Linear induction motor-equivalent-circuit model

J. Duncan, C.Eng., M.I.E.E., Mem.I.E.E.E.

Abstract: An equivalent-circuit model of a linear induction motor is developed, using the rotary-motor model as a basis. The rotary-motor model is modified to account for the so-called 'end effect' and is used to predict output thrust, vertical forces and couples. These predictions are checked against test results of a practical motor used on a prototype transit vehicle.

List of symbols Ri R21 L1 L2i Lm D v d T c^i co 2 P I2\ Im 11 12 e hea ime Imea = = = = = = = = = =

= = = =
=

= =

MMF = 1

primary resistance secondary resistance referred to primary primary leakage inductance secondary leakage inductance referred t o primary magnetising inductance motor effective length motor velocity distance between midpoints of secondary and priary current layers motor pole pitch supply frequency slip frequency number of poles in motor secondary slip current referred to primary magnetising current primary current instantaneous secondary eddy currents referred t o primary average value of i2e over motor length instantaneous effective magnetising currents average value of effective magnetising current over motor length magnetomotive force

The equivalent circuit is essentially a performance model and generally does not require internal machine design data. The machine data can be obtained from the machine terminals by measurement and is then incorporated into the model. The linear motor tends to present more difficult problems of measurement, but in principle the same kind of data can be obtained as in the case of the rotary motor. In the case of the linear motor, it is also possible to measure the vertical forces under controlled static conditions to obtain motor constants for calculating these same forces under dynamic conditions. The process of testing is not discussed in this paper owing to lack of space and the author's lack of direct involvement in the testing. The motor used to test the model is the motor destined to be the first large-scale revenue application of traction linear motors in the world. To the author's knowledge, it is also the only linear motor which has been subject to static testing to establish the equivalent-circuit components.
.primary

tOMOl

Introduction secondary sheet secondary back iron Fig. 1 Side view of linear induction motor 2 Linear induction motor

The existence of a reasonably accurate equivalent-circuit model of the rotary induction motor has been of significant benefit to engineers who design and use these motors. There is, however, no similar generally recognised model for the linear induction motor. Most of the existing models depend on field theory and can show reasonably accurate performance predictions. They do not, however, provide the convenient conceptual frame of reference that the equivalent circuit provides. The complexities of the field theory appear to rule out a simple equivalent-circuit model. Laithwaite, in Reference 1, p. 58, suggests that '. . . these (end) effects cannot be readily expressed in terms of components in an equivalent circuit', while Reference 2, p. 83, concludes that '. . . conventional circuit analysis has proved to be incapable of modelling the asymmetric magnetic and electrical conditions which prevail in the linear motor'. This paper attempts to develop a simple equivalent circuit for the linear motor by using only the usual equivalent circuit components and avoiding the use of field theory. The resulting model accounts for many of the experimentally established characteristics of linear motors.
Paper 2154B (PI), first received 25th February and in revised form 11th August 1982 The author is with Transportation & Vehicle Systems, Transportation Technology & Energy Division, Ministry of Transportation and Communications, 1201 Wilson Avenue, 3rd Floor, Central Building, Downsview, Ont. M3M 1J8, Canada IEEPROC, Vol. 130, Pt. B, No. 1, JANUARY 1983

The linear induction motor (LIM) is shown in Fig. 1. The primary is simply a rotary-motor primary cut open and rolled flat. The secondary usually consists of a sheet conductor with an iron return path for the magnetic flux. The primary and secondary together form a single-sided linear induction motor. The secondary iron return path may be replaced by a suitable airgap and a second primary to produce a double-sided arrangement. Single- or double-sided motors which use sheet secondaries are electromagnetically equivalent, and both may be analysed by the methods presented in this paper.
3 Rotary-motor equivalent circuit

