You are on page 1of 14

Proceedings of Plastics Pipes XI

Munich, Germany, 3rd - 6th September 2001

Soil-Structure-Pipe Interaction with Particular Reference to Ground Movement Induced Failures


J.L. OLLIFF and S.J. ROLFE
Montomery Watson

D.C WIJEYESEKERA and J.T. REGINOLD


University of East London

ABSTRACT
Soil-pipe interaction studies generally recognise the significance of deformations in the pipe due to soil loading, but not differential ground and structure movements, which can induce excessive stress concentrations in the pipeline. Plastics pipes can suffer failure due to such movements, though their flexibility makes them less vulnerable than rigid pipes. This study examines the problems of interaction between pipelines and the surrounding soil medium, subjected to differential ground movement.

A procedure for predicting pipeline settlements, based on strip foundation theory is presented, and a detailed analysis is made of failures encountered by a group of submarine plastics pipelines.

INTRODUCTION
Structural pipeline design is a subject that has been in and out of fashion over the years. The first description of the behaviour of buried flexible pipelines under load was given by Clarke(1) in 1897. Most of the research work in the next 40 years was on the behaviour and design of rigid pipelines in the cross-sectional direction, most famously by Marston, Spangler(2) and Schlick. The first analysis of flexible pipe behaviour taking account of both soil and pipe stiffness was published by Lazard(3) in 1935. The behaviour of a buried pipeline will depend very much on how its stiffness compares with the stiffness of the native soil in which it is to be buried. Although traditionally rigid materials are thought of as being concrete, clay and asbestos cement, and flexible materials are thought of as being the plastics, the differentiation is not as simple as this. The response of the pipes under load will be largely dependent on the behaviour of the native soil. If the pipes are of medium stiffness (say 20,000N/m2), and buried in a stiff

PREVIOUS PAPER | CONTENTS | INDEX | SEARCH | NEXT PAPER

1128

Plastics Pipes XI

Fig. 1 Soil at the sides of the pipeline compresses more than soil over the pipeline, and due to friction, increases the load on the pipeline.

soil (such as a dense gravel of Es , soil modulus 150MN/m2, K, soil stiffness = 111 MN/m2) then the pipes will exhibit predominantly flexible behaviour, i.e. they will tend to deflect on loading. If, however, the same 20,000 N/m2 stiffness pipes are buried in a soft soil (such as a very soft clay of Es 15 MN/m2, K 11 MN/m2), then the pipes will exhibit predominantly rigid behaviour, i.e. they will tend to settle into their foundation on loading. Pipes exhibiting rigid behaviour are those which attract a backfill pressure which is higher than the overburden stress ( * H) value. In analysis, the overburden pressure on a rigid pipeline, H, is multiplied by a load concentration factor, C1 (closely related to the CC used in the U.K. and American tradition), which is greater than unity. The reason for the load on a rigid pipeline being greater than gH is because the soil to the sides of the pipes in the trench tends to compress more than the pipes themselves, and hence by friction, additional load is placed on the column of soil above the pipeline, see Figure 1. Equilibrium of the pipeline demands that the vertical load (pressure multiplied by diameter) is matched by the foundation reaction, and the increase in vertical load on the pipes above the overburden value therefore increases the pressure on the foundation soil beyond the previously existing overburden value. In addition, this load is transferred to the underlying soil only over a small arc at the base of the pipeline, meaning that the greater load distributed over a smaller area inevitably leads to the settlement of the pipeline into the underlying soil. The resulting settlement of the pipes into the foundation soil, in turn reduces the value of C1 until a new equilibrium is reached. This process of interaction was taken account of by Marston and Spangler in their settlement-deflection
PREVIOUS PAPER | CONTENTS | INDEX | SEARCH | NEXT PAPER

Soil-Structure-Pipe Interaction with Particular Reference to Ground Movement

1129

ratio (rsd), which included a term (sf) denoting the settlement of the pipeline into its foundation.

