Asimse rinc siearaooara:Us

by E. N. BROMHEAD", MSc, DIC, BSc, CEng, MICE

THE HITHERTO EXPENSIVE and time- Fig. 1 shows both the prototype and pro- wheel clutch and handwheel were there-
consuming nature of ring shear tests to duction shear cells in partly sectioned fore added to the worm drive in the pro-
determine the residual strength of soils elevation. ducti'on device.
has prevented the test from becoming a During the initial trials, it was found that
routine procedure in commercial labora- Sample preparation with the ariginel "built up" design for the
tories. This has had the result that the The lower porous platen and confin- confining rings, platens and centring pin,
ideal "three-pronged attack" on landslip ing rings are assembled and a remoulded it was almost inevitable that cumulative
problems by sample kneaded into the annular cavity. manufacturing tolerances gave rise to an
(a) back analysis, The top of the sample is then struck off eccentricity in alignment. Accordingly, the
(b) ring shear tests on representative level with the top of the confining rings, production machines are built around the
strata, and and the upper platen placed in position, central drive shaft and this problem does
(c) direct shear tests on actual slip sur- located on the centring pin, which is well not arise. The opportunity was taken to
faces, greased. With the water bath (which is a combine the confining rings and lower
is rarely fully developed. Since a number push fit over the '0'-ring seal) in position, platen and also to improve the general
of studies have shown that the ring shear the sample is flooded with distilled water. drainage ducts. This composite assembly
tests may possibly provide the lower If a very soft sample has 'been prepared, is positively located on the central shaft
bound strengths in any investigation, this it is quite important to allow time for it and held down by knurled nuts. Sample
omission is of concern. to come into equilibrium under the load preparation is aided by the simplicity with
However, ring shear testing cannot be- of the upper platen and then under the which the platen/confin'ing ring assembly
come a routine procedure outside the re- load of each item af the lever loading sys- can be removed from the water bath.
search laboratory until a simple, robust, tem, since if too much load is applied too For tests at low normal stresses the tor
inexpensive apparatus which has a fairly quickly there may be excessive 'squeeze'r que transmitted through the sample may
large potential through-put of tests in the loss of soil through the clearance be- be quite small and it was found that the
working week is developed. This Paper tween the upper platen and confining proving rings lacked sufficient sensitivity.
describes such a device, which has re- rings. However, with some soils, it may The production units have the proving
cently been built and evaluated at the be preferable to get a load on to the rings mounted in turrets with quick release
Schoal of Civil Engineering, Kingston Poly- sample to prevent swelling. clamps so that the effective length of the
technic and which in its fully developed The sample is then consolidated under torque arm can be changed at will, and
form is now avail'a'b'le commercially. loads which provide the desired normal the torque arm was redesigned with posi-
effective stress on horizontal planes, and tive stops at tw'o positions so that the
Description and general principle the coefficient:of consolidation can be rings could be brought to bear at similar
An annular soil sample Smm thick evaluated by monitoring the vertical set- radii without measurement. These turrets
with inner and outer diameters of 70mm tlement. When the soil is fully consoli- also all'ow a variety of load cells to be
and 100mm respectively is confined radi- dated it may be sheared at the appropriate fitted, and they swing out of the way for
ally between concentric rings. It is com- rate. sample preparation.
Other detail design improvements in-
pressed vertically between porous bronze
loading platens by means of a lever load- Developments from the prototype clude a more compact drive (again iden-
ing system and dead weights. It was found that the absence of a clutch tical with that for the appropriate pro-
In the prototype this lever system was mechanism in the prototype led to con- prietary direct shear machine) and the
adapted from a proprietary 5:1 ratio hanger siderable difficulty being experienced with provision of a scale 'of degrees so that the
system complete with lever levelling jack setting up and with gear changing. The amaunt of rotation can be seen at a
originally part of a direct shear apparatus. lack of manual control over the rotation glance. Fig. 2 is a photograph showing the
This system was, however, deleted from was another source of annoyance. A pin general arrangement.
that manufacturer's range along with the
machine in question and replaced by a new
Production
machine with a 10:1 ratio counterweighted
system. Accordingly, the production mach- Torque arm/loading platten assembly
ines were adapted to use the new com- Prototype
ponents.
A rotation is imparted to the base plate Lifting lug l2l with
locating studs and
and lower platen by means of a variable retaining nuts at 90
speed motor and gearbox driving through
a worm drive. This causes the sample to
shear, the shear surface forming close to Samp
conta
the upper platen (which is artificially
roughened to prevent slip at the ptaten/ ~Central ising Pin
soil interface) . Water bath
The settlement of the upper platen dur-
ing consolidation or shear can be mont- Protractor
tored by means of a sensitive dial gauge
/~
drum
bearing on the top of the load hanger.
Torque transmitted through the sample r" )
is reacted by a pair of matched load mea-
suring devices (sensitive proving rings in
3i
Confining rings
I'. 'iI
Turntable
the prototype) bearing on a cross arm.

