Professional Documents
Culture Documents
1977 Prater1977 Shaft Lining
1977 Prater1977 Shaft Lining
1977 Prater1977 Shaft Lining
Various theories for determining the earth pressure on shaft linings in cohesionless soils are
discussed, and results are presented for a Coulomb-type analysis with a conical sliding surface.
The assumed shape of the failure surface approximates closely the one given in published results
obtained by the method of characteristics. The simplicity of the cone permits an investigation ofa
number ofparameters. e.g. the earth pressure coefficient on radial planes, which turns out to be a
decisive parameter in the analysis, and accounts for the widely differing published values for
earth pressures on shaft linings. Certain theories c o ~ ~lead,
l d especially at greater depths. to rather
conservative designs.
A similartheory is also presented for earth pressureson shafts in cohesive soils. In this case the
possibility of base failure must be considered as well, and it is shown that this might be the
deciding failure mechanism.
Cet article presente une disc~~ssion des diverses theories relatives a la determination de la
poussee des terres sur les rev2tements de puits dans les sols pulverulents, et les resultats d'une
analyse de type Coulomb avec surface de glissement conique. La forme supposee de la surface de
rupture est une bonne approximation de celle d i j i publiee et obtenue par la methode des
caracteristiques. La simplicit6 du c8ne permet I'Ctude d'une quantiti de paramttres, tel que le
For personal use only.
coefficient de poussee des terres sur des plans radiaux qui s'avkre ttre le paramktre dtcisif de
l'analyse et e x p l i q ~ ~les
e grandes differences entre les valeurs publikes de poussCes des terres sur
des rev2tements de puits. Certaines theories peuvent conduire a des designs particulierement
conservateurs. specialement a grande profondeur.
Une theorie similaire est presentee pour les pressions des terres sus les revttements de puits en
milieu coherent. Dans ce cas la possibilite de rupture par soulkvement du fond doit egalement Ztre
considCree et on montre que $a peut 2tre le mecanisme qui controle la rupture.
[Traduit par la revue]
Can. Geotech. J., 14, 91 (1977)
rface
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by UNIVERSITY OF ARIZONA LIBRARY on 12/19/12
-
Shaft diameter D
Depth h 7.5 m
2- 1
= 3.55 m
""I
I
For personal use only.
E f f e c t i v e e s r t l ~prossure
lo)
Shaft dimeter 0 - 3.55 m
FIG. 4. Variation of the earth pressure for different shaft depths (after Steinfeld 1958).
ative portion of the pressure distribution given and the ratio h/r, which derives from [2] not
by Steinfeld's theory. What is less satisfactory, [I]. Schulz's results, as remarked by Walz
however, is the fact that he revises a part of (l973), do not lead in principle to a different
Steinfeld's theory which is correct, that is, the distribution of earth pressure.
relation between the angle of the slip surface It is assumed by Walz (1973) that Kara-
94 CAN. GEOTECH. J. VOL. 14, 1977
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by UNIVERSITY OF ARIZONA LIBRARY on 12/19/12
FIG.5. Earth pressure coefficient k , as a function of shaft geometry 11/r for different friction
angles 9 (after Karafiath, earth pressure at rest - Hungarian standards).
FIG.8. Relationship between shaft aspect ratio rz and the inclination of the failure surface or
+
for different friction angles and = 1 - sin a.
FIG.9. Relationship between depth factor h / r and the earth pressure coefficient k,, for
p = -+, h = 1 - sin % a n d A = 1.
For personal use only.