TRADITIONAL APPROACH
Generally speaking, if structural pipeline design is done at all, it is limited to the cross-sectional design. Settlement is rarely, if ever, considered, and when it is, the chances are that the chosen method of analysis will be a weight comparison calculation. In this traditional approach the weights of the pipeline full of water, the pipe embedment and the backfill are compared with the weight of the soil mass they replace in the trench, with the difference in weights leading to a settlement prediction. This approach is wrong because it overlooks one crucially important point - the pressure redistribution. As an analogy, why must ladies not wear stiletto heels in halls with wooden floors? Because the heels mark the floors. Why do the heels mark the floors? Are the ladies too heavy to walk on the floor? The reason the heels mark the floors, of course, is because the ladies' weight is transferred to the floor over a very small area. The same principle applies to pipelines. As well as attracting a pressure greater than the previous overburden value, which was present before the excavations began, stiff pipelines often transmit their load to the underlying soil over a width which is less than the pipeline diameter. Hence the settlement which occurs is likely to be much greater in magnitude than the value predicted from a weight balance approach, which assumes uniform pressure distribution on the trench bottom. There are a number of ways of mitigating the effects of pipeline settlement. Both Barnard4 and Leonhardt5 assumed a zone of influence around a buried pipeline, in which the pipe-soil interaction occurred, extending a maximum of 2 pipe diameters from the pipeline in every direction. If the native soil in this zone was very soft, therefore, it could be removed to a depth of 2 pipe diameters beneath the pipeline, and replaced with an incompressible material in order to minimise settlement. This might be a reasonable solution with a small diameter pipeline, but with large diameter pipelines the cost could be prohibitive. Rigid, cement mortar pipe joints, as widely used in the nineteenth century, and often well into the twentieth century, were eventually realised frequently to crack. The solution was the adoption of flexible mechanical joints, sealed by rubber rings, and joints of this type were widely adopted for use with PVC and GRP pipes. The success of such joints in solving many problems led most engineers to believe they could solve all problems. This resulted, for example, in recommendations that two such flexible joints should be provided, close together, where pipelines approach and enter structures.6 The mistaken belief was that these two flexible joints, and the short length rocker pipe between them, would satisfactorily accommodate differential settlement between the pipeline and the structure. What was overlooked, of course, was that the shorter the rocker pipes were made, the greater would be the angles through which their joints would have

PREVIOUS PAPER | CONTENTS | INDEX | SEARCH | NEXT PAPER

1130

Plastics Pipes XI

to rotate to accommodate a given amount of differential settlement. Short length may therefore have reduced the risk of longitudinal bending failures in rocker pipes, but it did this at the expense of increasing the risk of joint damage. The vulnerability of short pipes to damage by settlement is well known, and is frequently observed in closedcircuit television surveys of clay pipe sewers.7 The mechanism by which damage occurs has been described by Rolfe.8 The traditional approach therefore exacerbates the problems of differential settlement. At the same time, the trend towards higher ring stiffnesses which has been followed in the design of some plastics pipes, for example GRP and profile-wall polyethylene, has brought these pipes into the semi-rigid category. Such pipes, when installed in soft soils, then attract backfill pressure concentrations which cause them to settle into their foundation soils. The abilities of different types of pipe to accommodate the shear forces, bending moments, curvatures and joint rotations caused by settlement and differential settlement, vary greatly. This variation is evident between the different types of plastics pipe. The low elongation at break of GRP pipes makes them unsuitable for accommodating the longitudinal curvature which tends to develop as pipelines try to follow ground settlements. Polyethylene pipes, on the other hand, having very large elongations, are ideally suited for use where large settlements, and differential settlements, will occur. The widespread lack of awareness of these problems, amongst consulting engineers, contractors, and even some pipe manufacturers, often results in the adoption, or at least proposal, of inappropriate pipe materials. The consequent failure, delays, and extra costs, are harmful to all parties.

STRIP FOUNDATION APPROACH (OLLIFF 1994)9


Awareness of the patently incorrect assumption of uniform pressure distribution implicit in the traditional approach, led the first two authors' company to develop a more rational, and safer approach. This approach consists of taking the vertical soil load on the pipeline, as calculated for the normal cross-sectional design, and estimating the resulting settlements, by analysing the pipeline as a strip foundation responding to that load. The procedure was presented to the CEN committee attempting to draft a "common European method" for structural pipeline design in 1994.9 The essential steps in the analysis are as follows: 1. Calculate pipe-soil stiffness ratio (e.g by Greatorex10, 11) using Eqn. 1

n = E / (105 S + 0.8 E )
0.48

(1) (2)

2. Estimate vertical soil pressure concentrations factor using Eqn. 2

C1 =1 + 0.585 (1 n )