The rotor induction motor (RIM) equivalent circuit is well known and may be represented as shown in Fig. 2. The component values used in this circuit represent per-phase values and are referred to the primary. To obtain the total motor performance, the outputs given by the circuit must be multiplied by three. In many modern applications, the supply frequency becomes a variable, as shown in Fig. 2. The iron loss component tends to become more complex with variable supply frequency, but its influence on motor performance is generally small. In 0143-7038/83/010051 + 07 $01.50/0 51

the case of the linear motor, the use of a large airgap leads to low flux densities which tend to make the iron loss very small, so the iron loss component may be neglected without significant loss of accuracy. The RIM equivalent circuit, however, does not account for the behaviour of the linear motor, and this failure is usually attributed to a phenomenon known as the 'end effect'.
1*3,

L21

applied to the primary and the secondary sheet is replaced gradually and continuously. With 3-phase currents applied to the primary, both the MMF per unit length and flux density are sinusoidally distributed along the motor length, and change position with time. It is convenient, and does not affect the argument, to represent these waves as RMS or effective values. If all the primary slots are filled, this provides a uniform MMF per unit length of motor and flux density over the primary length. As the primary winding turns per unit length of motor is fixed, the primary magnetising current Im may be used to represent the MMF per unit length of motor, which, in turn, gives rise the airgap flux. If no saturation is present, the magnetising current will also represent the flux density. The presence of half-filled slots will be dealt with later in Section 4.5.

Fig. 2

Rotary induction motor equivalent circuit

4 4.1

End effect Magnetising branch inductance

The main difference between rotary motors can be seen in Fig. 1. In a LIM, as the primary moves, the secondary is continuously replaced by a new material. This new material will tend to resist a sudden increase in flux penetration and only allow a gradual build up of flux density in the airgap. The entry of new material and its influence on the flux will modify the motor performance. To obtain a suitable LIM equivalent circuit, it will be necessary to quantify the influence of the entry of new material on the airgap flux distribution and account for the end effect. This can be done by examining the magnetising branch of the equivalent circuit. The primary shown in Fig. 1 can be supplied with direct current to produce a flux wave which is stationary in space. This flux wave can be established with the secondary conducting sheet removed. If the secondary conductor is replaced instantaneously, the flux will tend to disappear at time zero and then rise to its original value with the passenge of time. This mechanism is illustrated in Fig. 3. Time has been normalised, using T2 as a basis, so that x = t/T2 At time zero, the secondary eddy current in the sheet will grow very rapidly to mirror the primary currents, so nullifying the primary MMF and reducing the flux to nearly zero. The rise of the secondary eddy current is controlled by the sheet leakage time constant L2i/R2l This time constant, however, is, small (about 5%) compared with the magnetising or mutual inductance time constant, so that for practical purposes it may be considered negligible.This means that the eddy current at time zero and referred to the primary winding is nearly equal to the primary magnetising current lm and will then decay exponentially with time. The time constant for this decay will be the total secondary time constant:

2 3 normalised time

Fig. 3 Rise of total airgap flux

T2 =

(Lm+L21)/R2l

(1)

The secondary or sheet eddy current opposes the primary magnetising current Im. In an induction motor, this eddy current corresponds to a secondary current which resists the build up of the airgap flux and is in phase opposition to Im. This secondary eddy current referred to the primary is designated I2e and will lag the referred secondary slip current I2l by about 90. This condition is similar to a rotary motor in a constant start-up transient, where the rotor resists the build up of the airgap flux. It is now necessary to examine what happens to the secondary eddy currents when 3-phase alternating currents are
52

It is now possible to examine the gradual replacement of the secondary sheet by considering a small element as it enters and moves through this uniform magnetic field. The small element will initially generate maximum eddy currents and reduce its flux density to zero at the leading edge of the motor, and will then allow the flux density to grow exponentially with a time constant T2 as the eddy current dies away. As this element of rail passes beyond the primary, the primary MMF seen by the rail will disappear and be reflected in the exit rail eddy currents as they try to maintain the gap flux. These exit rail eddy currents will decay very rapidly with a very short time constant related to the leakage inductance of the rail. These transient changes at the entry and exit ends of the motor are shown in Fig. 4, where they are plotted in terms of T2. If this intuitive explanation of the influence of T2 is accepted, then it can form the basis for quantifying the end effect. Changes to account for this can then be made to the classical equivalent circuit. The spatial distribution of the flux density along the motor length will, however, depend on the primary speed relative to the secondary sheet. The distance moved by the primary in one secondary time constant is vT2 where v is the primary velocity. For a given primary velocity, distance can therefore be expressed in terms of the secondary time constant T2. The primary length on this normalised time scale can be obtained as follows. The time taken for the motor to traverse a point on the
IEEPROC, Vol. 130, Pt. B, No. 1, JANUARY 1983