Fig. f. Partly sectioned general elevations comparing the prototype and production
'Senior Lecturer, Kingston Polytechnic, Kingston
models
upon Thames, Surrey

40 Ground Engineering
Interpretation of results
Bishop et al. (1972) discuss the influence
of a variety of distributions of shear stress
across the sample in a ring shear test on
the measured torque and hence on the
angle of shearing resistance calculated
assuming a uniform shear stress distri-
bution. In some cases, errors of about 10%
may arise. It is, however, much more like-
ly that the distribution is close to the
uniform one particularly where the sam-
ple is narrow in comparison to its dia-
meter.
Conventionally, we take a uniform shear
stress when residual strength has 'been
reached and hence the torque T transmit-
ted through the sample is given by:

T = z» w (R, —R,»)»

where R, and R, are respectively inner

and outer sample radii.
(It should be noted that the non-uniform
strain distribution greatly encourages pro-
gressive failure in brittle soils and it is
therefore unlikely that any ring shear ap-
paratus will give a reliable peak shear
strength. The brittleness index cannot
therefore be fully defined on the basis of
ring shear tests alone).
Since the torque is given by the mean
load on the proving rings (or load cells) Fig. 2. Photograph showing general arrangement of production machines
multiplied by the distance between them:

i.e. T = (F, + F,)L/2

Thus: Differences in the stiffnesses of the (and double drained) that extremely rapid
3 (F, + F»)L proving rings can throw an unbalanced rates of shear are possible while still al-
T load onto the centring pin and cause fric- lowing full drainage to occur. However,
4m. (R»' R,») tion. This will also be affected by uneven higher rates of shear may induce rheo-
settlement of the loading platen. How- logical or viscous shear strength elements.
The normal effective stress 0's given ever, such friction will have two compon-
ents — one tending to reduce the n'ormal
One meth'od of ensuring that the re-
sults are o'btained in terms of effective
by:
load carried by the soil sample and the stresses is to test so slowly that the sam-
P other tending to increase the torque trans- ple is fully drained at all times through
mitted to the proving rings. The effects the peak (which is of course a remoulded
m (R» R») are therefore largely s"If-compensating, as peak since even if an "undisturbed" sam-
in fact are all the similar types of side ple is used, its upper surface is rernoulded
where P is the total vertical load friction noted as (b) and (c) above. Fur- during preparation) and on to the resi-
thermore, the likely uribalanced loads will dual state. This is a fairly slow process
Hence: only give rise to small amounts of torque even in the present device.
due to friction if the centring pin is light- An alternative and more rapid procedure
'
3 (»+F») (R» Rr )L ly greased and of an appropriate diameter. is as follows:
r/»»'— tan y The possibility of tilting and its resulting
4 (R»» —R,»)P (i) Consolidate the sample to the first
effects is reduced by the presence of the normal load, and monitor settlement
centring pin. In the prototype the taper to determine the time for full consoli-
Sources of error of the loading platen was designed to dation (t~).
The errors that can arise in the use of prevent 'squeezed'oil from being trapped (ii) Shear the sample undrained for about
this apparatus do so from: but was found to be an unnecessary re- 5 revolutions to form a shear surface;
(a) inaccuracies in load measurement, finement. Soil extruded from the outside release the torque.
(b) friction on the centring pin, of the sample can of course be removed (iii) Reconsolidate to rid the sample of
