltage
.
,

0 ( in kv ) betwee n c onducto r and return c ond uc tor


. for the ,
2 54 E N GI NE E RI N G M A THE M ATI CS .

TAB L E IX .

A . L um i no s i t y c h a ra c t e ri s t i c of 40
-w at t t un gt s en i nc a nd e s ce nt la m p .

L h o ri zo n t a l c a nd e p o l w e r.

p w a t ts i np u t .

B . C o r o na lo ss o f g
h i h -v o t a l g e t r a ns m i s s i o n li ne ; at 00 c yc e s l
1 89 5 In l g th
en o f co nd u c t o r .

3 10 m . d i s t a nc e b e t w e e n c o nd u c t o rs .

No 0 0 0 . s e v e n-s t r and c a b l e, cm . d i am e t e r .

"1 3 °
C 76 2 . cm . b ar o m e t e r ; s un s h i ne .

e = k i lo v o lt s b e t w e e n co nd u c t o r s , e fi e c t iv e
'
.

p ki lo w a t ts lo s s .

ob 0 1
1 81 . 0 221 0 .
8 70
. 6 44
.

o b 00 1 86 . 2 2 2 7 -0 1 0 66 219 0 8 31
o -
. . .

a
0 1 92 6 1 3 -2 5
-
f 1 .

o 0
0 a

f 20 0 . 6 1 89 0 .
2 10
.

o fie 0
1

20 8 . 6 1 95 0 .

o i n00 21 6 . 0 2 03 8 .

C . Vo l um e -p re s s u r e c h ara c t e ri s t i c o f d ry s t e am at i t s b o iling —
p o i nt .

t = d e gr e e s C .

P = p r e ss ur e , i n k g p e r c m . .
2

V= vo lu m e , i n m 3 p e r kg . .

0 61 2
.

0 3 79
.
0 1 67
.

0 319
.
256 E NGI N E E RI NG M A TH E M A TI CS .

( G) .
(The phase di ff erences here are re f erred to the ma x imum
o f the f un damental : with waves o f di ff erent freq u encies the ,

phase di ff erenc es natur ally change f rom point to point and in ,

speaking of phase di fference the reference point on the wave


,

Th i r d H arm onic

FI G 85
. . E ff e ct of S m all T hird H ar m o nic .

must th u s be given For instan ce in C the third harmonic is


.
,

in phase with the f undamental at the ma x im um point of the


latter but i n O pposition at its z ero point )
,
.

The equati ons o f these waves are :


A: y= 1 00 c os B ,

C : y= 1 00 cos {1 + 1 0 cos
E : y= 1 00 cos cos (3fi+ 9 0 deg ) .

G: y= 1 0 0 cos fi+ 1 0 cos (3 B+ 1 80 deg )


,
.

1 00 c os fl — 1 0 c os
E M P I RI CAL C UR VE S . 2 57

y
= 1 00 cos fi+ 3 0 c os 3 5
y = 1 00 cos
fl + 3 0 c os ( 38 4 -
45 deg ) .

y= 1 00 cos
fi + 3 0 cos (3 04 -
90 deg ) .

y = 1 00 cos fl + 30 cos deg )


y = 1 00 cos
fi + 30 cos deg ) .

— —
1 00 c os fl 30 cos 38

Th i r d Harm onic

FI G 86
. . E ff e ct of L arg e T hird H ar m o nic .

In all these waves one cycle O f the tri ple harmonic is given in
,

dotted lines to indicate its relative position and intensity and


, ,

the ma x ima o f the harm onics are indicated by the arrows .


2 58 E NGI N E E RI NG M A T H E M A TI CS .

As seen with the harm onic in phase or in o pposition ( C and


,

G ) the waves are symm etrical ; with the harmoni c ou t of phase


,

the waves are unsymmetrical of the s o called saw tooth ,


-
“ ,

type and the saw tooth is on the rising side of the wav e with a
,

lagging on the decreasing side with a leading triple harmonic


, .

Th i r d H ar m oni c Fla t Ze r o Re v e rsa l

FI G 87
. . lat Ze ro
F a nd Re v e rs al b y T hird H arm o nic .

The latter are sho wn in I) E F ; the former have the sam e shape
, ,

b ut reversed that is rising and decreasing side of the wave


, ,

interchanged and there f ore are n ot shown


,
.

The tri ple harmonic in phase w ith the f u ndamental C gives , ,

a pea ked wa ve with flat ze ro and the peak and the flat z ero
,
2 60 E NGI NE E RI N G M A TH E M A TI CS .

At 0,the w ave is entir el y h ori z ontal at th e z ero that is remains , ,

z ero for an appreciabl e time at the reversal I n t h is fig ur e the


.
,

three harmoni cs are sho w n separately in dotted lines in their ,

relative intensities .

A tri pl e harmoni c in opposition t o the f u ndamental (Figs .

85 and 86 C ) is c haracteri z ed by a flat t o p an d steep z ero an d ,

Fi f t hH ar m o ni c

'
E fi e ct Of L arg e Fi fth H arm o n ic .

w ith the increase of the th ird harmonic the flat t op devel ops ,

into a do u bl e peak (Fig 86 C ) w hi le steepness at the point o f


.
,

reversal increases .

The si m ple s aw to oth prod u c ed by a tri ple harmoni c in


,

q u adrat ur e w ith the f u ndam ental is sho w n in Fig SSE With . .

increasing tri ple harmoni c the h u mp of the saw tooth bec omes
,
E M PI RI CAL C UR VE S . 261

more prono u nced and c hanges to a secon d an d l ower peak as ,

sho w n in Fig 86 This figur e gives the variation of the saw


. .

tooth shape f rom 45 to 45 deg phase di ff erence : With the phase.

o f the third harmoni c s hi fting from i n phase to 45 deg l ead the -


.
,

fl at z ero by m o ving u p on the wave has f ormed a h u m p or saw


, ,

tooth low down on the decreasing ( and with 45 deg lag on the .

increasing ) side of the w ave At 90 deg l ead t h e saw tooth has


. .
,

moved up to the middl e of the down branch o f the wave and ,

with 1 3 5 deg lead has moved still further up forming pr ac t i


.
, ,

,
8
cally a second lo wer peak With 0 deg lead or opposition
1 — . .

o f phase— the h u m p of the saw tooth has moved u p t o the


top and formed the second peak— or the flat top with a lower
, ,

third harmonic as in Fig 85 G


,
. .

Figs 88 and 89 give the e ff ect of the fifth harmonic super


.
,

imposed o n the f u n damental of 5 per cent in Fig 88 and of 2 0 ,


.
,

per cent in Fig 89 Again A gives the f u ndamental sine w ave


. .
,

C the e ff ect of the fi fth harmoni c in opposition w ith the f u nda


mental E in q u a d rat ure (lagging ) and G i n phase One cycl e
,
.

of the fifth harmonic is sho wn in dotted lines and the max ima ,

of the harmo ni cs indicated by the arrows .

The equations of these waves are given b y

s = 1 00 cos 3
y
g y = 1 00 c os 3 — 5 c os 5 3

a y = 1 00 cos 3 — 5 cos deg ) .

g = 1 00 c os c os 5 3
y

s = 1 00 cos 3
y
a = 1 00 cos 3 20 cos 5 3
y
w .
y=
— 1 00 cos 3 20 cos deg ) .

G : y= 1 00 cos c os 5 3

In the distortion caused by the fif th harmoni c (in opposi


tion to the f u ndamental ) flat top (Fig 880 ) or do u ble peak (at .

higher val u es of the harmonic Fig is accompanied by flat,


.

z ero ( or at very high val u es o f the fi f th harmonic doubl e rever


, ,

sal at the z ero similar as in Fig 87 d) while i n the dist ortion


,
.
,

by the t hird harmonic it is accompani ed by sharp z ero .

With the fi f th harmonic in phase w ith the f undamental a ,

peaked wave res ults with steep z ero Fig 88G an d the transi ,
.
,
2 62 E NGI N E E RI NG M A TH E M A TI CS .

tion from the steep z ero to the peak with larger val ues o f th e
,

fifth harmoni c then devel ops into two additional peaks thus
, ,

givi ng a trebl e peaked wave Fig 88G with steep z ero The
,
.
, .

be ginning of trebl e peakedn ess is n otic eabl e already i n


88 G with only 5 per c ent of fi f th harmonic
,
.

Th i r d H ar m oni c
t o
fi f h H arm ni c

T hird a nd F ifth H arm o n ic .

With the seventh harmoni c the trebl e peaked wave would


,
-

be acc om panied by fiat z ero and a q u a d r u pl e peaked wave


,
-

wo uld give steep z ero (Fig .

The fi f th harmonic o u t o f phase with the f un damental again


gives saw tooth waves Figs 88 and 89 E b u t the saw tooth
-
,
.
,
2 64 E NGI N E E RI NG M A TH E M ATI CS .

While characteristic of th e e ffect of the fifth h arm onic is


C oin cidenc e of peak with stee p z ero or trebl e peak of flat
, ,

to p or do u ble peak with flat z ero or d ou ble reversal and do u bl e ,

saw tooth
-
.

Th i rd H arm o nic
Fi f t h H arm oni c

FI G 9 2
. . T hird a nd i th
Ff H ar m o nic .

Bythus combining thir d and fi f th harmoni cs o f proper


val ues they can be made to ne utrali z e each other s e ff ect in any
,

one of their characteristics but then accentuate each other in


,

the other characteristic .

Th u s peak and flat z ero o f the tri ple harmonic combined w ith
peak and st eep z ero of the fifth harmoni c gives a peaked wave ,

with n ormal sin u soidal appearanc e at the z ero val u e ; c om b in


EM PI RI CAL C UR VE S . 2 65

ing the flat tops or do u bl e peaks o f both harmonics the flat ,

z ero o f the on e neutrali z es the steep z ero o f the other an d we ,

get a flat top or do u bl e peak wi th n ormal z ero Or by com .

bining the peak o f the t hi rd harmonic with the flat top of the
fifth we get a w ave w ith n ormal t op b u t steep z ero an d w e get a , ,

wave with n ormal top b u t flat z ero or do u ble reversal by com


, ,

bining the triple harmonic peak wi th the fi fth harmonic flat top .

Th u s any of the characteristics can be prod u c ed separately


by the combination o f the thir d and fi fth harm o ni c .

B y combining third and fi f th harmonics o ut o f phase w ith


f unda e nt al such as give single or do u ble saw tooth shapes the
m — -
,

vario u s other s aw tooth shapes are prod u ced and still f ur ther
-
,

saw tooth shapes by combining a symmetrical (in phase or i n


-
,

opposition ) thir d harmonic with an o ut of phase fi f th or ,

inversel y .

These shapes produc ed by the s u perposition under di fferent ,

phase angles of fi f th and third harmonics on the f u ndamental


, ,

and their grad u al change into each other by the shifting in


phase of one o f the harmonics are shown in Figs 90 9 1 and 9 2 ,
.
,

for a third harm onic O f 1 0 per cent and a fi f th harmoni c of 5 ,

per cent of the fundamental .

I n Fig 9 0 the third harmonic is in phase in Fig 9 1 in q u adra


.
,
.

ture lagging and in Fig 92 in opposition w ith the f undamental


,
. .

A gives the fundamental B the f u ndamental w ith the third har ,

monic only and C D E F the w aves resulting from the s u per


, , , , ,

position o f the fi f th harmo ni c on the combination of funda


mental and third harmoni c given as B I n C the fi f th har monic ,
.

is in opposition in D in quadrat ure lagging in E in phase and


, , ,

in F in q u adrature leading .