rail is Tv = D/v
(3)

It will be noted that Qv is a constant velocity depending on the basic motor dimensions and secondary material:

where D is the motor length. The time Tv, in terms of the secondary time constant T2, is

Qv =

DR2l/(Lm+L2l)

(5)

Q = TJT2 = DR2ll{Lm+L2l)v

(4)

Therefore, Q is dimensionless but represents the motor length on this normalised time scale. On this basis, the motor length is clearly dependent on the motor velocity, so that, at zero velocity, the motor length is infinitely long. As the velocity rises, the motor length will effectively shrink. The curve of ime between x = 0 and x Q represents the distribution of the MMF per unit length along the motor length. This concept, therefore, provides a means of describing the distribution of effective MMF along the motor length for any velocity.

The primary magnetising currents have already been used to represent the MMF per unit length of motor. In a similar way, the secondary demagnetising eddy currents, when referred to the primary, appear as antiphase currents or negative MMF per unit length. The value of the secondary eddy currents i2e varies along the motor length. However, when these currents are reflected into the series-wound primary winding, they will appear as an average current in antiphase to the primary magnetising current. These average values are developed in eqns. 6 and 7 and are shown in Fig. 4. The average value of the eddy current per unit length, i2e, is
Im rQ

{l-exp(-0} Q

hea = -^ |o exp(-x)dx = In

(6)

The average value of ime, the effective magnetising current per unit length, is
Q.

l-exp(-Q) Q

(7)

The demagnetising effect of the secondary eddy current I2ea can be represented by means of an inductance connected in parallel with Lm and carrying the eddy current I2ea, as shown in Fig. 5a. This inductance, however, is simply shunting current away from the magnetising inductance Lm and contributes nothing to excitation. The size of this inductance must be chosen to produce the same loss of excitation as the average value of i2e over the motor length. The required value of the parallel inductance is, therefore:
J

m'meal*2ea

'-'i

Q l-exp(-<2)

(8)

a.

This parallel circuit may be replaced by an equivalent series circuit, shown in Fig. 5b, where the total inductance becomes 11 - exp (-Q) Q (9)

and carries the magnetising current 'm = Jmea + hea (l0) Notice that, as the velocity tends to zero, or the motor length tends to infinity, the magnetising branch inductance becomes Lm and the linear motor becomes equivalent to the rotary as the end effect disappears.
4.2 Rail eddy-curren t losses

Fig. 4 Distribution of effective MMF at velocity > 0

If no significant saturation exists, then the MMF distribution will correspond to the flux-density distribution. The resulting flux distribution is similar to that found experimentally in References 3 to 5 and others. When this paper was in its final revision, the author's attention was drawn to Reference 7. F.C. Williams, et al. appear to have used a similar approach to the problem of flux distribution and derived the same curve as shown in Fig. 3. It is now clear that a high velocity will lead to a significant loss of flux at the leading edge of the motor. At zero velocity, however, the loss of flux will be negligible. The reduction of end effects, therefore, will require a high value for Q. The value of Q therefore indicates the motor's ability to resist the loss of output due to end effects. Unfortunately, this means using a motor with a low velocity, high electrical circuit resistance R2\ and high magnetic reluctance or low Lm. The ratio Q used above is the inverse of the 'goodness factor' developed by Laithwaite [6], and this fact may go some way towards explaining the poor performance of linear motors.
IEEPROC, Vol. 130, Pt. B, No. 1, JANUARY 1983

There is another aspect of the secondary eddy current which must be taken into account. When the eddy current circulates in the secondary, it will assume the same kind of path as that used by the slip current J21, but displaced by nearly half a pole pitch. The resistance of this circuit is R2l and ohmic losses will be produced by the eddy currents. These losses may be evaluated by determining the RMS value i2e over the motor length.