Active Rankine
Pressure at rest (plane wall KO)
Terzaghi
Berezantzev
Lorenz, X = 0
Steinfeld, X = 1
Karafiath, X = 0 . 5
Steinfeld, X = 0 . 5
The author, X = 0 . 5
Some results obtained by Karafiath were pre- results to changes in the coefficient h is made
sented in Fig. 5. The influence of the coeffi- clear by the results shown in Table 1, taken
cient h is clearly seen in Fig. 9. It was men- from Schulz's paper and modified. For all cases
tioned earlier in the paper that for h = K,,, the the following values are adopted: -y = 2.0
active earth pressure coefficient, k , is indepen- t/m3, (P = 30°, and r = 2.0 m.
dent of the aspect ratio n. This is inferred from It is evident that for realistic results h should
+
[I 11. By assuming that h = tan (a p)/tan CY be chosen in the range KJi < h < KO.The same
(= KJi for k = -1 and CY = 45' +(P/2), k , model is used by Steinfeld, Lorenz, Karafiath,
+
becomes tan (CY p)/tan CY and thus the earth and the author, but Lorenz by choosing h < KL1
pressure exhibits a hydrostatic form. If a value obtains a result about 15 times greater than
+
of less than tan (CY p)/tan CY is chosen, the with h = 1 - sin a, and 3.5 times greater than
earth pressure exceeds the active Rankine pres- the active Rankine pressure in the above ex-
sure. Lorenz (1966) set h equal to zero and ample for a depth of 20 m. The deviation from
obtained values of k , which are very probably the Rankine value increases more and more
much too conservative. The sensitivity of the with depth, while theories with h > K.\ show
CAN. GEOTECH. J. VOL. 14, 1977
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by UNIVERSITY OF ARIZONA LIBRARY on 12/19/12
FIG.10. The earth pressure force at different depths for a shaft with diameters 8 and 16 m
and h = 1 - sin 9.
a decrease in comparison with the active Ran- Let us suppose we want to investigate a
kine earth pressure. shaft of diameter 4 m and depth 20 m in a
The influence of X on the earth pressure is cohesionless soil characterized by = 2 t/m3
further brought out in Fig. 10 where the in- and @ = 30". The water table is assumed not
fluence of the radius, i.e. the arching action, is to enter the problem. We shall assume that the
also to be seen. It should be noted that the coefficient x = KO = 0.5.
force shown in Fig. 10 is the total force acting If we are interested in the extent of the
at each depth. The pressure distribution over failure zone we may resort to Fig. 8. For a
the depth must be calculated in a step wise depth h = 20 m (i.e. n = 0.1) the potential
manner. The construction for a shaft of 16 m slip surface is inclined at approximately 70.2'.
diameter is shown in Fig. 11, the smooth It can be shown by a graphical construction
curves being interpolations of the steps in the that this slip surface encloses all other poten-
pressure diagrams. A further worked example tial slip surfaces above it, and thus the assumed
is given below. The active Rankine pressure cone mechanism is valid for this depth.
lines are included for comparison purposes. A Now using Fig. 9 the value of h/r at which
comparison between the results obtained here theoretically the earth pressure force becomes
and those of Berezantzev (1958) and Lorenz zero is 12.75, i.e. at about 25.5 m depth. Fur-
(1966) is presented in Fig. 12. ther use of Fig. 9 shows that the earth pressure
coefficient k , corresponding to 20 m depth is
Worked Example 0.07. Thus, using [ l o ] ,the total force is 28
A guide to the use of the figures presented t/m circumference of shaft. If instead of the
for Part 1 of this paper is now given. First of total force the maximum pressure is required,
all it should be noted that the computations a graphical construction like the one shown in
for the shaft must be made for various depths Fig. 11 must be performed. Working in incre-
and the procedure is incremental. ments of (say) 2 m depth the force for each
PRATER 99
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by UNIVERSITY OF ARIZONA LIBRARY on 12/19/12
For personal use only.
increment is found as follows. Beginning with undrained, unconsolidated state, a state which
the top 2 m the force is calculated as above may exist directly after loading. This case does
and then divided by the depth increment to not previously seem to have been dealt with in
give the average pressure for the increment, the literature, but an approximate solution
and so on for each load difference between could be obtained using one of the arching
successive increments. The computational steps theories for trenches dug in purely cohesive
are outlined in Table 2. soils (Prater 1973). The same mode of failure
The incremental procedure shows that the is assumed here as was previously assumed
maximum force is not in fact 28 t/m, but with non-cohesive soils.
about 30.5 t/m. The maximum pressure of
about 2.7 t/m2 is reached at a depth of about
Method of Computation
The assumed mode of failure, with the
10 m, and zero pressure is obtained at about
forces acting on the sliding mass, is shown in
17.5 m depth.