(c) binding between the upper platen during the test with a soft brush or a jet undrained pore water pressures from
and the confining rings due to tilting, of water from a wash bottle. (ii) above; the value for tb can be
or 'squeezed'oil becoming trapped, In practice it is found that the main slip updated from this stage in some
(d) side friction if the slip surface forms surface forms close to the upper platen, tests.
below the top platen/soil interface, but that other slip surfaces can and do (iv) Select a shearing rate and shear the
and form more deeply in the soil. It seems un- sample on the pre-formed slip surface
(e) the possible non-uniform mobilisation likely, therefore, that side friction on the A flat topped stress strain curve us-
of shear stress across the sample. confining rings is a significant factor. ually results and providing that the
The apparatus was originally designed Smear type ring shear devices, particu- sample is sheared for a time at least
to use commercial grade proving rings and larly tHose of very simple construction as long as ts and the measured shear
these are of course quite adequate for are apparently more subject to these re- stress d'oes not change we can be
most purposes. Their accuracy is fairly latively unquantifiable effects than the sure that the sample is fully drained.
limited at low loads and for tests on very Imperial College device. However, a sys- The shear rate can be changed at this
weak materials or at very low normal ef- tematic study by La Gatta (1970) showed point to investigate viscous effects.
fective stress it may be necessary to resort that the largest potential source of error (v) The sample is then re-consolidated
to more sensitive electrical load cells, but was in the assumption of uniform mobi- to the next normal effective stress,
an alternative method of overcoming the lisation of shear stress ~on the slip surface, providing a new value for ts, and
problem is to adjust the proving rings so and this is also found to be so in the the procedure of step (iv) fol'lowed.
that they react closer to the axis of the present case. This process is quite simply explained
device. This increases the leads to be car- to technicians with minimal skills and has
ried and brings them into a better part of Selection of appropriate strain rate been found by experience to give results
the proving ring's working range. In this apparatus, the sample is so thin identical with those of the first method.
42 Ground Engi'neering
 0.35
1
q
Failure planes adjacent
~e imperial College tests tp pile shafts
0.30— ~5 Kingston polytechnic tests showing range
& No. of tests at each stress level
(continued from page 39)
- 16
stress on AC is applied, then in the limit
i~ h hear stress tends to zero the
oe 025— - 14
element ACDB may be regarded as bein eing
4 in an incipient state of failure. The incipi-
0
0 5 ent state is represented by the failure
5 4 12 ID
circle with centre 0" and failure point
0.20— F" 1b.
in Fig. It is clear that for this case
'8
K is equal to K,, the coefficient of pass-

—
10 ive earth P ressure,
2c
0.15— ~ + '=Kv=N
N
V

0 where N = tan'45 + d'/2)

0.10 I I o i o I
- 6
0 200 400 600 800 The bounds for K are therefore 1<K<K,.
Normal effective stress kN/m 2 The intermediate case is represented
'
in Fig. 1b by the failure circle with centre
Fig. 3. Comparison of results of rin shear tests on Gaultu c/ay (from the formation at 0 and with failure point F. The centre of
Folkestone Warren) in the Imperial Co//ege andnd tthee presen
resent machines the failure circle 0 can take any position
bbetween 0'nd 0"
corresponding to
the upper and lower bounds. Its actua actual
position for any given problem will de-
pend on t'e assumption of the inclination