Lm

Fig. 5

a b Inductive element of linear motor magnetising branch

The RMS value of i2e is

motion: Motor thrust due to eddy-current loss

l-exp(-20| { 2(2
For the section of rail under the primary: Eddy current losses =

(11)

vQ The factor 3 sums the three phases of the motor. It should be noted that, with a very long motor, the magnetising branch resistance becomes zero, and so this thrust becomes zero, as it is in the case of the rotary motor. At zero velocity, however, the thrust will reach a limiting value of
3/;

05)

= tlR21

1 - exp ( - 20) 2Q

(12)

As the motor passes over the rail, the primary magnetising MMF seen by the rail disappears and this creates an equivalent reflected MMF and current in the rail which tries to maintain the flux (see Fig. 4). The magnetic energy at the point of exit dissipates in the ohmic resistance of the rail. These losses, however, can be represented by the time rate of change of magnetic energy as it leaves the motor gap. The airgap magnetic energy per unit length of the motor at the point of exit is given by Magnetic energy = 0.5(Lm + L2l) {Im (1 -exp(-<2)} 2 /> where Im (1 exp ( Q)) is the effective magnetising current at the point of exit. The power loss can now be obtained by multiplying the energy expression by the motor velocity v to give: exit power loss = 0.5(L m + L 21 )/ {1 - exp ( - Q) }2v/D
T2 n

(16) D and will undergo a change of sign as the velocity passes through zero. The change of sign shows a similarity to mechanical friction, and the eddy-current loss appears to be a magnetic friction.
4.3 Vertical forces

{l-exp(-0}2

Forces perpendicular to the plane of the airgap (i.e. radial forces) exist in rotary motors, but they only become significant as a result of nonuniform radial airgaps. In the case of the singlesided linear motor, these forces are always significant and must be considered in the design of the motor structure and mounting devices. The vertical forces can be separated into two components and a turning couple. The largest component is the attractive force between the primary and the secondary back iron. This force is due to the main flux crossing the gap, and is proportional to the product of the effective magnetising current squared and the inductance Lm, i.e. the stored energy contained in the magnetic gap. The RMS value of ime over the motor length is

2(2
(by substitution from eqn. 4) Adding eqns. 12 and 13 gives: Total ohmic loss due to eddy currents in rail

(13)

j|Lj o

(1_exp(_ (l-exp(-0)(3-exp(-0)|-; (17) 1 2(2

0.5

The vertical attractive force is (14)

= II R,

{l-exp(-0}
Q

To account for this power loss in the equivalent circuit, it is necessary to add a resistance to the magnetising branch, as shown in Fig. 6. This addition is sufficiently accurate, provided the inductive element dominates the magnetising branch impedance. However, as the supply frequency approaches zero, the model breaks down. This condition is analogous to the case of direct-current dynamic braking used in rotary squirrel-cage motors. In the rotary motor, the equivalent circuit does not provide a model for direct-current braking. The thrust due to these losses, which are passed across the gap, is obtained by dividing the total losses by the motor velocity. This thrust, by Lenz's law, will always act to oppose
R

The constant Ka is obtained from static tests on the motor and includes the multiplier 3 to convert from single phase to three phase. The second component is the repulsive force between the secondary slip current I2\ and its reflected current in the primary winding. This force is given by F - K
21

(19)

I,

where d is the distance between the midpoints of the primary and secondary current layers. Kr is a constant which includes the multiplier 3 to convert from single phase to three phase and may be determined from static tests. The deriviation of eqns. 18 and 19 and the tests to determine the constants are discussed in the Appendixes.
4.4 Vertical forces couple