We see here ro u nd tops with flat z ero (Fig nearly .

triangular waves (Fig 90E ) appro x imate half circles (Fig


.
,
.

9 2 E ) sine waves with a dent at the top ( Fig 9 2 C ) and vari


,
.
,

o us di fferent forms of saw tooth .

The eq u ations of these waves are :

s = 1 00 c os 3
y
a = 1 00 c os 3 + 1 0 cos 33
y
g y
= 1 00 c os c os 33 — 5 cos 53
a y = 1 00 c os c os 33 — 5 c os deg ) .

s = 1 00 cos 3 + 1 0 c os c os 5 3
y
2 66 E NGI N E E RI NG M A TH E M A TI CS .

= 1 00 c os
y 3
— 10
y
= 1 00 c os 3 c os deg .
)
= 1 00 c os — 10 c os 5 3
y 3 cos deg .
) —5

= 1 00 — 10
y c os 3 c os deg .
) — 5 c os
= 1 00 — 10
y cos 3 cos deg .
)+5 c os 5 3
= 1 00 c os — 10 c os deg ) +5 cos
y 3 .

s = 1 00 cos
y 3
w y = 1 00 c os —
3 10 c os 33
c = 1 00 c os — 10 c os 33 — 5 c os 5 3
y 3
b y= 1 00 cos 3 10 -
cos 33 — 5 c os deg ) .

m y
= 1 00 cos 3 — 10 cos c os 5 3

E VE N H ARM O N I c s .

1 61 . o f the wave —shape dis tortion of even har


Characteristic
monies is that the wave is not a symm etrical wave b u t the ,

two half waves have di fferent shapes and the characteristics ,

of the negative hal f w ave are O pposite to those of the positive .

This is to be e x pected as an even harmo ni c whi ch is in phas e


, ,

with the positive hal f wave of the f undamental is in O pposition ,

with the negative ; when l eading in the positive it is lagging ,

in the negative and inversel y ,


.

Fig 93 shows the e ffect of a sec on d harmoni c of 30 per c ent


.

o f the f u ndamental A s u perim posed in q u adrat ure 60 deg


, ,
.

phase displ ac ement 3 0 deg displ acement and in phase in


,
.
,

B C D and E r espectively
, ,
.

The equations of these waves are :


A: y= 1 0 0 c os 3 an d y = 3 0 c os (2 3 — 90 ) ’

B : y= 1 00 — 90
c os c os ( 3
2 )
— 60
C : y= 1 00 cos c os ( 3
2 )
— 30
D : y= 1 00 c os cos ( 3
2 )
E : y= 1 00 cos c os 23

Q uadrat ur e c ombination (Fig 93B ) gives a wave where the .

risi ng side is flat the decreasing side stee p an d inversel y with


, ,

the other hal f wave C and D give a peaked w ave for the on e a
.
,

saw tooth for the other hal f wave an d E c oincidenc e of phas e , ,

of f undamental an d secon d harmonic gives a c om bination of ,

on e peaked hal f w ave with on e flat top or double peaked wave - -


.
2 68 E NGI N E E R I NG M A TH E M A T I CS .

of a wave in w hich the on e hal f wave is a sh ort high peak the


,
,

other a l ong flat top by the s u perposition o f a second harmoni c


,

of per cent and a third harmoni c of 1 0 per cent both in


,

phase with the f u ndamental .

A gives the f undamental sine wave B and C the second and ,

third harmonic D th e combin ation of f u ndamental an d second


,

S e c o nd Th i rd Har m o ni c

FI G 9 4
. . Pe ak a nd lat T
F op b y Se co nd a nd T hird Harm o nic .

h armo ni c giving a do u ble peaked negative hal f wave and E the


,
,

addition o f the third h armoni c to the wave D Thereby the .

do u bl e peak of the negative hal f wave is fiat t e d to a l ong flat


top and the peak o f the positive half wave intensified an d
,

shortened so that the positive ma x im u m is abo ut two and one


,
~
E M P I RI CAL C UR VE S .

half times the negative ma x im u m and the n egative half wave ,

nearly 7 5 per cent longer than the positive half wave .

The eq u ations of these waves are given by :

A: y= 1 00 c os 3
B: y cos 2 3
C : y= 1 0 cos 33
D : y= 1 00 cos cos 23
E : y= 1 00 c os c os c os 33

H I G H HA R M O N I C S .

1 62Comparing the effect of the fi f th harmonic , Figs 88 and


. .

89 , with that O f the t h ird harmonic , Figs 85 and 8 6, it is seen .

1 0 if S e v e nt h Harm o n i c

E ff e ct Of Se v nth H ar m o nic
e .

that a fi f th harmonic even if very small is f ar easier distin


, ,

g u is h e d that
,
is merges less,
int o the f u ndamental than the third

harmonic S till m ore t his is the case with the seventh har
.

monic as shown in Fig 95 in phase and in O pposition of 1 0 per


,
.
,

cent intensity Thi s is to be e xpected : sine waves which do not


.

di ffer very m u ch in freq u ency s u ch as the f u ndamental and


,

the secon d or third h arm onic merge i nto each other and form a
,

res u ltant shape a distorted wave of characteristic appearance


, ,
2 70 E NGI N E E RI NG M A TH E M A TI CS .

whil e sine w aves o f very di ff erent freq u encies as th e fund am e n ,

tal and its el eventh harmoni c in Fig 9 6 when s u perimposed ,


.
, ,

remain distinct from each other ; the general shape of the w ave
is the f undamental sine and the high harm onics appear as ri p
,

pl es u pon the fundamental th us gi vi ng what may be called a ,

corr ugated sine wave B y co u nting the n u mber o f ripples per


.

FI G 96 . . W ave i n w hi ch E le ve nth H arm oni e Pre dom i na te s .

compl ete wave or per hal f wave the order of the harmo ni c
, ,

can then rapidly be determined For instan ce the wave sho w n .


,

in Fig 96 contains mainl y the eleventh harmonic as there are


.
,

eleven ripples per wave The wave sho w n by the oscillogram .

Fig 97 shows the twenty third harm onic etc


.
-
,

FI G 9 7. . C D 2 35 1 0 . Alt e rnator W av e w i t h S i ng l e H i g h H arm o nic .

Very f requently high harm onics appear in pairs o f nearly the


same freq u enc y and intensity as an eleventh and a thirteenth ,

harmonic etc I n this case the ri pples in the wave shape sh o w


,
.
,

ma x ima w here the t w o harmonics c oincide and n odes w here


, , ,

the t w o harmonics are in opposition The presence o f n odes .

makes the c o u nting o f the n u m ber o f ri pples per c om plete w ave


more d i fficult A c onvenient meth od o f proc ed ure in this c ase
.
272 E NGI N E E RI NG M ATHE M ATI CS .

72
t wo freq u encies : f —l— and f
g and as one of these f req u encie s
is the impressed engine f req u ency thi s a ffords a check
,
.

Wh ere the two high harmonics of nearly eq u al order as the ,

el eventh and the thir teenth in Fig 9 8 are appro x imately eq u al


.
,

in intensi t y at the n odes the ripples practically disappear


, ,

and between the nodes the ri pples give a freq u en cy intermediate


between the two c omponents : Appar ently the twel fth harm onic
in Fig 98 I n this case the two c onstituents are e asily deter
. .

mined : 1 2 — 1 = 1 1 an d ,

Where of the two c onstituent s one is greater than th e other


the wave still sh ows n odes but the rippl es d o n ot entirel y d is a p
, ,

pear at the n odes b u t merel y decrease that is the wave sh ow


, , ,

a sine with ri ppl es which increase an d decrease al ong the waves

FI G . 1 00 . C D 2 35 1 1 . Al te rnat or W av e w ith T W O Ne arl y E q u al H i g h


H ar m o n ic s .

as shown by the oscillograms 9 9 and 1 00 I n this case one o f .


,

the t w o high freq u encies is given by c o u nting the total number


of ripples b u t it may at first be in do u bt wh ether the other
, ,

c om ponent is higher or lo w er by the number of nodes The .

decision then is made by c onsidering the length O f the ri ppl e at


the n ode : I f the l ength is a ma x im u m at the n ode the secondary ,

harm onic is of higher frequen cy than the predominating on e ;


i f th e length o f the ri pple at the n ode is a minim u m the sec ond ,

ary freq u ency is l ower than the predominating one This is .

ill u strated in Fig 1 0 1 I n this fig u re A an d B re present th e


. .
,

tenth and twel fth h ar m Onic of a wave respectively ; C gives ,

their s u perposition with the lo w er harmonic A predominating ,

w hile B is only of hal f the intensity of A D gives the s u pe r pos i


.

tion o f A and B at eq u al intensity and E gives the s u per ,

position with the higher frequency B predominating That is .


,

the respective equations would be :


E M P I RI CAL C UR VE S . 273

A : y= cos 1 03
B : y= eos 1 2 3
C : y= eos c os 1 23
D : y= eos 1 03 + cos 1 23
E : y= 0 5 c os 1 03 + c os 1 2 3
.

As seen in C the hal f wave at the node is abnormally long


, ,

showing the preponderance of the l ower frequency in ,


E ab h or
S u pe r pos it i on o f Hig h Har m oni cs

FI G 1 0 1
. . Su pe p r o itio n of T w Hi g h
s o H arm o n ic of Vario u s I nt e ns iti e s
s .

mally short s h owing the preponderance O f the higher f req u ency


,
.

I n alternating c ur rent and voltage waves the appearance o f


-
,

two s u cc e ssive h igh harmonics is q u ite f requent For instance


'

if an alternating c ur rent generator c ontains 72 sl ots per pol e ,

this prod u ces i n the voltage w ave the t w o harm onics o f orders
E NGI N E E R I NG M A T H E M A T I CS .

2 72 — 1 and 2 a + 1 S uch is the origin o f the harmonics in the


.

oscillo grams Figs 99 and 1 00 . .

The nature o f the increase and decrease of the ripples and the
formati on of the n odes by the s u perposition of two a dj acent
high harmonics i s best seen by c ombining their e x pressions trig
o no m e tr i cally .

Thus the harmonics :

y, ( 2n= c os

a nd yz (
= c os 2 n + 1
)3
c ombined give the res ultant :
y = y + y2 1
—c os ( 2 n c os (2 n
—2
c os 3 c os 2 n3

that is give a w ave of frequency 2 n times the fundamental :


,

c os 2 n3 but w hich is n ot c onstant b u t varies in intensity


, ,

by the factor 2 c os 3 .

No t in f req u ently wave shape distorti ons are met which are
-
,

not due t o higher h arm onics o f th e fundamental wave but are ,

inc ommens u rable there w ith I n this case there are two entir ely
.

unrelated f req u encies This for instance occ urs i n the sec ond
.
, ,

ary c ircuit of the single phase inducti on m otor ; two sets o f


-

c urrents of the f req u encies f an d ( 2f — f ) e x ist ( where f is the


, , ,

primary freq u ency and f the frequency of sli p ) Of this natur e


, .
,

freq u ently is the distortion produced by s ur ges oscillations


, , ,

arcing grounds et c in electric circui ts ; it is a combination of


,
.
,

the natural freq u ency o f the circ ui t with the impressed fre
q u eney Tel ephonic currents comm only show such m ultiple
.

freq uencies whi ch are n ot h arm oni cs of each other


, .
2 76 E NGI NE E RI NG M A T HE M A TI CS .

The ne x t then in c ases where c onsiderable nu merical c ale u


,

lati o ns are required is the metho d of c alc u lation ,


The most .

c onveni ent one us u ally is the arrangement i n tabular form .