Fig. 13. The following forces are known:
This example could be reworked using Fig.
12 which was constructed for the same param- Weight of the sliding mass
eters as were used in the worked example.
Though for other values of a, etc. Fig. 9 must
be used.
Cohesive force
Part 11. Earth Pressure on Shaft Linings
(7
-- active Rankine
------ Berazantzev
-- - Karafiath / Prster
[ ~ - ~ ~ - l - ~ i n + )
d
Failure surface
n* = (1 - tan2 .)/
(cot a -
6c
If A is set equal to one,3 the following simple
relation is obtained
1181 n* = tan ( 1 - tan2 a )
which is independent of the term c/yh. Sub-
stituting for n* in [17] yields the active earth
FIG. 12. Total earth pressure force per metre of
pressure (per unit length of shaft circum-
circumference according to different authors (as a ference)
function of the shaft aspect ratio).
[19] P = (yh2/2) (1 - N/sin or cos3 or)
The equilibrium equation at the point of where N is the stability number (after Taylor),
incipient failure is N = ~ / ~ Equation
h . 19 may be rewritten
TABLE
2. Computational steps
FIG.15. Determination of the critical depth for a shaft in purely cohesive soil with differeat
shaft diameters ( h = 1).
in the presence of tension cracks a reduction Some further examples are given in Fig. 16.
in the value of h, of one third is to be ex- These examples show that hComay be many
pected. times larger than h,, the critical depth for a
The critical depth, on the assumption A = 1, long trench.
may be estimated with the help of Fig. 15. For The evaluation of the above values for hco
a specific shaft diameter the critical depth is does not take into consideration the possibility
reached when the curve for the depth, with of a base failure in the shaft. A simple check
parameter r, intersects the curves constructed on base stability can be made by using the
on the basis of [22] with parameter c/y. The theory outlined below - cf. the plane strain
results may also be presented in a dimension- theory of Terzaghi (1943, pp. 189-194). It
less form (Fig. 16). This figure is used in the could alternatively be assumed that the failure
following manner: surfaces for active and passive conditions are
Step 1. Calculate 6,: 8, = arctan (ry / c). identical - this would not, of course, be true
Step 2. Construct the straight line OA. for non-cohesive soils - and then utilize a
Step 3. From the point of intersection with formula for the bearing capacity of a pile in
the curve determine (hCoy/c)and thus he$ clay.
The procedure is illustrated by the following With reference to Fig. 17, the vertical pres-
example: sure, level with the base of the shaft, i.e. on
r=4.2m c/y=1.5 the surface cd is
Thus ry/c = 2.8, = arctan (2.8)
[23] a = yh, -
2nr(l + $)hsc
and hCoy/c= 9 rr2([l + $1' - 1)
hco= 13.5 m (cf. h, = 4c/y = 6.0) provided that a hard layer of soil or rock does
PRATER
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by UNIVERSITY OF ARIZONA LIBRARY on 12/19/12
For personal use only.
FIG.16. Determination of the critical depth of a shaft in purely cohesive soil (dimensionless
curves).
not impede the formation of the failure sur- depth based on side failure hco and on base
face de. failure h, for different values of shaft radius r
For base failure and c / y values.
The results shown in Table 3 indicate that
the critical depth obtained by considerations
where N , is bearing capacity parameter. It is of base failure may be smaller than that for
usual to assume that N , lies within the limits side collapse. The smaller value is obviously
6 < N , < 9 for round foundations. the deciding one. Usually, however, the differ-
It follows from [23] and [24] that ence between the critical depth values is not
significant.