21ZT,

Fig. 6 54

Equivalent circuit of linear induction motor

The attractive forces in the airgap are distributed along the primary length and are proportional to the flux density squared, which is derived from Fig. 4 and is shown in Fig. 7. This distribution, assuming no saturation, corresponds to the distribution of the attractive vertical force along the primary length.
IEEPROC, Vol. 130, Pt. B, No. 1, JANUARY 1983

It now becomes possible to establish the turning moment of the attractive force. This turning moment produces the 'bow wave' effect of lowering the trailing edge of the motor, relative to the leading edge, as the speed increases. As the vertical force has already been established in eqn. 18, it is now necessary to establish the equivalent point of application or centroid of the distributed forces. This can be done by treating Fig. 8 as a geometrical problem and establishing the centroid of the area under the curve. The actual forces are irrelevant to this problem, so it is convenient to normalise the curve on the basis of Im = 1.0. The centroid is established in the usual way by integrating the elemental moments of area and dividing the result by the area, as shown in Fig. 7:

for low-speed LIMs at less than 20m/s. Higher speed LIMs, or those with fewer poles, may demand a more complex treatment.
6-r
IT 3
1.0 3

0.5

motor length Fig. 8 5 Effective motor length Linear motor equivalent circuit

W=Imd-e X
_ centroid M

//
m
J*

The modifications to the rotary equivalent circuit developed in this paper make it possible to use an equivalent circuit model to calculate the motor performance and include such things as vertical force and the turning couple. The thrust of the motor due to slip currents is obtained in the classical manner of rotary equivalent circuits: Motor thrust due to slip currents
3/fi *2i

! \

,->

normalised time Fig. 7 Distribution of attractive vertical force along motor length

(24)

Motor thrust due to eddy currents -3/m*21 (l-exp(-0)

Area under curve

(25)

= Q + 2 exp ( - 0 - 0.5 exp ( - 2 0 - 1.5 Moment of area about point of origin

(20)

fQ v n I X^I

(CXp(

^2 i
X) )

ClX

Jo

The motor thrust due to slip currents will go negative as the slip frequency changes sign to change the motor to the generating mode. The thrust due to eddy-current losses will change sign with a velocity change of sign, so as to always oppose motion. From a given set of terminal conditions, the output power, thrust, speed and power factor can be determined for any value of slip frequency which avoids saturation. In addition, eqns. 18 and 19 will provide the vertical force. Eqn. 23 will enable the turning couple to be determined.
Test results

= 0.5Q2 + 2 ( 0 + 1 ) exp ( - 0

(21)

- (0.50 distance of ( - 0 - 1.75 The effective + 0.25) exp the 2centroid of the area from the origin is obtained by dividing eqn. 21 by 20, to give: M=

A single-sided linear induction motor has been developed for an urban transit system funded by the Government of Ontario.
(22)

O.502 + 2 ( 0 + 1) exp ( - 0 - ( 0 . 5 0 + 0.25) exp ( - 2 0 -1.75 Q + 2 exp ( - 0 - 0.5 exp ( - 2 0 - 1.5

This distance M is in terms of the T2 scale used in Fig. 8 and must be multiplied by D/Q to be converted to metres. The turning couple C is then referred to the midpoint of the motor, as follows: C = (MD/Q- 0.5D)Fva =
4.5 Effective motor length
AOHz

(M/Q-0.5)DFv

(23)

The previous discussion assumes filled slots; however, Fig. 1 shows that the slots at the end of the primary contain only one side of a coil. These half-filled slots extend about twothirds of a pole pitch from each end, and are practically unavoidable. This effect can be allowed for, especially on motors of reasonable length, by selecting an effective motor length. Fig. 8 shows the MMF profile obtained in a 6pole motor. The total length of this motor is (6 + 2/3)r or, in general terms, (P + 2/3)r. To allow for the half-filled slots, the effective motor length can be taken as (P+ l/3)r
IEE PROC, Vol. 130, Pt. B, No. I, JANUARY 1983