As e x ample c onsider the problem of c al c u lating the regu la


,

tion o f a —volt transmission line o f r = 60 ohms resist


,

anc e x = 1 3 5 ohms ind u ctive reactance and b = 0 00 1 2 c ond e n


, ,
.

sive su sceptance f or various values of non inductive ind u c tive


,
-
, ,

and c ond e nsive load .

S tarting wit h the c omplete equations of the long distanc e -

transmission line as given i n Theory and Calc u lati on of


,

Transient E lectri c Phenomena and Oscillations Section I I I , ,

paragraph 9 and c onsidering that f or every one o f the vario u s


,

power factors lag and lead a su fficient number of values


-
, , ,

have to be calculated to gi ve a curve the amount of work ,

appears hopelessly large .

However witho u t loss of engineering e x actness the e q ua


, ,

tion of the transmission line can be simplified by appro x im a


tion as discussed in Chapte r V paragraph 123 to the form
, , , ,

E 1 = E0 1+ ‘

l Z IO

1 +

I1 :
10 1+ + YE O 1 +

where E 0 ,
Io are voltage and current respectively at the step ,

down end ,
E 1 1 1 at the st ep —u p end of the line ; and
,

'

Z r+ jx 60 1 3 53 is the total line impedance ;


Y = g + jb + 0 00 1 2 j . is the total shunted line admittance .

Herefrom follow the n u merical val ues

5 15
2 2
1 + 0 0 36j
.

1 00 27
N UM E RI CAL CAL C UL A TI ONS .

Z
l
1

y 1 + 0 00 1 2 j)
. + o o1 2 j).

0 00 1 1 68j
.
-
l

hence s u bstit uting in


,
the f o llow ing eq u ations may be
w ritten

A+8 ;
1
'

0 03 03) E o =
E1 =
(0 9 1 9 (2 )
. .

11 C— D .
)
_
1 65 o
Now the work of calc u lating a series of n u merical
val ues 1 s cont in u ed in tab u lar f orm as f ollows : ,

1 . 1 00 P E R C EN T PO W E R -
FAC T O R .

E 0 = 60 kv . a t s t e p -d o w n e nd o f li ne .

A + 2 2 j kv
. .

D= 1 1 68 1 ) E 0 1 0 = am
°

3
. 7o l i
.
p .

E1 = + e
I o am p . B kv . e1 2 + 0 22 = e z .
— = t an e .

0 2 2j. 3 03 6 +
- 3 1 58 + l
2 6j.
l 4 Bi ‘ '

5 3j. 329 5 +
58 5 + 1 0 1j
. .
+ 0 l 73
.

4 5 + 1 0 5j
. . 5 9 6 + 1 2 7j
. .

+ 14 2
°

5 7 + l 3 2j
. . 41 + 0 2 53
6 8 + 1 5 8j
. . 6 1 9 + 1 8 0j
. .

t am
C am p .
42

— 90 1 49 1 5
j .

+ 90 9
20 1 7 5 + 70 8j
. . 50 1 3 + 30 6 = 5 3 1 9
40 36 8 + 1 4 j
. . 3 5 9 + 7 1 5j
. . 64 0 1
60 55 l + 2 2 j
. . 54 2 4-72 3 5 8165
80 73 5+ 2 9j
. .
+73 0j .

00 91 9 + 3 6j
. .
+ 0 811
20 1 10 3 + 4 3j
. .
+ 74 4 ] + 0 68 0 4
2 78 E NGI NE E RI NG M A TH E M A TI CS .

ci = 6o kv . a t s t e p - u p e nd o f li ne .

Re d . Fa c t o r ,

am p .
Po w e r - Fa c t o r .

C u rve s of 20 , c
o,
21 . cos 0, p l o t t ed i n Fi g . 86 .

2 . 90 PE R CE N T PO W E R F AC T O R -
,
LAG .

c os sin
0- j s in 0)

E1
(0 9 1 9 + ( 1 00 5 + A+B ’

1 1 08310 0 1 0 3
-

I1 (0 9 1 9 +
'
e D .

and now the table is calculated i n the same manner as under 1 .

Then correspo nding tables are calculated in the s ame ,

manner for power facto r


,
and respectively lag
-
, , ,

and for power factor - lead ; that is for ,

cos 0 + j s in -
0 6j ;
.

0 9+
O S + 0 6j ;
. .

O7+ .

The n curves are plotted for all seve n values of p owe r factor -
,

from lag to lead .

From these curves for a number of value s of 20 for ins tance


, , ,

40 60 80 1 00 numeric a l values of 21 ea cos 0 a re


, , , , , , ,
280 E NGI N E E RI NG M A TH E M A TI CS .

( a) E stimation of the magn itude of an e ffect ; that 13 ,

determ ining appro x imate numerical val u es w ithin 2 5 50 or , ,

1 0 0 per ce nt Very f requ ently s u ch very ro u gh appro x imation


.

is su fficient and is all that can be e x pected or calc u lated


,
.

For instance when invest igat ing the short c irc u it current o f an
,
-

electric generat ing system it is of importance t o know whether ,

this current is 3 or 4 t imes normal c u rrent or whether it is ,

40 to 50 times normal current but it is immaterial whether ,

it is 45 to 46 or 50 t imes normal I n st u dying lightn ing .

phenomena and in general abnormal voltages in electri c


, , ,

systems calculat i ng the di scharge capac ity of lightn ing arres


,

ters etc the magnit u de o f the qu ant ity is often su fficient In


,
.
,
.

calc u lat ing the crit ical speed of tur bine alternators or t he ,

natural period of osc illat ion of synchrono u s machines the ,

same applies since it is o f importance o nly to see that these


,

speeds are s u fficient ly remote f rom the normal operating speed


to give no tro u ble in operat ion .

( b ) Appro x imate calc u lat ion re quiring an accuracy of one ,

or a few per cent only ; a large part of engineering cale u


lat ions fall in this class especi a lly calc u lati ons in the rea lm o f
,

des i gn Altho u gh freq u ently a h igh er acc u racy co u l d be


.
, ,

reached in the calc u lation proper it wo u ld be o f no val u e , ,

s ince the data on which the cal c u lat ions are based are s us
ce pt i b le to variations beyon d control due to vari ation in the ,

material in the mechanical dimensions etc


, ,
.

Th us for instance the e xc iting c u rrent of i nduct ion motors


, ,

may vary by several per cent due to variations of the lengt h ,

of a ir gap so small as to be beyon d the lim its o f constr u ct ive


,

acc u racy and a calc u lat ion e x act to a fraction o f o ne per cent
, ,

w hile theoret ically possi b le th u s wo u ld be pract ically u se less


, ,

The calc u lation of the ampere —t u rns requ ire d f or the sh unt
fie ld e x c itat ion or for the series field of a dire ct c u rrent
,
-

generator needs on ly moderate e x actness as variat ions in the ,

magnet ic material in the spee d re gulat ion of the driving


,

power etc prod u ce di fferences amo u nt ing to severa l per


,
.
,
»

cent .

(0 ) E x act engineering calc u lat ions as for instance the , , ,

calc u latio ns o f the e fficiency o f apparat us the re gu lat ion o f ,

transforme rs the characterist ic c u rves of ind u ct ion motors


, ,

etc These are determ ined w ith an acc u racy f req uent ly amount
.

i ng to one tenth of one per cent and even greater


-
.
N UM E RI CAL CAL C UL A TI ONS . 2 81

e x act e ngi ne erm g calc u lat io ns the accurac y


E ven for m os t ,

of th e slide rule is
u s u ally su fficient i f intelli gently used t hat , ,

is used so as to get the greatest accuracy For accurate caleu


,
.

lations pre f erably the glass slid e sho u ld not be used b u t the
, ,

res u lt interpolated by the eye .

Thereby an acc u racy w ithin 4 pe r cent c an easily be main


t aine d .

Fo r most engineering c alc u lations logarithmic tables are ,

su fficient for th ree decimals i f intelligently u sed and as s u ch


, ,

tables c an be c ontained on a single page their use makes th e ,

c alc u lation ve ry m u ch mo re expeditio u s than tables o f more


d ec imals The same applies to trigonometric tables : tables
.

of the trigonometric functions ( not their logarithms ) of three


d ecimals I find most c onveni ent for most c ases given f rom ,

degree to d egree and u sing d ecimal f ractions of the d egrees


,

( not mi nu tes and


E x pedition in engineering c alc u lati ons t h u s requires the u se
of tools of no higher acc u racy than requ ired i n the result and ,

such are the slid e rules and the three d ecimal logarithmic and
,

trigonometric t ables The u se of these h ow eve r make i t


.
, ,

necce ss a ry to g u ard in the c alc u lation against a loss of accu rac y .

Su ch loss o f accu racy occ u rs i n s u btracting o r dividing tw o


terms w hi ch are nearly equ al in som e logarithmic operations , ,

sol u tion o f equ ation etc and i n s u ch c ases either a higher


, ,
.

acc u racy o f calc u lation m u st be employed — seven d ecimal


logarithmic tables e tc o r the O peration w hich lowe rs the
,
.
~

acc u racy avoid ed The latte r c an us u ally be d one For


,
. .

ins tance in divid ing 2 97 by 2 83 by the slide rule the proper


, ,

way is to divid e 2 97 by 283 and add the resu lt ,

to 1 .

I t is in the methods of c alc u lation that e xpe rienc e and j udg


ment and skill in e fficiency o f arrangement o f nu meric al c ale u
lati o ns is m ost marked .

1 67 While the c alc u lations are u nsatis f actory if not carried


.
,

out with the d egree o f exac tness w hich is f easible and d esirable ,

it is equ ally w rong to give nu merical val u es w ith a nu mber o f

T hi s O bvious ly do e s no t app ly to so m e cla e of e ngi n ri ng w ork ss s ee ,


in
w hi ch a m ch hi gh e r acc racy of tri g o no m e tric f u nctio ns i s r e qui r e d
u u
, as

tri go nom e tric surv e yi ng e t c , .


2 82 E NGI N E E RI NG M A TH E M A TI CS .

c iphe rs greater than the method or the p urpose of the calcu la


t ion w arrants For ins tance i f in the des ign o f a d irect cur rent
.
,
-

generator the calc u lated fie ld ampere t u rns are given as 97 38


,
-
,

such a numeri ca l val ue destroys the confidence i n the work o f


the calculator or designer as it implies an acc u racy greater
,

t han poss ible and there by sho ws a lac k of j u dgment


,
.

The number o f ciphers in whi ch the result o f calc ulation is


given should signify the e x actness I n this respect t wo
.

systems are in use :


( a ) Num erical values are given w ith one more decimal
than warranted by the probabl e error of the re su lt ; that is ,

the dec imal be f ore the last is correct but t he l ast dec imal may
,

be wrong by several u nits This method is us ually employed


.

in astronomy physics etc


, ,
.

(b ) Num erical val ues are given with as many dec i ma ls as


the acc u racy O f the ca lc u lation warrants ; that is the last ,

dec imal is probably correct within half a unit For ins tance .
,

an e ffici ency of 86 per cent means an e ffic iency between


and pe r cent ; an e ffic iency of per cent means an -

e ffic iency between and per cent etc This system ,


.

is generally used in engineering calculat ions To get accuracy .

of the last dec imal o f the res u lt the calc ulat ions then must
,

b e carried o ut for one more decimal t han given i n the res u lt .

