Some Aspects of Pile Response Near an Excavation

L.T. Chen B.E., MsEng., Ph.D.

Senior Research Associate, University of Sydney and Geotechnical Engineer, Coffey Partners International Pty. Ltd., Australia H.G. Poulos

AM, B.E., Ph.D., DScEng, FASCE, FIEAust., FAA Senior Principal, Coffey Partners l n t m t h m l Pty. Ltd. and Professor of.Ciuii Engineering, University of Sydney, Australia

Summary This paper analyzes the response of piles to excavation-induced lateral soil movements, focusing on braced excavations in undrained clay layers. The analysis involves a combination of the finite element method and the boundary element method, the former for simulating the excavation and the latter for analyzing the pile response. The analyses show that appreciable effects may be induced in existing piles by an adjacent excavation. The pile response, especially the maximum bending moment, may be increased significantly if the pile head is restrained fiom rotation, thus increasing the risk of possible structural damage of the pile. A larger soil tension capacity is also found to increase the pile response. For closely spaced piles, the group effect may influence the pile response if a large amount of soil movement is encountered. The extent of the influence depends on a pile arrangement and pile spacings.

1. INTRODUCTION Excavations for the construction of high rise buildings in congested urban areas have become increasingly prevalent. They may however cause damage to surrounding existing structures because of the soil movements they induce. For example, there have been reports of severe damage to piled foundations due to excavations (e.g. Finno et al, 1991; Amirsoleymani, 1991; Chu, 1994), and thus it is important to develop a better understauding of the problem so that appropriate measures can be taken to minimize any potential damage. The present paper focuses on the effect of excavations on the lateral response of piles. Although an excavation will cause both vertical and lateral soil movements, the latter component is considered to be more critical for adjacent piles as piles are often not designed to sustain significant

lateral loadings. For this reason, and as a first step, only the lateral soil movement will be cotsidered in the following analysis. The analysis involves a finite element program and a boundary element program. First, the finite element program is used to simulate the excavation procedure and to generate "fiee-field" soil movements. Second, these soil movements are used as input into the boundary element program for analyzing pile response. The main purpose of this paper is to investigate the effect of various factors on the lateral pile response (mainly bending moments and deflections), so that a better understanding of the problem can be achieved.

2. PROBLEM ANALYZED Figure 1 shows the problem addressed, of an existing single isolated pile situated near an

I

I 9

Lp = 22m d = 0.5m Ep =300OOMPa

/ / / / A / / / / , / / / / / / / / / / , / , / / , / / / , / / > / / / / / / / / , / , / / / / / / /

Figure 1 ' Standard ' Problem Analyzed

excavation. With the progresc; of the excavation, the surrounding soils will move 1:owards the excavation and these movements will induce bending moments and deflections in the pile. hl the present study the problem has been simplified as follows: the excavation is sufficiently long that a two dimensional plane str;rin analysis is applicable; the soil is a uniform saturated clay layer and is in an undrained condition during excavation. Some constant parameters have been selected for a 'standard' problem as shown in Figure 1. The definition of the parameters ;is as follows: B = half width of excavation, H = total thickness of soil layer, X = distance :from e:rtcavation face, 4, =

maximum depth of excavation, EL = stiffness of wall, s = strut stiffhess, cu= widrained shear strength of soil, E, = soil Young's madulus, y = unit weight of soil and wll, L, = length of wall, L, = pile length, d = pile diameter, E, =: pile Young's modulus.

3. FINITE ELEMENT AND BOUNDARY ELEMENT PROGRAMS A twedirnensional finite element program was used to simulate the p1.ane-strain excavation without the presence of the pile. The :kite element program used is named AVPUIL (Anidysis of Vertical Piles Under Lateral Loading) anrl has been described elsewhere (see Chen & Poulos, 1993; Chen & Poulos, 1994; Chen, :1994). 'In the program, eight- noded isoparametxic elements were used to model the soils and the suppoxting wall, while Goodman- type interface eleme11t.s were used to model the interaction between the soil and the wall. In the present study, the interface between the soil and the wall was assumed to he rougl~, i.e. no slip occurred. Struts were modelled ,as springs whose stiffness was assigned to an element nodt: corresponding to the strut position. The soil and the interface elements were modelled as elasto-plastic materials, obeying the Tresca failure criterion and a non-associated flow rule. The wall was modelled as a linear elastic material. In the fhite element simulation, the wall was assumed to be installad prior to excavation and to have no effect on surroundiqg soils. The excavation was carried out from. top to bottom in ten steps, with each step involving rcrmoval of a lm thick layer. Four levels of s.truts were simulated, the first being placed after the: first excavation step and the remaining three at steps 4 ,7 and 10. The boundary element prclgram used is named PALLAS (Piles And 1,atera.l Loading Analysis) and has been described e:lsewheire, for example, Hull (1992). PALLAS uses a simplified form of boundary element analysis in which the pile is idealized as an elastic beam and the soil as an elastic continuum, bur. with l i t i n g pressures at the pile-soil interface to allow consideration of local failure of the soil adjacent tc~ the pile. The program