0.0875 (1 n ) D H

PREVIOUS PAPER | CONTENTS | INDEX | SEARCH | NEXT PAPER

Soil-Structure-Pipe Interaction with Particular Reference to Ground Movement

1131

This formula closely reproduces the rigorous Marston-Spangler result, by combining Greatorex's solution (from Eqn. 1) for the settlement-deflection ratio, with the concentration factor proposed by Olliff.12 3. Calculate elastic settlement using Eqn. 3a (Spangler based methods) or Eqn. 3b (ATV/ ONORM based methods).

e = (C1 1)(2 T D ) H D E e = (C1 1) H D 0.6 E4


4. Calculate long-term consolidation settlement in cohesive soils using Eqn. 4

(3a) (3b)

c = 0.009 (LL 10 )T log (1 + C1 ) / (1 + eo )

(4)

This approach was first proposed by Olliff(9), and early verification of this approach was obtained by comparing predicted elastic settlements with observed initial settlements of a concrete pipe sewer in Hong Kong. The settlement predicted by Eqn. 3a was 180 mm, whilst observed settlements averaged 134 mm, but reached a maximum of 275 mm.

BEAM EFFECTS
The differential settlement between two points, distance l apart, is expressed in terms of angular distortion . This should not exceed the allowable angular distortion all.

max =

s1 s2 s2 s3 all l1 l2

(5)

Foundation settlements depend on the compressibility of the foundation, and the magnitude of the pressure concentration. The influence of these is likely to vary along the length of a pipeline, resulting in a variable settlement profile. These differential settlements cause bending moments in the pipe barrels, shear forces across the pipe joints, and angular rotation of the joints. Frequently, adverse bending moments and shears induced by differential settlement arise at positions where the pipelines approach structures. Most pipe joints can accommodate angular movements of 0.5 to 4.0, so that the amount of differential settlement a rocker pipe can accommodate is its length multiplied by the tangent of a relatively small angle which may amount only to a few millimetres. Furthermore, the settlements adjacent to structures can be relatively much larger, and accentuated by loose backfill material beneath the pipeline, or raised ground levels in the formation around the structure. Vertical displacement of rocker pipe = L Sin (6) where L is the length of the rocker, and the joint rotation.
PREVIOUS PAPER | CONTENTS | INDEX | SEARCH | NEXT PAPER

1132

Plastics Pipes XI

This relationship suggests that longer pipes are desirable to accommodate large differential settlements, but increasing the length of the pipe will also increase the shear forces at the joints, and the bedding moment in the pipe. The first free pipe can be regarded as a simply supported beam carrying a uniformly distributed load, in which case the maximum bending moment will be: M = 0.125 Pv.da.L2 (7)

In the case of flexible pipes, the limiting bending moment is more likely to be controlled by consideration of stability against buckling, than bending stress. According to Lundquist,13 the critical longitudinal bending moment for a pipe is given by the following formula: (8) Mc = K.E.rm.t2/(1-v2) The designer must first decide how much settlement is acceptable. This should include consideration of the following possible constraints: The bending strength of the pipes. (Note: the product standards for clay, concrete and fibre - cement pipes all include minimum requirements for the moment of resistance of small diameter pipes), The shear strength of the pipes. (Note: this may be particularly significant for the socket of mechanical joints and for butt fusion joints in polyethylene pipelines) and Leak - tightness of mechanical joints. (Note: some pipe product standards include requirements for shear force resistance whilst remaining watertight).

CALCULATION OF SETTLEMENT
Following Selvadurai14 (1984) and Fletcher and Herrmann15 (1971) we have the following for near surface and deeply buried pipelines
Authors Shallow Foundation Deep Foundation

Selvadurai (1984)

Fletcher and Herrmann (1971)


C (v ) 0.65 1 + v 2 K C (v ) E S

J.L.Olliff (1994)
Foundation Settlement

K=
K

0.65 E s 2 1 vS

) )

Z = (C1 1.0 )(2 T D0 )H D0 E 3


Where T = Thickness of bedding material under pipes

6 E S (1 + v S ) K C (v ) E S (1 v S )(3 4vS ) C (v ) 0.60 3 + 2v 2

The foundation stiffness parameter K depends on the elastic constants, Elastic modulus and Poissons ratio (ES and S) of the soil medium surrounding pipeline. It must be emphasised that these approximate expressions are valid for situations in which the soil/pipe material modular ratio satisfies the constraint (Es/E) < 0.01. Also, in these expressions the material parameters encountered are the elastic modulus and Poisson's