Figures 18 and 19 illustrate the influence of
the tangential stresses (for A = 1) using the
It is now possible to compare values of critical cone type failure to estimate the active earth
104 CAN. GEOTECH. J. VOL. 14, 1977
TABLE
3. Values of ciitical depth
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by UNIVERSITY OF ARIZONA LIBRARY on 12/19/12
Failure
surface
For personal use only.
earth pressure coefficient at rest (KO)when and Longinow (1964) have considered the
using the Coulomb method. case of a rough wall in their investigation of
The different forms of the pressure distri- the pressures acting on silos. They assume that
bution diagrams of Steinfeld and Befezantzev, the angle of wall friction is equal to the angle
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by UNIVERSITY OF ARIZONA LIBRARY on 12/19/12
drawn attention to by Walz (1973) has also of internal friction of the soil, in which case
been satisfactorily explained. The inconsistency the cylindrical surface becomes a characteristic
lies in the theory of Steinfeld (1958), wherein surface and the differential equations reduce to
it is incorrectly assumed that the earth pres- a simple form on the cylindrical surface. If
sure distribution may be obtained directly using wall friction is neglectcd, however, the results
the Coulomb method. In other words, the lie on the conservative side, and the above
method proposed by Steinfeld (illustrated in theory (or that of Berezantzev) may be ap-
Fig. 2 of this paper) is not acceptable, since plied. If linings to shafts and wells are driven
statically it is only possible to isolate the forces then it is possible that passive pressures may
shown in Fig. 1. Steinfeld partially corrects his be developed at the cutting edge. Only in these
pressure distribution diagram when he comes circumstances is it necessary to consider
to consider various depths of shaft (Fig. 4 ) . Lorenz's results, and then only for the cutting
Strictly speaking, however, his diagrams are shoe itself if a recessed lining is used (Fig. 6 ) .
incompatible: Figs. 3 and 4 should both give
the same pressure distribution for the same
shaft depth, were it possible to obtain the dis- Final Remarks
tribution directly (Fig. 3). The present paper has dealt mainly with
The Coulomb method indicates that below theoretical aspects of earth pressures in shaft
For personal use only.
a certain depth no earth pressure acts (Kara- linings. Some preliminary results of small-scale
fiath 1953a), whereas the limit equilibrium laboratory model tests in sand are reported by
theory used by Berezantzev (1958) does not Walz (1973) - work initiated by Prof.
give this result. The earth pressure according Lorenz.
to Berezantzev reaches a limiting value asymp- Case records where field measurements have
totically which albeit is much smaller than the been made are very scant in the literature.
active Rankine value at greater depths. In Kany (1972) does give some results for a 30
practice, there may in fact be a reduction of m diameter cylinder 40 m high. He found that
pressure at depths greater than some critical locally high pressures occurred when compact-
depth, due to arching action. A similar phe- ing the backfill material around the cylinder.
nomenon is known to exist for retaining walls These stresses, however, decreased during
not fulfilling the plasticity deformation condi- later functioning of the structure, and at the
tions (Terzaghi 1943, p. 66) and in slurry wall base of the cylinder the pressures were espe-
trenches (Prater 1973). For design purposes, cially small. This effect could be produced both
however, it would be advisable to ignore the by wall friction and the phenomenon of arching
reduction of pressure and dimension the shaft described in this paper.
lining for the maximum earth pressure (e.g. as If a shaft lining can deform to allow the
in Fig. 11 ) , as would in any case be common arching action to become fully effective it is
practice. If the pressure reduction is neglected, suggested that some account of reduced earth
Berezantzev's results and those obtained with pressure be taken according to the theories
the assumptions of Coulomb's method are referred to in this paper. But a considerable
comparable (Fig. 12). amount of engineering judgement is necessary
Certain investigators have simplified Bere- and for uncertain ground conditions the active
zantzev's theory of limit equilibrium. Walz Rankine pressures should probably be used,
(1973) has assumed that the slip lines are or some theory applicable to rough retaining
straight, corresponding to Rankine's pattern walls.
of slip lines. In this case the simultaneous par- NOTEADDED IN PROOF-A relevant paper by
tial differential equations for the stresses in the Walz appeared after the submission of this
plastic zone reduce to simple differential equa- work. Only the reference (see Walz 1976) and
tions, which may be easily solved. Costantino no discussion can be given here.