2 10 12 U 16 motor speed , m/s 18 20

Fig. 9

Thrust against speed curves

calculated test Constant current 465 A Tooth to reaction railgap = 12.6 0 . 8 mm

55

Table 1 : Linear-motor design data

Motor length Stack width Slots per pole Pole pitch Number of slots Phases Wound poles Connection Turns per coil Coil pitch Iron behind slots Secondary back iron Width Depth Conductor Resistivity Thickness Mean iron-to-iron gap Equivalent circuit test values

1.9 m 0.216m 12 0.287 m 79 3 6 Wye 4 7 0.0384 m 3 mm laminated steel 0.240 m 0.0381 m Aluminium 3.5SVcmat 20C 4.5 mm 17.1 mm

and rail resistance observed on test. Fig. 11 shows the total motor vertical force for the same conditions as Fig. 10.
7 Discussion

0.038 SI at 20 C 0.00104 H 0.109 a at 14C 0.0002 H 0.00449 H

The test results show that the equivalent-circuit model provides reasonable predictions of performance. There were a number of factors which assisted, and some which prevented closer results. The back iron used in the secondary was laminated using 3 mm thick steel. It was deep enough to avoid saturation over the range of interest. The static tests showed that the equivalentcircuit constants were not significantly influenced by frequency. These factors lead to simplified modelling. This would not be the case with solid or saturated back iron. On the other hand, the configuration of the magnetic gap is constantly changing and this leads to inaccuracies. When the vehicle is accelerated, the gap tends to close up at the trailing end of the motor. This is due to the reaction of the truck suspension system to the vehicle's accelerating forces. As the motor speeds up, the flux tends to move towards the trailing end of the motor producing even more closing forces on this end. The combined effect of these forces is to produce a tapered gap which varies with load and speed. The combination of the tapered gap and the MMF
300r

320

o 1 240 "5
E 200 -

120
2 80 2 4 6 8 10 12 14 16 motor speed , m/s A 6 8 10 12 14 16 motor speed,m/s

18

20 22

Fig. 11

Motor vertical force against speed

Fig. 10

Motor-terminal voltage against speed

1 - calculated test Constant current = 465 A Tooth to reaction railgap = 12.6 mm

calculated test Constant current = 465 A Tooth to reaction railgap = 12.6 mm

This motor has been subject to extensive static and dynamic testing. The static tests were used to obtain the equivalentcircuit component values and the vertical force constants. These are listed in Table 1, and were obtained using normal rotary motor tests. The open-circuit test was simulated by removing the secondary sheet, and gap search coils were used to separate the primary and secondary leakage inductance. The dynamic tests were carried out while the motor was driving a prototype transit vehicle on a test track. The results of the dynamic tests are shown along with the calculated curves in Figs. 9, 10 and 1 ]. Fig. 9 shows the motor thrust plotted against speed for a constant current at various supply frequencies. To illustrate the effect of the changes in the gap and rail resistance, two calculated curves are shown. These have been calculated using the maximum range of the average airgap estimated from observations made during the test and the maximum range of secondary resistance derived from ambient temperature changes recorded during the test. The variation obtained in the calculated curves with gap changes indicates a major source of difficulty in obtaining test results for a motor mounted on a vehicle. Fig. 10 shows the phase terminal voltage plotted against motor speed for constant current at various supply frequencies. In this case, the calculated values are shown for the mean gap
56

distribution will tend to produce an effective gap reduction which is larger than the observed reduction. A difference of 2 mm was observed on tests between the leading and trailing ends of the motor. The tapered gap, in practice, will tend to produce higher values of thrust, vertical force and terminal voltage than those obtained from the model. This result will be accentuated at higher motor speeds. This may explain some of the differences between calculated and test results in Figs. 9 to 11. The uncertainties and difficulties of measuring the airgap and its lack of uniformity on a vehicle-mounted linear motor are formidable and may prove intractable. If this is the case, then linear-motor models may have to be checked against laboratory tests where the airgap can be controlled. It should be noted that a single-sided motor used in a transit vehicle will encounter changes to the airgap and rail resistance which are much higher than those obtained in these tests. The linear motor must, therefore, be designed to produce the required output under worst case conditions especially on automated systems, where reliable propulsion-system performance is vital.
8 Conclusions

The modified simple equivalent circuit provides a performance model for the linear induction motor. This model appears
IEEPROC, Vol. 130, Pt. B, No. 1, JANUARY 1983

to be of sufficient accuracy if consideration is given to the typical airgap variations obtainable in practice. It also provides a conceptual frame of reference related to the rotary model, which leads to an improved understanding of the performance of linear induction motors.