PREVIOUS PAPER | CONTENTS | INDEX | SEARCH | NEXT PAPER

Soil-Structure-Pipe Interaction with Particular Reference to Ground Movement

1133

ratio of the soil surrounding the buried pipeline. The assumed linear elastic behaviour will be valid only for soils, which are strictly valid for small strains within the vicinity of the buried pipeline. It may be noted that although the ground displacements are large the ground strains around the pipeline region are expected to be small. The linear elastic moduli for the soil medium can be estimated from results of triaxial tests conducted on samples of the backfill material or the undistributed soil, Winterkorn and Fang16 (1974). The cell pressure in the triaxial test is taken as H0 where is the unit weight of the backfill or the natural soil. Typical results for Es and ns are given by Bowles17 (1977) and Selvadurai18 (1979).

CASE STUDY
This case study presents the failure, at the point of commissioning, of a network of offshore GRP pipelines, serving as intake pipelines installed in a powerhouse project. Information was gathered by one of the authors, who was employed by the contractors on the project, but the exact location details of the case study are not disclosed. Sandy silts and occasionally sands and silty sands cover the seabed which gently deepens reaching a maximum depth of 10 metres over the area that was actually investigated. Three main lithological units occur in the submarine strata over a depth of 40 m from the sea bed. Borehole investigations were carried out along the line of intake pipelines. The nearest borehole to the shoreline was located at a distance of approximately 500 m, which is similar to the soil properties of the near shore pump house foundation area. The field geological data together with the interpreted logs from seven boreholes indicate that the thickness and nature of the superficial and channel deposits vary over very short distances, characteristic of the near shore depositional processes in a channel. An uneven nature of the bedrock (marl) was observed from seismic surveys and showed the presence of channel deposits overlying the marl. The geotechnical investigations revealed that the superficial deposits are mainly composed of loose to very loose, soft to very soft dark grey clayey sandy silts with an abundance of seaweeds and shells in some places. These gave unfavourable geotechnical conditions with highly compressible soils having almost no cohesion and very low angles of friction. The channel deposits were a more favourable formation than the overlying superficial deposits due to the presence of cemented horizons with an absence of fines and organic material. The dark grey marl bedrock was therefore a relatively homogeneous formation with reasonable geotechnical characteristics. The presence also of a Khaki marl with higher plasticity and compressibility than the dark grey hard marl occurred in the transition zone accentuating the threat of differential settlement. The geotechnical properties for the superficial deposits overlying the marly bedrock is summarised as follows for the purpose of this paper:

PREVIOUS PAPER | CONTENTS | INDEX | SEARCH | NEXT PAPER

1134

Plastics Pipes XI

Consolidation characteristics of Cc = 0.46 mv = 0.6 1.5 m2/MN, Es = 0.6 -1.7MN/m2. Unconfined compressive strength qu = 0.290.35 MPa. Along the transition zone in the vicinity of the pumphouse, the trench for the pipeline was excavated to the marly bedrock. The minimum depth of the trench was 200 mm below the elevation of the pipeline invert. Ground investigations suggested that the required excavation to reach the bedrock level would not exceed 300 mm below the base of the pipeline, and in most cases, some excavation of the marl was required in order to satisfy the minimum bedding thickness requirement of 200 mm. The Filament wound GRP pipes of stiffness 2500 N/m2 were installed in 12m lengths. These were of 2.7 m internal diameter with a wall thickness of 30 mm, and utilised GRP sleeve joints. The pipe laying was started from the point of the intake structures and progressed towards the inlet area of the pumphouse. The design did not accommodate flexible joints or rocker pipe connections at the interface of the pipe with the structure in the transition zone. The water test was carried out on the cooling water intake system by flooding the Pumphouse by opening one intake pipeline, while the other two pipelines were isolated using their stop gates. When the initial flooding was completed up to -5.20 m level, the other two pipelines were made active to successfully flood the pump house to 0.00 m level. The three 2.7 m diameter GRP pipes failed at the forebay inlet area as a consequence of the flooding of the cooling water pumphouse. The failure took place at a distance of 36.28 m from the inlet wall of the pumphouse, and was consistently at the crown of the joints fracturing both the pipe and the coupling. The mean thickness of the backfill layer (including the armour layers) was 4.46 m. Assuming that the average unit weight of the backfill was 19 kN/m3, and that of the in-situ soil was 16.5 kN/m3. The net additional pressure at the invert of the pipeline was 11.15 kPa. The settlement of the underlying deposits was computed using the equation:

Bulk density: 1.6 1.9 Mg/m3 Clay fraction of 30-80 % Liquid limit of 30-50% Plastic Limit of 0-27%

=H

1 Es

(9)

The ground settlement thus estimated was 30 mm for 2 m (H) of soft soil increasing to 90 mm for 6 m (H) of soft soil beneath the pipelines.