For ins tance when calculat ing the e ffic iency by adding the
,

vario us percentages of losses data l ike the following may b e


,

g i ven :

Total
E ffic i e nc v 1 00
Appro x i mate ly

is obvio us that t hro u gho ut the same calculat io n the


It
same degree of acc u racy mus t be observed .

I t follows herefrom that the val u es

2 4;
2 84 E NGI N E E RI NG M A THE AI A TI C S .

po ints near the desired val ue so as to get at least a short piece


,

o f curve i nc luding the desired point .

The main advantage and f oremost p u rpose of c ur ve plott ing


,

thu s is to show the shape of the f u nct ion and thereby give ,

a c learer concept ion of it ;


b ut for recording n u merical
val ues and deriving n u mer
,

ical val ues from it the plotted


,

c u rve is in f erior t o the table ,

d ue t o the limited acc u racy


possible in a plotted c u rve ,

and the f u rther inacc u racy


res u lt ing when draw i ng a
curve thro u gh the p lotted cal
c ulat e d po ints . To some
e x tent the nume rical val ues
,

as taken f rom a plotted c u rve ,

depend on the partic u lar


kind of c u rve ru le used in FI G 1 03 Com p ound ing Cur ve
. , .

p lotting the c u rve .

I n general c u rves are used f or t w o di ff erent purposes and


, ,

on the p urpos e for which the curve is plotted sho u ld depend ,

the method of plotting as the scale the z ero val ues etc
, , ,
.

When c u rves are used to


illustrate the shape of the
f unct ion so as to show how ,

m u ch and in what manner a


q u antity varies as function
of another large divisions of ,

incons pic u ous cross sect ion -

ing are desirable but it is ,

essential that the cross


sectioning sho u ld e xtend to
the z ero val ues o f the func
t ion even if the numeri cal
,

val ues do not e xtend so


FI G 1 04
. . Com p ounding Cu rve far S ince O therwise a wrong
. ,

impression wo u ld be con
f erred As ill ustrations are plotted in Figs 1 03 and 1 04 the
. .
,

compo u nding c u rve o f a direct c u rrent generator The arrange


-
.
N UM E RI CAL C ALC UL A TI ONS . 2 85

ment in Fig 1 03 is correct ; it shows t h e relative variation


.

of voltage as f u nction o f the load Fi g 1 04 in which the


. .
,

cross sectioni ng does not be gin at t h e scal e z ero con f ers the
-
,

FI G 1 0 5
. . Cu rve Plotte d to s how Cha racte ri stic S hape .

FI G 1 06. . Cu rve Plot t e d for U se as D e s i gn D ata .

wrong impression that the variation o f voltage is far greater


than it really is .

When c u rves are used to record n u me ri cal val u es and


derive them from the c u rve as f or i nstance is commonly the
, , ,
2 86 E NGI NE E RI NG M A TH E M A TI CS .

c ase with magneti zation curves it is unnecessary to have the z ero ,

o f the f unction co inc ide with the z ero of the cross sec t ioning but -
,

rather pre ferable not to have i t so i f thereby a better sc ale of ,

the cur ve can b e sec u red I t is desirable however to us e su ffi


.
, ,

c i e nt lysmall cross secti oning to make it possible to take num er


-

ical values f rom the curve with good accura cy This is ill u s .

tra t e d by Figs 1 05 an d 1 06
. B oth sh ow the magnetic charac .

teris t ic of so ft steel f or the range a b ove B


, 8000 in which it is ,

usually employed Fig 1 0 5 shows the proper way O f plotting for


. .

showing the shape of the function Fig 1 06 the proper way o f ,


.

plotting for use of the cur ve to derive num erical values there from .

FI G 1 0 7
. . a
S m e Fun ctio n Plotte d to D i ffe re nt S cale s; I is corre ct .

1 69 . Curves
shou ld be p lotted in such a manner as to sho w
the q uant ity which they represent and its variat ion as well as , ,

possible Two features are des irable h e re for


.

1 To use such a scale that the average slope o f the curve


.
,

or at least of the more important part o f it does not di ffer

"
,

much from 45 deg H ereby variat ions of c u rvature are best


.

show n To i llustrate this the e xp onential f u nct ion y= c


.
,
is c

pl otted in thr ee diff erent scales as c u rves 1 II II I in Fig 1 0 7 , , , , . .

C ur ve I has the proper scale .

2 To us e such a scale that the t otal range of ordinates is


.
,

not m u ch di fferent from the total range of abscissas Th u s .

when plotting the power factor of an ind u ctio n motor i n -


,

Fig 1 08 curve I is pre f erabl e to c ur ves II or I I I


.
, .
28 8 E N G I N E E RI N G M A TH E M A TI CS .

pa p er b u t not log 23 or (P or logarith m ic paper u sed etc as


, , , , .
,

this w o u ld not sho w the shape o f the relation U s in g


f or instan ce se m i logarithmic pa p er that is w ith lo ga ri t h
-
, ,

mi c absciss ae and ordinary ordinates the plotted c u rve w o u ld ,

sho w the shape o f the relation y=f ( lo g etc The u se o f .

lo g arithmic p a p er or the u se o f y or x/E as coordinate etc is


,
2
, ,
.
,

j u stified only w here the p u rpose is to sho w the relation bet w een
y and lo g 23 or bet w een y and x or bet w een y and et c as is
2
.
, , ,

the case w hen inve s ti g atin g the e qu ation o f an e m p ir cal c u rve ,

or w hen intendin g to sho w some p artic u lar f eat u re o f the relation


y f Th u s f or instance w hen p lottin g the p ow er p cons u med
by corona in a hi g h p otential transm ission line as f u n ction o f the ,

line volta g e e by u sin g x/ f) as ordinate a strai g ht line res ults


, ,
.

Also w here so m e p art i c u lar f u nction o f one o f the coordinates ,

as lo g 23 gives a more rational relation it m a y be u sed instead


, ,

o f it Th u s f or instan ce in radiation c u rves or w hen ex p res sin g


.
,

velocity a s f u nction O f w ave len g th or f re qu enc y or ex p re ss in g ,

atten u ation o f a w ireless w ave etc the lo g o f w ave len g th or


,
.
,

f req u en cy that is the g eo m etric s ca le ( a s u s ed in the theory o f


, ,

so u nd w ith the octave as u nit ) is more rational and there f ore


,

pre f erable .

S ometi m es the val u es o f a relationshi p extend over s u ch a


w ide ran g e as to make it i m p o ss ible to re p re s ent all o f the m in
one c u rve and then a n u m ber o f c u rves m ay have to be u s ed
, ,

w ith di ff erent scales I n s u ch ca s es the lo g arith mi c scale O f ten


.
,

b rin g s all val u es w ithin one c u rve w itho u t i m pro p erly cro w din g ,

and es p ecially w here the p u r p o s e o f c u rve plottin g is not so


m u ch to sho w the s ha p e o f the relation as to record f or the p u r ,

p ose o f takin g n u m erical val u es f ro m the c u rve the latter ar ,

ran g ement that is the u se o f lo g arithmi c or semi lo g arith m ic


, ,
-

p a p er may be de s ira b le Th u s the mag neti c characteristi c o f


.

iron i s u sed over a ran g e o f field inten s ities f ro m very fe w a m


p ere t u rns per c m in tran s f or m ers
. to tho us ands o f a m p ere ,

t ur ns in tooth den s itie s o f rail w ay m otor s and the m a g neti c


, ,

characteristic th u s is either re p resented by three c u rves w ith


di ff erent scales o f ratio s 1 ,
as sho w n in Fi g 1 09 or .
,

the lo g o f field intensity u sed as absci s s as that i s se m i lo g arit h , ,


-

mic p a p er w ith lo garith m ic scale as ab s ciss ae and re g ular scale


, ,

as ordinate s as sho w n in Fi g 1 1 0
,
. .

I t m u st be reali z ed that the lo g arith m ic or g eometrical s c a le


N UM E RI CAL CAL C UL A TI O NS . 89

FI G 1 09
. .

FI G 1 1 0
. .
2 90 E NG I NE E RI NG M ATH E M ATI C S .


in eq u al divisions re p re s ent not eq u al val u es o f the
whi ch
u antity b u t eq u al f ractions o f the u antity — is some w hat
q , q
less easy to read than co m mon scale H o w ever as it is the same .
,

scale as the slide r u le thi s is n ot a serio u s obj ection


,
.

A disadvanta g e o f the lo g arith m ic scale is that it cannot


extend do w n to z ero and relations in w hich the entire range
,

do w n to z ero req ui res consideration th u s are not well s ui ted ,

f or the u se o f lo g arith m i c scale .

1 71 Any en g ineerin g calc u lation on w hich it is w orth


.

w hile to devote any ti m e is w orth being recorded w ith s u ffi,

cient com p leteness to be g enerally intelli g ible Very o f ten in .

makin g calc u lations the data on w hi ch the calc u lation is bas ed ,

the s u bj ect and the p u r p ose o f the calc u lation are g iven i nco m
p le t e ly or not at all since they are f a m iliar to the calc u lator a t
,

the time o f calc u lation The calc u lation th u s w o u ld be u nin


.

t e llig ib le to any other en g ineer and u s u ally becomes uni nt e lli


,

g i b le even to the calc u lator in a fe w w eeks .

I n addition to the na m e and the date all calc u lations sho u ld ,

be acco m p anied by a complete record o f the O bj ect and p ur pose


o f the calc u lation the apparat u s the ass u mptions made the
, , ,

data u sed re f erence to other calc u lations or data employed


, ,

et c in short they sho u ld incl u de all the in f ormation req ui red


.
, ,

to make the calc u lation intelli g ible to another en gineer w itho u t


f u rther in f orm ation besides that contained in the calc u lations ,

or in the re f eren ces g iven therein The small a m o u nt o f ti me .

and w ork req u ired to do this is ne g li g ible compared with th e


i ncreased u tility o f the calc u lation .

Tables and c u rves belon g in g to the cal c u lation sho u ld in


the same w ay be co m p letely identified w ith it and contain
s u fficient data to be intelli g ible .

1 7 1 A E n g ineering investi g ations evidently are of n o val u e


.
,

u nle s s they can be co m m u nicated to those to w hom they are o f

interest Th u s the en g ineering re p ort is an essential and i m


.

portant p art o f the w ork I f there f ore occasionally an en g ineer


.

w ork that he hates to


,
w aste
,

or scientist i s met w h o is so m u ch interested in the investi g atin g
the time o f m aki n g p roper and

com p lete re p orts this is a very f oolish attit u de since in general


, ,

it destroys the val u e o f the w ork .

As p ractically every en g ineerin g investi g ation is O f interest


and i m portance to di ff erent classes o f people as a r u l e not one , ,
2 92 E NG I NE E RI NG MATH E M ATI CS .

( c) The general re p ort This sho u ld g ive the results that is


.
, ,

ex p lain w hat the matter is abo u t in p lain and practically non ,

techni cal langu ag e addressed to laymen that is non -en gineers


, , , .

I n other w ords it sho u ld be u nderstood by any intelligent non


,

technical man .

S u ch g eneral report w o u ld be materially shorter than the


g eneral en g ineering report as it w o u ld omit all details and , ,

merely deal with the g eneral problem p u rpose and sol u tion , .

I n g eneral it is advisable to combine all three reports by


, ,

havin g the scientific record preceded by the general engineerin g


rep ort and the latter p receded by the g eneral re p ort R o u g hly
, .
,

the general report w o u ld u su ally have a len g th o f 20 to 40 p e r


cent o f the g eneral en g ineerin g report the latter a le ngth of 1 0 ,

to 2 5 p e r cent o f the co m p lete scienti fic record .