From eqn. 18 at zero velocity, 3 (27)

Acknowledgments

Evaluation of Ka, however, requires exact knowledge of the effective gap g and the influence of the slots. Ka can, however, be determined from static tests to be discussed in Appendix 11.3.
7 7.2 Repulsive forces

The test results used to assess the model were obtained by Spar Aerospace Ltd. and Metro Canada Ltd. for the Urban Transportation Development Corporation, who have kindly agreed to their use in this paper.

The formula which gives the force between two parallel circular wire loops is presented in many standard text books as F = 2nd

10

References

1 LAITHWAITE, E.R.: 'Linear electronic motors' (Mills & Boon, London,1971) 2 WILLIAMSON, S., and SMITH, A.C.: 'Field analysis for rotating induction machines and its relationship to the equivalent-circuit method', Proc. IEE, 1980, 127, (2), pp. 83-90 3 SKOBELEV, V.E.: 'Limiting factors in operation of linear motors at superhigh speeds', Rail Int., Feb. 1973, pp. 269-276 4 SKOBELEV, V.E.: 'Influence of longitudinal fringe effect on the operation of high-speed traction induction motor', ibid., Dec. 1974,pp.767-781 5 SKOBELEV, V.E.: 'The problem of using the asynchronous linear motor for high-speed ground transport', ibid., June 1977, pp. 297308 6 LAITHWAITE, E.R.: 'The goodness of a machine', Proc. IEE, 1965, 112, (3), pp. 538-541 7 WILLIAMS, F.C., LAITHWAITE, E.R., and PIGGOT, L.S.: 'Brushless variable-speed induction motors', ibid., 1957, 104A, pp. 102118

where i{, i2 are the loop currents, L is the length of wire in the loop and d is the distance between loops. In the case of the linear motor, the slip current I2l and the reflected secondary current I2 correspond to the loop currents i x , and i2 to provide eqn. 19:
Fvr Kr
'21

These currents are in opposite directions and so provide a repulsive force. The constant Kr presents a complex expression related to the coil geometry, but it may be obtained by means of a suitable test.
71.3 Test to determine Ka and Kr

11

Appendixes

When the motor velocity is zero, eqn. 18 becomes tva

This Section deals with the derivation of eqns. 18 and 19 and the method of establishing the constants Ka and Kr from test results.
7 7.7 Attractive forces

KaLmIm
**va tvr

The total attractive force is given by

The stored energy in the motor airgap at zero velocity and with no secondary sheet present is given by S =
'm1I m
2

KaLmIm

Kr

(28)

This energy is proportional to the effective magnetic gap g. If this gap is now changed by 8g, by means of a force Fva, the input energy due to this change will equal the change in the stored energy. Thus

rFT

va

LI T T ^ -'ml

(26)

The linear motor is restrained in a static condition with a variable-frequency supply current, and with facilities for measuring the total vertical force Fvt. A known current is now supplied at two different frequencies to obtain a significant difference in the relationship between Im and I2l. The force Fvt is measured at these two frequencies. From the equivalent circuit for velocity equals zero, it is now possible to determine Im and l2x for each of the applied frequencies. Eqn. 28 may therefore be repeated with different values for Fvt, Im and I2X. As Lm and d are already known, the two unknowns Ka and Kr may be determined. Eqn. 27 may provide a useful check on the effective

IEE PROC, Vol. 130, PL B, No. 1, JANUARY 1983

57