Volume changes during shear I h b en found that the roughness )or fissured clays the failure plane may
S ettl e ment of the upper loading platen the ~pp~r plat~~ is 8 significant f~~t~~

',
be ~98~m~ o P h d
h' may be morritored by means of t if the sail has an appreciable silt fraction,
( ant direction of the orientation o t e
single sensitive vertical dial gauge) u or
two components: the dilation or contrac- It may be similarly s own a
tion of the soil due to shear and the loss Condusions soils where K,<1 the values of K will
of material due to 'squeeze'. In practice sjm be bounded such that K„<K<1, where
it is impossible to separate these two. K i th t of cti
pie in design and inexpensi e in co s r c- ffi i rth

Practical experience
Proving tests have been car ied r'ut
cia s, from
arj
on raP'ore
o . Tests on both remoul ed and
Is ed'amples ca be a e q i e
'th those
sophisticated
obtained from much
equipment.
Comparing the construction with a dir
se
ure d a selection as fo fissured stiff I
canno t be made. In the interim it has be en
(1g7
„
ImPerial College ring shear mac ine s ect shear (shear box) apparatus, it will
b h h h h i ommon. appears
Although the ring shear has some extra
Folkestone e
m onents,
comPO the capital costs for the two
d ic s are broadly similar, and due to o the rom ese
simplicity of operation of the ring rin sshear, surmise a,
n costs need not be excessive. for
In view of this it is expected a r'
shear testing to determine the residual K
that rin tr ngth parameters of soils will become
a routine operation in commercial te s t'in t e ini ia
shear test results tended to underestimate
the field mobilised shear strength. It ae.and as
own, its values depend, amonost
other factors, on the determination of the

'-
imi
Howeve,withth
in
p
records.
e and len 9th y
i

a full ring shear

Acknowl amen
Thee p rototyPe was bui 'I t in
s
'h e School.
f 'I re plane. For heavily fissured clays
a pre erre
of Civil Engineering, ingston oy may bee p resent. In the application of the
I I b M R St S io Tch i-
e rin shear tes cian, and a number of t h e P roving tests
', stu y

'f
A. H M .

ri htl
ve the hi hest liquid limit.
be recognised as absur The authorh
rfo
R B. Higginson
d b
and
M

'ateful
is gra e u to Mr
Mr...J.
. N. D. Lambert.
am
F. Lupini
ress-

to correlate field conditions wit a sin-

i 'es, or for that matter with
and ro esaor
ia'I College
.
for permission o q
h t al f the comparative test. He would
a d
erties over anY
similar situation in relation to
s'ike
g
I

to acknowledge the help of the design

f. Burland, J. B. (1973): "Shaft ric ion o

i
staff of W y keha Farrance Engi ering
shear
h tests
t t should
s ou be recognised. It is
surely time t o a ccord the ring shear test
Ltd ~specially that of e ing. Mr..
ive stress methods of calculating skin friction
equ a'I treatment. Two tests ove g n iles in soft clay". Ground En-
qineering, Vol. 10, N o.. 3 April 1977 pp.
0 f four or five normal stresses
G. E., Ger ga, V, K.,
uahy be car i
I971li 'A new 3. Parry, R. H. G, fk
ing week. nd s Ii o o h d skin friction on piles in s
ren th'.'. Geotechnique
i i

e p m
measurement
.
of residual strengt eo ec ni Gro d E i ee n, o
73-328. pp. 33-36.
R. L., si Parlay,
by rotst'on she~~ t~est D J 1969) "Joi ts d fi s e i on-
86
nickel were trie, e P " " arvsrd Soil MechasnicsS
bridge, Massachusetts.
No C iio Cf

cOme by the use Of ffat ChrOmed braSS, dd' l observations on the

en'.. tion of K —
thee lateral earth pressure co-
coastal landslides of Folkestone Warren'. Q. J1. efficient for s a ric i
brOnZe Or Stain'leSS Steel aS apprOpriate. Eng. Geol in press.
44 Ground Engineering