NUMERICAL MODELLING
In numerical modelling, the quantitative description of physical phenomena is established with a system of ordinary or partial differential equations valid in a certain region (or domain) with the imposition of suitable boundary and initial conditions. The

PREVIOUS PAPER | CONTENTS | INDEX | SEARCH | NEXT PAPER

Soil-Structure-Pipe Interaction with Particular Reference to Ground Movement

1135

modelling acts to educate the intuition of the design engineer by providing a series of cause-and-effect examples. Finite Difference Analysis (FD) is one of the simplest forms of such discretization processes. In a field such as geomechanics, the field data (such as in-situ stresses, material properties, and geological features) will never be known completely. However, if extensive field data are available, then these can be incorporated into a comprehensive model that can yield design information directly. More commonly, however, the data-limited model does not produce such information directly, but provides insight into mechanisms that may occur; the designer can then do simple calculations, based on these mechanisms, which estimate the parameters of interest or the stability conditions. For the numerical modelling, the central pipeline out of the three was considered for Flac3D modelling, as illustrated in Figure 2. The pipeline is modelled as a structural element in which flexure takes place only in the longitudinal direction. The techniques proposed by Selvadurai18 (1979), can be used to examine the flexural interaction between both near surface and deeply embedded pipelines located in a transition zone. The trench section for the central pipeline was assumed rectangular. The pipeline was considered as a continuous beam fixed at both ends. The sign convention of compression as negative and tension as positive was adopted. The model is symmetrical about the global X-axis. The pipes, couplings, soil properties and boundary conditions are defined for a transition zone of 9 pipe lengths. The model was subjected to different settlement conditions for analysis in order to study the development of critical stress regimes. A soil stiffness, K, of 572.6 kN/m2 (based on the Selvadurai,14 1984) was adopted in the numerical analysis. The corresponding settlement of the structure (weight of 1360 Tonnes) and the pipeline are 149.1 and 19.5 mm respectively. The differential settlement of 129.6 mm occurring over a transition length of 9 pipe lengths was modelled. Figure 2 shows the results of the modelling to indicate a stress concentration to occur at the pipe crown at a third of the modelled transition length. The failure that occurred in the field was also at the third coupling and affected both the coupling and the crown of the pipe (see Figures 3 and 4). The model output analyse study illustrates that, the stress concentration at the crown of the pipe is due to the induced settlement. Similar patterns of stress concentrations are observed at the crown level of the joint couplings as illustrated in the Figures 3 and 4. According to Leonhardt's adaptation of Jaky's theory, the additional soil stress caused by the presence of the pipes, will be dissipated over a distance of about 2 pipe diameters. Consequently, if the compressible foundation soil at the bottom of a trench is removed for this additional depth, and replaced by gravel, or well-compacted sand, the settlements will be minimal. Such a solution is likely only to be practicable with small diameter pipes, and an alternative, which may be used with larger diameter pipes, is to construct a reinforced concrete slab over the full width of the trench, to redistribute the pressure concentration.

PREVIOUS PAPER | CONTENTS | INDEX | SEARCH | NEXT PAPER

1030 Plastics Pipes XI

Fig. 2 "Maximum stress distribution on the soil pipe model"

PREVIOUS PAPER | CONTENTS | INDEX | SEARCH | NEXT PAPER

Soil-Structure-Pipe Interaction with Particular Reference to Ground Movement 1031

Fig. 3 Fractured GRP coupling surface and the Numerical modelling output.

PREVIOUS PAPER | CONTENTS | INDEX | SEARCH | NEXT PAPER

1032 Plastics Pipes XI

Fig. 4 Fractured GRP pipe surface and the Numerical modelling output.