106 CAN. GEOTECH. J. VOL. 14, 1977
BEREZANTZEV, V. G. 1958. Earth pressure on the cylindri- wall subject to the pressure of a cohesionless loose
cal retaining walls. Brussels Conf. on Earth Pressure medium. 5th Eur. Conf. Soil Mech. Found. Eng., Ma-
Problems, 11, pp. 21-27. drid, pp. 95-99.
COSTANTINO, C. J. and LONGINOW, A. 1964. The theory of STEINFELD, K. 1952. Der raumliche Erdwiderstand von
limiting equilibrium for axisymmetric problems: a com- Kreiszylindem und seine Vergrosserung durch Riit-
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by UNIVERSITY OF ARIZONA LIBRARY on 12/19/12
parison with experiment on silo skin friction. Proc. telung. Dissertation, Tech. Hochschule, Hannover.
Symp. on Soil-Structure Interaction, Arizona, pp. 1958. Uber den Erddruck auf Schacht- und Brun-
583-592. nenwandungen. Contribution to the Foundation En-
KANY,M. 1972. Measurement of earth pressures on a gineering Meeting, Hamburg. German Soc. of Soil
cylinder 30 m in diameter (pump storage plant). 5th Eur. Mech. Found. Eng., pp. 111-126.
Conf. Soil Mech. Found. Eng., Madrid, pp. 535-542. TERZAGHI, K. 1936. A fundamental fallacy in earth pres-
KARAFIATH, L . 19530. Erddruck auf Wande mit kreisfor- sure computations. 111 Contributions to Soil Mechanics:
migem Querschnitt. Bauplanung und Bautechnik, Ber- 1925-1940, Boston Soc. Civ. Eng., 1940, pp. 277-294.
lin, 7, pp. 319-320. 1943. Theoretical soil mechanics. J. Wiley &Sons,
-19536. On some problems of earth pressure. Acta NY. pp. 202-215.
Tech. Acad. Hung., Budapest, pp. 327-337. WALZ,B. 1973. Apparatur zur Messung des raumlichen
LORENZ, H. 1966. Offene Senkkasten. Grundbautaschen- Erddruckes auf einen runden Modell-Senkkasten.
buch. W. Emst & Son, pp. 795-798. Baumaschine und Bautechnik, 20, pp. 339-344.
PRATER,E . G. 1973. Die Gewolbewirkung der Schlitz- 1976. Active soil pressure on a cylindrical caisson
wande. Der Bauingenieur, 48, pp. 125-131. compared with model measurements. (In German.) 6th
SCHULZ, M. 1970. Berechnungdes raumlichen Erddruckes Eur. Conf. Soil Mech. Found. Eng., Vienna, 3, pp.
auf die Wandung kreiszylindrischer Kolper. Disserta- 669-672.
tion, University of Stuttgart. WESTERGAARD, H. M. 1940. Plastic state of stress arounda
SNARSKY, A. S. 1972. Design of an axisymmetric retaining deep well. J. Boston Soc. Civ. Eng. 27, pp. 1-5.
For personal use only.
This article has been cited by:
1. Tatiana Tobar, Mohamed A. Meguid. 2011. Experimental Study of the Earth Pressure Distribution on Cylindrical Shafts.
Journal of Geotechnical and Geoenvironmental Engineering 137:11, 1121. [CrossRef]
2. Y. M. CHENG, S. K. AU, Y. Y. HU, W. B. WEI. 2008. ACTIVE PRESSURE FOR CIRCULAR CUT WITH
BEREZANTZEV'S AND PRATER'S THEORIES, NUMERICAL MODELING AND FIELD MEASUREMENTS. SOILS
AND FOUNDATIONS 48:5, 621-631. [CrossRef]
3. Y. M. Cheng, Y. Y. Hu, W. B. Wei. 2007. General Axisymmetric Active Earth Pressure by Method of Characteristics—
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by UNIVERSITY OF ARIZONA LIBRARY on 12/19/12