The bearin g o f the three classes o f reports may be u nderstood


by ill u stration on an investi g ation w hi ch appears o f commercial
u tility and there f ore is s u b m itted f or ind u strial develo p m ent to
,

a man ufact u ring corp oration ; the financial and general adm inis
t ra t i ve p o w ers o f the cor p orations to w ho m the investi g ation is
,

s u bmitted w o u ld read the g eneral re p ort and i f the matter


,

appears to them o f s u ffi cient interest f or f u rther consideration ,

re f er it to the en g ineerin g de p artm ent The general re p ort th u s .

m u st be w ritten f or and intelli g ible to the non -e n g ineerin g


,

ad m inistrative heads o f the or g ani z ation The administrative .

en gineers O f the en g ineering departm ent then per u se the g eneral


en g ineerin g report and this report t h u st m u st be an en g ineerin g
,

report b u t g eneral and not req u ire the kno w led g e o f the s p ecialist
,

in the p artic u lar field I f then the concl u sion derived by the
.

ad m inistrative en g ineers f rom the readin g o f the g eneral en gineer


i ng re p ort i s to the e ff ect that the matter is w orth f u rther con
sideration then they re f er it to the specialists in the field covered
,

by the investi g ation and to the latter finally the scientific record
,

o f the investi g ation a p peals and is st u died in makin g final report


on the w ork .

I nversely w here nothing b u t a lengthy scientific report is


,

s u bmitted as a r u le i t w ill be re f erred to the en g ineerin g de p ar t


,

ment and the general engineer even i f he co u ld w ade thro u gh


, ,

the len g thy re p ort rarely has i m mediately t im e to do so th u s


, ,


lays it aside to st u dy sometime at his leis u re and very O f ten
this time never co m es and the entire matter drops f or lack of
, ,

p roper repre s entat i on .


N UME RI C AL CAL C UL ATI ONS . 2 93

Th u s it is o f the u t m ost i m portance f or the en g ineer and th e


s cienti s t to be able to p resent the res u lts o f his w ork no t only by
,

elaborate and len g thy scientific report b u t also by report o f ,

moderate len g th intelli g ible w itho u t di ffic u lty to the g eneral


,

en gineer and by short statement intel li g ible to the non -en gineer
, .

d . R e li ab ility of Num e ri cal C al cu l ati ons .

1 72 The mo s t i m p ortant and essential req u irement o f


.

n u merical en g ineerin g calc u lations is their absol u te reliability .

W hen makin g a calc u lation the most brilliant ability theo , ,

r e t i ca l kno w led g e and practical experien ce o f an en g ineer are

made u seless and even w orse than u sele ss by a sin g le error in


, ,

an im p ortant calc u lation .

R eliability o f the n u merical calc u lation is o f vastly g reater


i m p ortan ce in en g ineerin g than in any other field I n p u re .

mathe m atics an error in the n um erical calc u lation o f an


e x am p le w hich ill u strates a g eneral p ro p osition does not detract ,

f ro m the interest and val u e o f the latter w hi ch is the main ,

p u r p ose ; in physics the g eneral la w w hich is the s u bj ect o f


,

the investi g ation re m ains tru e and the investi gation o f interest ,

and u se even i f in the n um erical i ll u stration o f t he la w an


,

error is made With the most bril liant en g ineerin g desi g n


.
,

ho w ever i f in the n u merical calc u lation o f a sin g le str u ct u ral


,

m e m ber an error has been made and its stren g th thereby cal e u ,

lated w ron g the rotor o f the machine flies to p ieces by ce ntri f ugal
,

f orces or the bridge colla p ses and w ith it the re p u tation o f the
, ,

en g ineer The essential di ff erence bet w een engineerin g and


.

p u rely scientific c a clu la t i o ns is the ra p id check on the correct


ness o f the calc u lation w hich is u s u ally a fforded by the pe r
,

fo rm a nce o f the calc u lated str u ct u re— b u t too late to correct


errors .

Th u s ra pidity of calc u lation w hile by itsel f u sef u l is o f n o , ,

val ue w hatever com p ared w ith reli ability— that is correctness ,


.

O ne o f the fir s t and m ost im p ortant req u ireme nts to sec u re


reliability is neatness and care in the e x ec u tion o f the calc u la
tion If the calc u lation is m ade on any ki nd o f a sheet o f
.

p a per w ith lead p encil w ith f req u ent striki n g o u t and correct
, ,

i ng o f fig u res etc it is p ractically hopeless to e x pe ct correct


,
.
,

re s u lts f ro m any m ore extensive calc ulations Th us the w or k .


2 9 3a E NGI N E E RI NG MATH E MATI CS .

sho u ld be done w ith pen and ink on w hite r u led pa per ; if ,

chan g es have to be made they sho u ld pre f erably be made by


,

erasin g and not by striking o u t I n general the appearance o f


,
.
,

the w ork is one o f the best indications o f its reliability The .

arran g e ment in tab u lar f orm w here a series o f val u es are cal e u
,

lated Off ers considerable assistance in improving the reliability


,
.

1 73 E ssential in all extensive calc u lations is a complete


.

system o f checking the res u lts to ins ur e correctness ,


.

O ne w a y i s to have the same calc u lati on made independently


by t wo di ff erent calc u lators and then compare the resu lts ,
.

Another w a y is to have a fe w points o f the calc u lation checked


by so m ebody el s e Neither w ay is satis f actory as it is not
.
,

alw ays p os s ible f or an en g ineer to have the assistance O f another


en g ineer to check his w ork and besides this an engineer sho u ld
, ,

and m u s t be able to make n u meri cal calc u lations so that he can


absol u tely rely on their correctness witho u t somebody el s e
assistin g him .

I n any more i m p ortant calc u lations every o peration th u s


sho u ld be per f ormed t w ice pre f erably in a di fferent ma nner
, .

Th u s w hen m u lti p lying or dividing by the slide r u le the m u lti


, ,

plication or division sho u ld be repeated mentally a ppro x i ,

mately as check ; w hen adding a col u mn o f fig u res it sho uld be


, ,

added first do w n w ard then as check u p w ard etc


, , .

W here an exact calc u lation is req ui red first t h e magnit u de ,

o f the q u antity sho u ld be estimated i f not already kno w n , ,

then an a pp roxi m ate calc u lation made w hich can f req u ently ,

be done m entally and then the exact calc u lation ; or inversely


, , ,

a f ter the exact calc u lation the res ul t may b e checked by an


,

a pp roximate m ental calc u lation .

W here a series o f val u es is to be calc u lated it is advisable ,

first to calc u late a fe w individ u al points and then entirely , ,

independently calc u late in tab u lar f orm the series o f val u es


, ,

and then u se the previo u sly calc u lated val u es as check O r .


,

inversely a f ter calc u latin g the series o f val u es a fe w points


,

sho u ld inde pendently be calc u lated as check .

W hen a series o f val u es is cal c u lated it is u s u ally easier to ,

sec u re reliability than w hen calc u lating a sin g le val u e sin ce ,

in the f ormer case the di ff erent val u es check each other There .

f ore i t is al w ays advisable to calc u late a n u mber o f val u es ,

that is a short c u rve branch even i f only a sin g le point is


, ,
APPE ND I X A .

NOT E S O N T H E T HE O R Y O F FU NCTI O NS .

A G e n e ra l Fu n c ti o ns
. .

1 75 . The most general algebraic Ox pression of powers of


2: and y,

.
)y

)
y
" =
(1)
is the i mplici t f u nc ti on I t relates y and at so t ha t to
a nalytic .

eve ry value of 2: there correspond n val ues of y and to every ,

value of y there correspond m values of x if m is the e x ponent ,

o f the h ighest power of a: in

As suming e x press ion ( 1 ) solved for y (which us u ally cannot


be carried out in final form as it requires the solution of an ,

equa tion of the nt h order in y with coe fficients which are ,

e x pressio ns of x) the explici t ana lyti c fu ncti on


, ,

y (2 )
is obtained I nversel y solv ing the imp li cit fun ction ( 1 ) for
.
,

at that is from the e xp lic it f unction


, ,
e x press in g 2: as
funct io n o f y gives the reverse fu ncti on o f
,
that is
x (3 )
the general algebraic f unct ion in its implicit f orm
In ,

or the e x pli cit f orm or the reverse f unc tion x and y


are ass u med as general num bers ; that is as comple x qua n ,

ti ties ; thus ,

x= zrl + s ;

=
u yi + 79 2 .

and likew ise are the coe fficients coo, am


APPE N D I X A . 2 95

I f all
the coe fficients a are real and x is real the corre , ,

s po nd ingn val ues of y are e ither real or pairs o f conj u gate ,

comple x imaginary q uantities : yl + jy2 and yl jyg .

1 76 For n = 1 the implicit f u nct ion ( I ) solved f or y gives


.
, , ,

the rati onal fu ncti on ,

2
a oo + a o 1 r + a og x
-
11 1 1 1 23

and if in this f u nct ion (5 ) the denominator contains no a


s, the
i nteger fu nc ti on ,

=a + a 2 25 2
+ am x m
y o .
,

is O btained .

For n = 2 t h e implicit fu nction ( 1 ) can be solved for y as a


q u adratic eq u at ion and thereby gives ,

a l lx + a 1 2x
2 ““
l kt "
v
y a ns:
2

tha t is t h e e x p licit f orm (2 ) o f equ ation ( 1 ) contains in this


,

case a square root .

For n > 2 the e x plic it form y= f (2 ) e ither becomes ve ry


,

complicated for n = 3 and n = 4 or cannot be produced in


, ,

finite form as it re qu ires the sol ution o f an eq u ation of more


,

than the fourth o rder Nevertheless y is still a f u nction of .


,

x and can as such be calc u lated by approx imation etc


, ,
.

To find the valu e yl which by f u nct ion ( 1 ) corresponds t o ,

x = xl

,
Taylo r s theorem o ffers a rapid appro x imation Sub .

s t it ut ing 2 1 in f u nction ( 1 ) gives an e x pression which is of

the nt h order in y thus : and the problem now is to


,

find a val u e y] which makes F =0 .


,

Ho w ever ,

h?
F( r l , y1 ) + 11 "
l“
dy d y 2 2
.

wh ere b = yl — y is the di ff erence between the correct val u e y;


and any chosen val u e y .
2 96 E NGI N E E RI NG M A TH E M A TI C S .

Neglecting the higher orders of the small q u antity


a nd cons idering that F( x l y1 ) = 0 _
,
gi ves ,

F( I 1
'
7 y)
dF( I y)
,
17

dy

and he refrom is obt ained y1 = y+ h as first appro x imation , .

Us ing this value o f y1 as y in ( 9 ) gives a second appro x imation ,

which u s ually is s u fficiently c lose .

1 77 Ne w f u nctions are defined by the integrals of the


.

analyt ic f u nctions ( 1 ) or and by the ir reverse f u nct ions .

They are called Ab eli a n i ntegrals and Abeli an fu ncti ons .

Thu s in the most general case the e x plicit f u nction


corresponding to ( 1 ) being

9
the inte gral ,

f ( O dd s

then is the general Abelian integral and its reverse function , ,

the general Abe lian f u nct ion .