PREVIOUS PAPER | CONTENTS | INDEX | SEARCH | NEXT PAPER

Soil-Structure-Pipe Interaction with Particular Reference to Ground Movement

1139

A further possibility is to support the whole, or part, of the pipeline on piles. If this solution is adopted, however, account must be taken of the increase in the vertical soil pressure on the pipeline (due to the increased projection ratio). None of the measures described above should be expected to eliminate differential settlements completely, and designers should therefore ensure that flexible joints are provided at adequate intervals. This is particularly important at the approach to structures, where the first pipe flexibly jointed at each end should ideally be capable of accommodating the differential settlement without exceeding the allowable angular rotation of the joints.

CONCLUSIONS
Established pipeline design procedures frequently ignore or underestimate the settlements of soil masses, pipelines and structures. Settlement damage is frequently observed in old pipelines, and is the most common cause of structural failure in new pipelines. Analysis of pipelines as strip foundations can provide a useful estimate of likely settlements. Differential settlements between pipelines and structures are particularly dangerous. Pipeline design should include analysis of settlements, and the provision of measures to limit them and/or enable the pipelines to accommodate their effects. The ability to accommodate settlements should be taken account of during the pipe material selection process. Major pipelines should be designed with the assistance of comprehensive ground investigations, and more rigorous methods of analysing settlements. The effective modulus of a pipeline foundation will vary from place to place, reflecting inconsistencies in the placing and compaction of bedding material, variations in bedding thickness, and in sub-grade properties. Differential settlement is a serious hazard, needing consideration at the pipeline design stage. Pipeline, embedment, trench backfill and native soils can be modelled as an elastic system.

REFERENCES
1. 2. 3. D.D. Clarke, The Distortion of Rivetted Pipe by Backfilling, Proc. ASCE, 1897 M.G. Spangler, Underground Conduits: An Appraisal of Modern Research, Trans ASCE, 113, 1948. R. Lazard, Ouvrages Circulaires Place en Terre, Travaux 33, 1935.

PREVIOUS PAPER | CONTENTS | INDEX | SEARCH | NEXT PAPER

1140 4. 5.

Plastics Pipes XI R.E. Barnard, Design and Deflection Control of Buried Steel Pipe Supporting Earth Loads and Live Loads, Proc ASTM, 1957. G. Leonhardt, Die Belastung von Erdverlegten Rohren unterschiedlicher Steifigkeit unter Bercksichtigung des Verformungsverhaltens des umgebenden Bodens (Loading of buried pipe of varying stiffness with emphasis on the deformation of the surrounding soil), 3R International, 1977. Civil Engineering Specification for the Water Industry, 5th Edition. M.P. OReilly, R.B. Rosbrook, G.C. Cox and A. McCloskey, Analysis of Defects in 180 km of Pipe Sewers in Southern Water Authority, Research Report 172, TRRL, 1989. S.J. Rolfe, The Effects of Pipeline Settlement, World Water and Environmental Engineering, 22(6), 1999. J.L. Olliff, Pipeline Foundation Design, Document TC164/165/JWG1/TG1, CEN, 1994. C.B. Greatorex, Ductile Iron Pipelines-Embedment Design, Publication PJF268 Section 5, Stanton & Staveley, 1979. C.B. Greatorex, Personal Communication to J. Olliff, 1990. J.L. Olliff, European Structural Design Standardisation for Sewers and Water Mains: A Report on Progress, Pipeline Management 93, 1993. E.E. Lundquist, Strength Tests of Thin-Walled Duralumin Cylinders in Pure Bending, Technical Note 479, Nat. Advisory Committee on Aeronautics, 1933. A.P.S. Selvadurai, The Flexure of an Infinite Strip of Finite Width Embedded in an Isotropic Elastic Medium of Finite Extent, International Journal of Numerical and Analytical Methods in Geomechanics, 8, 1984. Fletcher and Herrman, Elastic Foundation Representation of Continuum, Journal of the Engg Mechanics Division, Proc. ASCE, 97, 1971. Winterkorn and Fang, Hand book of Foundation Engineering, VAN Nostrand Reinhold, New York, U.S.A., 1974. Bowles, Foundation Analysis and Design, McGraw Hill, New York, U.S.A., 1977. A.P.S. Selvadurai, Elastic Analysis of Soil - Pipe Interaction, Development in Geotechnical Engineering Series, Elsevier Scientific Publication Co., 17, 1979.

6. 7.

8. 9. 10. 11. 12. 13. 14.

15. 16. 17. 18.

PREVIOUS PAPER | CONTENTS | INDEX | SEARCH | NEXT PAPER

You might also like