= 1 function 2 gives the rational f u nction


( )
a I n the case n , , ( )
f5 ) and its special case the inte ger f u nct ion
,

Fu nction ( 6 ) can be integrated by po w ers of x ( 5 ) can be .

resolved into part ial fract ions and t her eby leads to inte grals
,

of the follow ing forms


2 98 E NGI N E E RI NG M A TH E M A T I CS .

and as reverse f u nct ions thereo f there are obtained the , tri go
nome tr ic fu ncti ons .

x = s in 2 ,

I n the case of the pl us sign integral ( 1 6 ) gives


,

— log {V 1 l arc sinh


=
2—
+ 11 x r
.
,

and reverse f unct ions thereo f are the hyperb oli c f u nc ti ons ,
_z


s 1 nh z :

cosh z .

The trigonometric f unctio ns are e x pressed by the series

sin z= z

a 2
4
z

e 14 19 .

as seen in C hapter I I paragraph 5 8,


.

The hyperbolic f u nct ions by subst itut ing ,

the series can be e x pressed by the series :


3 5
2 2 27

2 4 6
z: 2 2

In the ne x t case =3 or 4,
1 79 .
, p

Co d

already leads beyond the elementary f u nct ions that is ( 2 3 ) , ,

cannot be inte grated by rationa l logarithmic or are functio ns , ,


APP E N D I X A . 2 99

b ut gives a new class of f unctions the elli pti c i ntegrals and , ,

the ir reverse f unctions the elli ptic fu nc ti ons so called becau se


, , ,

they bear to the ellipse a rela tion similar to that which the ,

trigonometric f u nctions bear to the c ircle and the h yperbolic


functions to the equilateral hyperbola .

The integral (23 ) can be resolved into elementary functions ,

and the three classes o f elliptic integrals :

I
I

di :

( 23 — x
) (1

(These t hr ee classe s o f integrals mav be e x pressed in several


di fferent forms ) .

The rever e fu nctions of the elliptic integrals are given by


s

the elliptic functions :


x/ E= sin am u , c
( );
V1— r = eos

);
c l
known respectively as sine amplitude cos ine amplitude delta
, ,
-
,
-
,

amplitude .

E lliptic f unctions are in some respects similar to trigo


no me t ric functions as is see n b ut they are more general
, , ,

depend ing as they do not only on the variable cc b ut a ls o on


, , ,

the constant c They have the interes ting property of being


.

d ou b ly peri odic The trigo nometric f u nctions are periodic w ith


.
,

the periodicity 2 77 that is repeat the same val u es after e very


, ,

change of the angle by The elliptic functio ns have t wo


periods p ; and 292 that is , ,

sin am (u + u p 1 + mpg c) = sin ,


am u , c
( ) etc
, .
(2 6)
hence increasing the variable u by any m u ltiple of e ithe r
,

period p 1 and pg repeats the same val ues


,
.
3 00 E NGI NE E RI NG M A THE M A TI CS .

The two periods are given by the equations ,

I
dr
I

2 Vr ( 1 x) (1 e z
z
)

dr
'

2 V r (1 — zr
e )
1 80 . E llipt ic f u nctions can be e x pressed as ratios of two
infin ite series and these series which form the n u merator and
, ,

the de nom inator of the elliptic f u nction are called theta fu nc ,

ti ons and e x pressed by the symbol (9 th u s ,

sin (
am u , c )

cos am u, c
( )

Aam ( a , )
c ill —
0
2

and the fo u r 0 f u nctions mav be e x pressed by the series


30 (:c) 1 — 2q cos 2x + 2q 4
cos 42 — 2g9
cos
25
01 0 3) = 2ql /4
sin 12
— 2q9 /4
sin 3 2: + 2 9? sin 5 2:
25
=
02 05 ) 2 q1 /4
c os 2: cos 3 a3 + 2 g7 cos
03 (x) 1 + 2 q cos 2 x 2
+ q4
cos 42: + 2 99 cos 6x
w here

q
“and
e a
P2
( 30 )
PI
the c ase o f inte gral f u nction
In where p > 4 similar ,

inte grals and their reverse f u nctions appear more comple x ,


302 E NGI N E E RI NG M A TH E M A TI CS .

TOevery point P = x y of the e llipse then corresponds a ,

point P 1 = 2: yl on the circle w hich has the same absci ssa r


, , ,

and an angle 3 = AOP 1 .

The a re of the ellipse from A to P then is given by the , ,

integral ,
— z l
( 1 c z dz

where

z sin 2
0 and c

is the e ccentricity of the e llipse .

FI G 9 3 . . Re ctification of E lli ps e.

Thus the problem leads to an e lliptic integral of the first


and of the second class .

For more complete disc u ssion o f the elliptic inte grals and
the elliptic f u nct ions re i e re nce mus t be made to the te x t books
,
-

o f mathematics .

B . Spe c i a l Fu ncti o ns .

1 83 . Numerous special f unctions have been derived by the


e x igencies o f mathematical problems ma inly of astronomy but , ,

in the latter decades also of physics and o f engineer ing Some .

o f them have already been disc ussed as special cases of the


general Abelian integral and its reverse f u nction as the e x po ,

ne nt i al, tri gonometric hyperbolic etc f u nctions


, ,
.
,
.
AP PE N D I X A 3 03

Fu nctions may be represented by an infinite series of terms ;


that is as a s um o f an infinite nu mbe r o f terms w hich pro
, ,

gressively decrease that is approach z ero The denotation of


, , .

the terms is comm only represented b v the s ummation sign 2 .

Thus the e x ponentia l functions may be written when ,

defi ning ,

m=l
; M = 1 X2 X3 X4 X H . X n,

2 00 ”
$ fl?
e
x =1 Zn
,
0 0
,
0

0 n

wh ich means that terms are to be added f or all val ues of n


I
,

from n = 0 to n = oe
The trigonometric and hyperbolic f unctions may be w ri tten
in the form

fi n x = 2

w ig

cos h x=
I}
. I
4
IE
Functions also may be e x pressed by a series of factors ;
that is as a prod u ct o f an infinite series o f factors which pro
, ,

gre ss ive ly approach unity The product series is commonly


.

represented by the symbol


Thus for instance the sine fu nct ion can be e x pressed in the
, ,

form ,

sin x=x 1
l
l

1 84 . I nte gration of known f u nctions frequ ent ly leads to new


func tions . Thus from the ge n er al algebraic funct ions w ere
304 ENGI NEE RI NG M A TH E M A TI CS .

derived the Abelian f u nctions I n physics and in e ngineering


.
,

integration o f special f u nctions in this mann er frequ ently leads


to new special f unct ions .

For instance in the st u dy of the propagation through space


, ,

of the ma gnet ic field o f a cond u ctor in w ireless te le graphy, ,

lightning protection etc we get new f u nctions I f i = f (t)


,
.
,
.

is the current in the conductor as function o f the t ime t at a


, ,

d istance a: from the conductor the ma gnetic field lags by the


x
t ime tl )
whe re S is the speed of propagation (velocity of
S
light ) . Since
the field inte nsity decreases inversely propor
t io na l to the distance x it th u s is proportional to
,

and t he total magnetic flu x then is

If the current is an alternating current that is f (t) a , ,

trigonometric fu nct ion o f time e q u ation (42 ) leads to the


,

f u nctions ,

If
the current is a d irect current rising as e x ponent ial ,

function of the t ime equat i on (42 ) leads t o the f u nction


, ,
ENGI NEE RI NG M A TH E M A TI CS .

The definite integral ,

Z f (x ,
u , 1) ) dx ,

is no t a function of at but st ill is a function of,


u, v
may be a new funct ion .

1 86 For instance let


.
,

“ —l
y e q ;
then the integral ,


s
x
x
u
dx,

is a new function of u called the gamma fu ncti on


,
.

S ome propert ies of this function may be derived by part ial


integration thus ,

if n is an integer number ,

—1 — —n
) (u ( u n) (
F u ),
and since

if u is an integer n u mber then , ,

'
I (u ) =
| u — 1 .

C . E xpo ne nt i al , Tri go no m e tri c a nd Hype r b o l i c Fu n cti o ns .

(a) FUN C T I O N S OF RE AL V ARI AB L E S .

The e x ponential trigonometric and hyperbolic func


1 87 .
, ,

tions are defined as the reverse f u nct ions of the integrals ,

arc s in x;
APPEN D I X A . 30 7

x = sin u ,

sinh u ;

cosh u .

From ( 5 7 ) and ( 58 ) it follows that

sin 2
u + cos 2
u = 1 .

From ( 60) and (6 1 ) it follows that

cos 2
hu — sin 2h
u 1 .
( 63 )
— :r —u
Substit u ting ) for a: in gives ) instead of u,

and therefrom ,

— sin u ( 64)

Substit ut ing -
u ) f or u in reverses the sign of 32,

that is ,
—u —
sinh ) sinh u .

Su bstituting ) f or x in ( 58) and


— x
does not change
the value o f the square root that is , ,

cos —
u ) = cos u,

cosh u ) cosh u ,

Which sign ifies that cos u and cosh u are e ven fu nc ti ons
sin u and sinh u are odd fu ncti ons .

Adding and s u btractin g (60 ) and gives

e
i u = c o sh u t sinh u .
3 08 ENGI NEERI NG M A TH E M A TI CS .

(b ) FU N CT I O N S OF I M AG I NAR Y V AR I AB L E S .

= x gives
x8 8 . Su bst it u t ing in ( 5 6 ) and
,
sr = jy th,
u s y j ,

u u

x = s in u x = s inh u =

; 7

he nce , j u hence ju ,

y sinh ju y sin ju ;

1— y
2 = cos
ju ;

Resu b st it ut ing x in both


e


f
= s 1 nh u
x = s 1n u x


V1 x
2 =
c os u = c o sh ju
514 ?”
5 + f
2

Add ing and subtract ing ,


= c os h
e
i fu = c o s u t
f s in u ju i s i nh ju
= cos
s c o Sh u i s inh u ju iF j S in ju .

( ) F
c U N C T I O N S OF CO M PL E X V AR I AB L E S
1 89 . It is :
cu i w u
s e
i zv e
u
( cos vi j s in v) ;
310 ENGI NEERI NG M A TH E M A TI CS .

e
i “ = cosh u i s i nh

e
i fv = c o s v
i j sin v = co sh jvi j sinh jv;
5

(cos vi i sin v) ,

sin jv j sinh v j

sin ( u j; j v) = sin u cosh vi —


j cos u sinh v

v
S


sm
.

u t ] co s u ;
2


f
?

"
s A
cos u = cosh J u =

a”

e
cos j v = cosh v
2

( u i j v) = cos u cosh vZF S in u j sinh v


—v v _ _v
e s €
cos u ? 7
2 2

sinh u

sinh j v j sin I)
2

sinh (u j: jv) = sinh sin v



u cos 1) i f cosh u
u _ _u
€ e
cos 1) j: 7
2

c osh u cos ju ;

cosh jv cos v

c osh (u i jv) cosh u cos 1) j: j sinh u sin 1)

cos vi ]
2
APPEN D I X A . 311

And f rom (b ) and (d ) , respect ively (c) and ( c) , it follows that

sinh (u i jo) = j sin ( i v— ju ) i j s in ( o t ju ) ;


cos h (u j: jv) cos ( I) ? ju )
.

Tables o f the e x ponential funct ions and the ir logarithms,


and of the hyperbolic fu nct ions w ith real variables are given ,

I n the followi ng Ap p endi x B.


APPE NDIX B .

TW O T AB L E S OF E XP O NE NT I AL AND H YPE R B O L I C
FU NCT I O NS .

T AB L E I .

e= 2 7 1 83
.
, lo g e =0 . 4343 .

—3
X 10 X X

0 95 1
314 ENGI NEERI NG M A TH E M A TI CS .

T AB L E I I — C onti nu ed .

E X PO NE NTI AL AND H Y PE R B O L I C FU NCTI O NS .

b g e+
z lo g mnh . z

O 20. 0 0 8 6 8 59
. 9 9 13 14 1
. O 8 1873 0 20 134
.


H 28 40 3 H 0 3 1 42 0 2 5 26 1
O 25. 0 1 085 7 4
.
D
C . 89 1 4 2 6 . O 7 7 88 0
. . .

869 7 1 2 O 7 4082

H 0 45 3 4

O 30. 0 1 3 0 2 88
.
D
C . b . 3 4 9 86 . . 0 3 0 45 7
.


O 35 O 84 7 99 7 H 4 190 7 O 7 04 69 0 6 1 88 0 357 19

. 0 1 5 200 3
. . . . b . .


O 8 2 6 28 2 O 67032 H 08 1 08 0 4 107 6

O 40. 0 1 737 18
. . t . 4 9 1 83 . . .

D 8 04 5 6 7 H 56 83 1 O 63 7 63 0 4 65 34

O 45. 0 1 9 5 43 3
. ( . . . b . 1 0 29 7 .

O 0 2 1 7 1 47
. O 5 2 108
.

. 4 3 4 29 4

N0 . 8 6 8 5 89 0 1 3 5 34
.

n
c H 0 8573 6 o
o 9 14 2 6 4 1 2 1825
. 0 0 8 20 8 5 6 1 3 23 wU ‘


. . .
.

i
>
H 3 0 28 8 3
.
o
o . 69 4 1 1 7 2 0 0 85 5
. 0 049 7 9 7
.
1 0 06 7 7
. 10 0 178
. wO
v o
3 3 1 1 54 wM

b D . 5 2 00 3 0 o . 47 99 7 0 . 0 0 30 1 9 7
. 16 571 8
. 1 6 54 2 6
.

:
t H 7 37 1 78
.
o
o . 2 6 282 2 . 59 8 3 0 0 1 83 1 6
.
2 7 3083
. 27 . h O
i

H 9 54 3 2 5 h M
o
u
C . o . 04 5 6 7 5 . 0 170 0 0 1 1 109
.
45 0 1 4 1
.
4 5 0030
. t


0 O . 1 7 1 47 2 . 203

. 7 13

. 3 42 9 4 5

. 6 8 5 8 90
I ND E X

Ab e lia n int e g ral s a nd fu nctio n s, 3 05 B a e of lo g arithm 2 1


s ,

Ab s ol u t e nu m b e r , 4 B i no m ial se ri s w ith s m all q u a nti


e

val of fractio nal e xp re io n 49


ue ss ,
ti e s 1 9 3 ,

of g ne ral nu m b e r 3 0
e ,
th ore m i nfinit e se ri e s 5 9
e , ,

Acc rac y lo ss of 2 8 1
u , ,
of tri g o no m e tric fu nctio n 1 0 4 ,

of a pp ro xi m atio n sti m ate d 2 00 e ,


B iq adratic p arabola 2 1 9
u
,

of calc latio n 2 7 9
u ,

of cu rv q uatio n 2 1 0
e e ,

of transm i s io n li n q u atio n 20 8
s e e s,

Additio n 1 ,
C alc u latio n acc u rac y , ,
2 79
of g e ne ral nu m b e r 2 8 ,
ch e cking of 2 9 1 ,

a nd s u btractio n of tri g o no m tric e nu m e rical 2 58 ,

f nctio ns 1 02
u ,
r e liability 2 7 1 ,

Al g e bra of g e ne ral n m b r or c m u e o C ap acity 6 5 ,

p l e x qu a ntity 2 5 ,
C hain fractio n 2 08 ,

Al g braic e xp r ss io n 2 9 4
e e
,
C ha ng e of c u rv e l w 2 1 1 2 3 4 a , ,

functio n 7 5 ,
C haract e ris tic s of e xp o ne ntial cu rv e s ,

Alt e rnati ng c u rr n t a nd volta g ve e e c 228


tor 4 1 ,
of p arabolic a nd hyp rbolic c u rv e es ,

fu nctio ns 1 1 7 1 2 5 , ,
223
w av e s 1 1 7 1 2 5
, ,
C harg i ng c u rr e nt m axi m u m of con

Alt e rnatio ns 1 1 7 ,
d e ns r 1 7 6 e ,

Alt e rnator hort circ u it c u rr nt p


s e ,
a C h e cki ng calc u latio n 2 9 3 a s,

p ro x i m a t e d , 1 9 5 C i p h e rs nu m b e r of in calc u latio n
, ,
s,

Anal ytical calc u latio n of xtr e em a ,


2 82
1 52 C ircl e d e fin ing tri g o no m e tric fu nc
f nctio n 2 9 4
u , tio ns 9 4 ,

Ang l e s e e al o P has e a ngle


,
s . C o ffici e nts nk no w n of i nfinite
e ,
u ,

App ro x i m atio n calc latio n u ,


2 80 s e ri e s 6 0 ,

b y chai n fractio n 2 0 80 ,
C o m bin atio n of e xpo ne ntial f nc u

App ro xi m atio n g ivi ng ( 1 s s ) a nd tio ns 2 3 1 ,

(1 s ), 201 of g e ne ral nu m b e rs 2 8 ,

of infin it e e ri e s 5 3 s ,
of v ctors 2 9
e ,

m e thod s of 1 87 ,
C o m p ariso n of e xp o n ntial and h y e

Arbitrary co ns tant of s ri s 69 7 9 s e e
, , p e b o lic c u rv s 22 9
r e ,

Ar e a of tria ng l 1 06 e, C o m p l m ntary ang l e in trig o no


e e s

Arra ng e m e nt of nu m e rical calc la u m e tric fu nctio ns 99 ,

tio ns 2 7 5 ,
C o m p l e x i m ag i nary q u antiti es s e e ,

Attack m e thod of 2 7 5
, , Ge ner al nu mb e r .
3 16 I N DE X

C o m p l e x qu antity , ,
17 De c i m al s ,
nu m b e r of ,
in calc u latio ns ,

alge bra ,
27 2 82
see Ge nera l nu mb er . inlo g a rithm ic tabl e 2 8 1 s,

C o nju g at e nu m b e rs 3 1 ,
D e fi nite i nte g rals of tri g o nom e tric
C o ns ta nt arbitrary of s e ri e s
, ,
6 9, 7 9 fu nctio ns 1 03 ,

erro rs ,
1 86 De g re es of acc u racy 2 79 ,

factor ith p arabolic a nd h yp r


w e D e lta am p litu d e 2 99
-
,

b lic c u v e s 2 2 3
o r ,
D i ff re ntial e q u atio ns 64
e ,

p h e no m e na 1 0 6 ,
of e l ctrical e ng i ne e ring 6 5 7 8 86
e , , ,

t rm s of c u rv e e qu atio n 2 1 1
e , of s e co nd ord e r 7 8 ,

of m p irical c u r ve 2 3 4
e s, D iffe re ntiatio n of tri g o no m e tric
i n e x p o ne ntial c u rv s 2 3 0 e ,
functio ns 1 03 ,

w ith e x p o n n tial c u rv 22 9 e es, D i ph a nt ic e qu atio ns 1 8 6


o ,

i n p arabolic a nd h yp e rbolic D is tort d l e ctric w ave s 1 08


e e ,

c u rv es 22 5 , D i tortion of w ave 1 39
s ,

C o nve rg e nc y d e te rm in atio ns of D ive rg e nt s e ri e s 5 6 ,

p ot e ntial s ri e 2 15 e s, D ivi s io n 6 ,

of se ri e s 5 7 ,
of g e ne ral nu m b e r 42 ,

C o nv r g nt s e ri e s 5 6
e e , w ith s m all qu a n titi e 1 90 s,

C ore l es s b y p ote ntial s e ri e s 2 1 3 ,


D o bl e ang l es in tri g o no m e tric
u

c u rve e valu ation 244 ,


fu nctio ns 1 03 ,

C o se cant fu nctio n 9 8 , p e ak d w av e 2 5 5 2 6 0 2 66
e , , ,

C o s h f unctio n 3 05 , p e riodicity of e lli p tic fu nctio n s,

C o sine am p lit d e 2 9 9
-
u ,
2 99
co m p o ne nts of w a v 1 2 1 1 2 5 e, , s cal e ,
2 89
fu nctio n 9 4 ,

s e ri e s 8 2 ,

ve rse d fu nctio n 9 8 ,

C otang e nt fu nctio n 9 4 ,
e, 21
C ou nti ng 1 ,
E ff ici e nc y m axim u m of alt rnator e ,

C u rre nt change c u rv e e val u atio n 1 62


241 of i m pu l e tu rbi ne 1 5 4 s ,

of di tort e d voltag w av e 1 69
s e ,
of i nd u ctio n g e ne rator 1 7 7 ,

i npu t of ind u ctio n m otor ,


ap of tra ns form e r 1 55 1 7 4 , ,

p ro x i m a t e d , 1 9 1 E l e ctrical e ng i ne e ri ng di ff r e ntial ,
e

m axi m umof alt r atin g tra ns


e n e q atio ns 65 7 8 8 6u , , ,

m i s io n circ u it 1 59
s ,
E lli p s e l e ng th of arc 3 01
, ,

C u rv e s ch e cking calc u latio ns 2 93 b


, ,
E lli p tic i nt e g ral s and fu nctio n 2 99 s,

e m p irical 2 09 ,
E m p irical c u rv e s 2 09 ,

law cha ng e 2 3 4
, ,
e val u atio n 2 3 3 ,

ratio nal e q u atio n 2 1 0 ,


e q u atio n of c u rv e 2 1 0 ,

u se of 2 84 ,
E ng i ne e rin g di ff e r e ntial e q u atio ns ,

re p orts 2 9 0 ,

E q u ilat e ral h yp e rbola ,


2 17
Data o n calc u latio ns and c u rv e s ,
27 1 E rrors co ns tant 1 8 6
, ,

d e rive d fro m c u rve ,


2 85 n m e rical 2 9 3 b
u ,

De ci m al e rror 2 9 3 b ,
of ob s e rvatio n 1 80 ,
3 18 I N DE X
In co m m e nsu rabl e w ave s 2 5 7 , Lo g a ri t h m at io n, 2 0
I nd t e rm inate co e ffici nt
e m e thod e s, ,
of g e ne ral b rs 5 1
nu m e ,

71 L o g a rith m ic c u rve s 22 7 ,

In d e t e rm i nate co ffici nts of infi e e functio ns 2 9 7 ,

nit e s e ri e s 6 0 , p ap e r 2 33 2 87
, ,

I ndivid u al 8 s, s cal e 2 88
,

I nd cta nc e 65
u ,
tabl e s nu m b e r of d e c im al s in 2 8 1
, ,

I nfinit e se ri e 52 s, L o s of c u rv e i nd u ctio n m otor 1 8 3


s ,

val e s of c u rv e s 2 1 1
u ,

of e m p irical c u rv e s 2 33 ,

I nfl ctio n p oi nt s of c u rv e
e 1 53 s,

I m p e da nc e v e ctor 4 1 ,
M ag ne tic characte ris tic o n se m i
I m p licit a nal ytic fu nctio n 2 9 4 ,
log arithm ic pa pe r 2 88 ,

I nt g ral f u nctio n 2 9 5
e ,
M ag ne tit e arc volt am p e re charac ,
-

I nt e g ratio n co ns ta nt of s ri 69 7 9 e es, , t e ri s t ic , 2 39
of di ffe r ntial e q atio n 65
e u ,
characte ri tic e val u atio n 2 46
s , ,

b y i nfinit e s ri s 6 0 e e ,
M a g nit u d e of e ff ct d t e rm i natio n e ,
e
,

of tri g o no m e tric fu nctio ns 1 0 3 ,


2 80
I nt lli g ibility of calc u latio n 2 83
e s, M ax i m u m , se e E x tr e m u m .

I nt e rc e p t s d e fini ng ta ng e nt a nd c o
,
M axi m a , 1 47
tang e nt fu nctio ns 9 4 ,
M c L au ri n s s e ri e s w ith sm all

qu an
I nvol u tio n 9 ,
titi e s 1 9 8 ,

of ge ne ral nu m b e rs 44 ,
M chanis m of calcu lati ng e m p irical
e

I rratio nal nu m b e rs 1 1 ,
cu r ve s 2 37 ,

I rratio nality of r e p r e e ntatio n b y s M thod s of calc u latio n 2 7 5


e ,

p ot e ntial s e ri e s 2 1 3 ,
of i nte rm e diate co e ffici e nts 7 1 ,

of l e as t s qu ar e s 1 7 9 1 86 , ,

of attack 2 7 5 ,

M i ni m a 1 47 ,

M i ni m u m se e E xt e mu m
,
r .

M lti p l e fr e qu e nci s of w av es 2 7 4
u e ,

M u lti p licand 3 9 ,

M lti p licatio n 6
u ,

Le a t s q u are s
s ,
me thod of 1 7 9 , ,
1 86 of ge ne ral nu m b e rs 39 ,

L i m itatio n of m ath e m atical r e p re w ith s m all q u a ntiti e s 1 88 ,

s e nt a t io n, 40 of tri g ono m e tric fu nctio ns 1 02 ,

of m e thod of l e as t s qu are 1 8 6 s, M u lti p li e r 39 ,

of p ot e ntial se ri e s 2 1 6 ,

L i m iti ng val u e of in finite e ri e 54 s s,

L i ne ar n m b e r 33
u ,

s ee Pos i ti ve a nd Nega ti ve nu m b er . Ne g ative ang l e s i n tri g o no m e tric


L i ne calc u latio n 2 7 6 ,
functio ns 98 ,

q u atio ns app roxi m at e d 2 04


e , ,
e xp o ne nt s 1 1 ,

L o g arith m of e x p o ne ntial c u rv e 22 9 ,
nu m b e r 4 ,

asi nfi nit e s e ri e s 63 ,
Nod e s i n w av e s hap e 2 5 6 2 7 0 , ,

of p arabolic and hyp e rbolic c u rve s , No n p e riodic c u rv e s 2 1 2


-
,

22 5 N o l e e ffi ci e nc y m axim u m 1 50
zz , ,

w ith sm all q u a ntiti e s ,


1 97 Nu m b e r g e ne ral 1
, ,
I NDE X 3 19

Nu m e rical calcu lations 2 7 5 ,


Plott ing of c u rve s 2 1 2 ,

val ue s of tri g onom e tric functio ns , p ro p e r a nd i m p ro p r 2 86 e ,

1 01 of e m p irical c rv e 2 34 u ,

Polar c o ordi nat es of g e ne ral


- nu m
b e r, 2 5 , 2 7
e xp re ss io n of g e ne ral nu m b r e ,
2 5,
O b s e rvati on e rrors 1 80 , ,

O ctav e as lo g arith m ic s ca le ,
2 88 ol hase re latio n r e du ci ng tri g o
P yp ,

no m e t i c se ri e s 1 3 4
'

O dd fu nc t o ns , 8 1 , 9 8 305 ,
r ,

riod c
pe ,
1 22 of tri g o no m e tric fu nctio n 1 0 4 s,

har o ic s
m n in sym m e trical w ave , s y t m of p oi nt s or v e ctor s 4 6
s e ,

1 17 Po itiv e nu m b e r 4
s ,

aratio n 1 2 0 1 2 5 1 3 4
sep , , ,
Pot e ntial se ri e s 5 2 2 1 2 , ,

O m iss io n in calc latio ns 2 9 3 b


s u , Po we r factor m ax i m u m of in d u ctio n
O p e rator 40 ,
m otor 1 49 ,

O rd e r of s m all q u a ntity 1 88 ,
m ax i m u m of alt e rnati ng tra ns
Os cillati ng fu nctio ns 9 2 ,
m i ss io n circ it 1 58 u ,

Ou tpu t see Po we
,
r. of g e ne rator 1 6 1 ,

of h u nt e d r e si tanc e 1 5 5
s s ,

of s torage batt e ry 1 7 2 ,

of transform e r 1 7 3 ,

“ of transm iss io n li ne 1 6 5
5 add e d a nd s ubtracte d i n
,
a nd
no t ve ctor p rod u ct 42
7
,

tri g o no m tric fu nctio n 1 00 e ,


of hu nt m otor a ppro xi m at e d 1 89
s , ,

a pp roxi m ate d b y chai n fractio n ,


w ith s m all q a n titi e s 1 9 4 u ,

2 0 80 Probability calc u latio n 1 8 1 ,

P airs of hi g h harm o nic 2 7 0 s, Prod u ct s e ri e s 3 0 3 ,

P arabola co m m o n 2 1 8
, ,
of tri g o no m e tric f nctio n 1 02 u s,

Parabolic c u rv e s 2 1 6 ,
Pro je ctio n d e fi ni ng co i n f nctio n
,
s e u ,

Parall e lo g ram la w of g e ne ral nu m 94


be r 2 8 s, P roje ctor d e fin i ng s i ne fu nctio n
, ,
94
of ve ctors 2 9 ,

P ak e d w av e 2 5 5 2 5 8 2 6 1 2 6 4
e , , , ,

P n d u l u m m otio n 30 1
e ,

P rc e nta g e cha ng e of p arabolic a nd


e Q u adrants , s i g n of tri go no m e tric
h yp e rbolic c u rve s 22 3 ,
fu nctio ns 9 6 ,

P riodic c u rv e s 2 5 4
e , Q u adratu r e nu m b e rs ,
13
d e ci m al fractio n 1 2 , Q u art e r s cal e 2 89 ,

p h e no m e na 1 0 6 , Q u ate rnio ns 2 2 ,

Pe riodicity do bl e of e lli p tic f nc


,
u ,
u

tio n 2 99 s,

of t ri g o no m e tric functio ns 9 6 ,

P rm e ability m ax i m m 1 48 1 7 0
e u , ,
Ra u s div e ctor of g e ne ral nu m b r 2 8 e ,

Pha se a ng l e of fractio nal xp r s io n e e s ,


R a ng e of co nve rg e nc y of s e ri 5 6 e s,

49 R atio of variatio n 2 2 6 ,

of ge n ral n e um b r e ,
28 R atio n al e qu atio n of c u rv e 2 1 0 ,

Plai n nu m b e r ,
32 fu nctio n 2 9 5 ,

see Gener al nu mb er . Ra tio nality of p ot e ntial s e ri es 2 1 4 ,


320 IN DE X
Re al nu m b e r 2 6 , S ha p e of c u rv e s 2 1 2 ,

R e cta ngu lar c o ordi nat e s of g e ne ral p ro p e r i n p lotti ng 2 86


-
,

nu m b e r 2 5 ,
of e xp o ne ntial c u rv e 22 7 2 30 , ,

Re d u ctio n to ab s ol u t e val u s 4 8 e ,
of functio n b y c u rv e 2 84
, ,

R latio ns of h yp e rbolic tri g o no


e of hyp e rbolic fu nctio ns 2 32 ,

m e tric and e xp o ne ntial f u nc of p arabolic and h yp e rbolic c u rve s,

t i Ons , 309 2 17
Re lativ e ne ss of sm all q ua ntiti e s 1 88 ,
S harp ze ro w av e 2 55 2 6 0 2 6 5 , , ,

Re liabilit y of nu m e rical calc u latio ns , S hort circ u it c u rr e nt of alte r nator ,

2 93 app ro xi m ate d 1 9 5 ,

Re p orts e ng i ne e ri ng 2 90
, , S i g n e rror 2 9 3 0,

R s i s ta nc e 65
e ,
of tri g o no m e tric fu nctio ns 9 5 ,

R e s ol u tio n of v e ctor 2 9 s, S ilico n s t e e l i nv e s ti g atio n o f h ys t e re


,

Re v e r s al b y ne g ative u nit 1 4 , s is c u rv e 2 4 8 ,

do u bl e at e ro of w av 2 5 8 2 6 1
,
z e, , S i m p lificatio n b y a pp ro xi m ati on 1 8 7 ,

R e ve rse fu n ctio n 2 9 4 , S i ne a m p lit u d e 1 9 9


-
,

R i g ht tria ng l e d e fi ni ng t ig o no m e t r co m p o ne nt of wave 1 2 1 1 2 5 , ,

ric fu nctio ns 9 4 ,
f unctio n 9 4 ,

R i pp l s in w av e 45
e ,
s e ri e s, 82
b y hi g h har m o nic s 2 7 0 ,
v e rsu s fu nctio n 9 8 ,

R oots of g e ne ral nu m b e rs 45 ,
S i ne f u nctio n 305 ,

e xp r es s d b y p e riodic chai n frac


e S lid e r u l e acc u rac y 2 8 1 ,

tio n 2 086 ,
S m al l qu a ntiti e s a p p ro x i m atio n 1 8 7
, ,

w ith s m all q u a ntiti e s 1 9 4 , S p cial f u nctio ns 3 0 2


e ,

of u nit 1 8 1 9 46 , , , S q u ar e s l e a s t m e thod of 1 7 9 1 8 6
, , , ,

R otatio n b y ne g ativ e u n it 1 4 , S t e a m p ath of t u rbi ne 3 3 ,

b y q u adrat u re u nit 1 4 , S ubtractio n 1 ,

of g e ne ral nu m b e r 2 8 ,

of tri g o no m tric fu nctio ns 1 02


e ,

S u m m atio n s e ri s 3 03 e ,

Su p e r p o s itio n of hi g h harm o nic s 2 7 3 ,

Saddl e p oi nt 1 6 5 , Su pp l e m e ntary a ng l e s i n tri g o no


S w tooth w av
a - 2 46 e, ,
2 5 5 , 2 58 , 2 6 0 , m e tric f u nctio ns 99 ,

2 65 S u rg i ng of s ynchro nou s m achi ne s,

Sc alar 2 6 2 8 3 0
, , , 3 01
S cal e i n c u rv e p lotti ng p ro p r nd ,
e a Sym m e trical c u rv e m axi m u m ,
1 50
i m pro p e r 2 1 2 2 86 , , p riodic f u nctio n 1 1 7
e ,

fu ll do uble half e t c 2 8 7
, , ,
.
, w av e 1 1 7
,

S ci e ntific e ng i ne e ri ng r e cord s 2 9 1 ,

Se ca nt f u nctio n 98 ,

Se co nd har m o nic e ff ct of 2 66 ,
e ,

Se co ndary e ffe cts 2 1 0 ,


T ab u lar form of calc latio n 2 7 5 u ,

p h e no m e na 2 34 ,
T a ng e nt fu nctio n 9 4 ,

Se m i lo g arith m ic p a p e r 2 8 7 T a ylor s s e ri e s w ith s m all q u a ntiti e s



-
, ,

Se ri e s e xp o ne ntial 7 1
, ,
199
i nfinit e 52 ,
T e m pe ratu re w av e 1 3 1 ,

tri go no m e tric 10 6 ,
T e m p orary u s e of p ot e ntial se ri e s ,

S e v e nth harm o nic 2 62 ,


2 16

You might also like