can consider both a single pile and a group of non- identical piles. One of the most important input parameters for PALLAS is the limiting pile-soil contact pressure, p,, which has been found to have a major effect on the pile response (e.g. Poulos, 1973; Chen, 1994). In the present study, p, for single isolated piles has been assumed to be 9c, (where c, is the undrained shear strength of soil), while p, for closely spaced piles will be discussed below.

Lateral soil movement y(mm)

0 -20 -40 -60 -80 -100 0

Figure 2. Soil movement profile at X=lm

4. RESULTS FOR 'STANDARD' PROBLEM Figure 2 shows the computed soil movement profiles at a distance of lm away from the excavation face, for four stages of excavation. It can be seen that the soil movement increases with increasing excavation depth, and the rate of increase becomes more rapid as the soil approaches failure. The maximum soil movements corresponding to the four stages of excavation are plotted against the distance X in Figure 3, and they are shown to decrease with X, as expected.

Figure 3. Relationship between y,, and X

Profiles of pile bending moment and lateral shown in Figure 4(a), the pile deflection follows deflection corresponding to the soil movements closely the free-field soil movement, indicating that shown in Figure 2 are presented in Figure 4. As the pile is relatively flexible. The bending moment

Deflection (mm) Bending moment (kN.m) 0 -20 -40 -60 -80 -100 -100 100 300

0 0

25 (a) Deflection profile

25 @) Bending moment profile

Figure. 4 Pile pesponse for 'standard' problem

Deflection (nun) Bending moment (IcNm)

-20 -1 5 -1 0 -5 0 -300 -200 -100 0 100 0 0

25

(a) Deflection and bending moment profiles

-250 J 1 Head condition

(b) Maximum bending moment

Figure 5. Effect of pile head condition on pile response

profiles shown in :Firure 4@) have a double curvature, with the maximum values increasing with increasing depth of excavation. It should be noted that the pile response shown above does not include any deflections or bending moments which may have already existed before the excavation.

5. PARAMETRIC S'I'UD1E;S There are a number of factors which may affect the pile response, notably construction procedures and supporting conditiom; of the excavation, the soil and pile properties, etc. Pculos & Chen (1995a) have studied some o:f the Ezctors and their results have revealed that: a) piile response (bending moment and deflection) decreases with increasing c, (and EJ but increases wit11 the stability factor N, (N,= yhlc,,); b) pile respom: decreases with stiffer excavation support conditi.ons (i.e. larger wall stifhess E L and strut stiffiess k, and smaller strut

b spacing s); and c) pile bentling moment increases with increasing pile tf arneter, while the deflection generally follows the soil movement unless the pile diameter is larger than lrrt. The effect of some other factors is discussed below.

5.1 Effect of Pile Head Con.dition Laboratory tests on n~odel piles have revealed that the pile head condition may have a remarkable effect on pile response due to lateral soil movements (e.g. Chen, 1994). This effect was investigated in the present study via four different pile head conditions, i.e. freelunstrained, fixedrestrained, fixe&'unrestrained, and fieelrestrained, and results are shown in Figure 5. It can be seen the effect may be remarkable on the bending moment, especially at pile head level, if the pile head is fixed against rotation. The most significant situation i.s when the pile head is fixed against both rotation and .translation. This result may indicate that particular care should be taken

b for existing piles whose head is fixed andlor restrained. To minimize the possibility of damage,

Deflection (mm)

0 -20 -40 -60 -80 -100

I . * ,

(a) Ileflectior~ profile

pile heads should, if possible, be left fiee, if a potential exists for soil movements around the pile is anticipated in future.

5.2 Effect of Soil Tension Capacity In the above studies the soil has been assumed to have a full tension capacity. However, a real soil may have only a limited tension capacity, and so, to study its effect, a case was analyzed in which the soil was assumed to have no tension capacity. For the two extreme cases, it was found that a larger soil tension capacity tended to give a larger soil movement and to consequently result in a larger pile response, as shown in Figure 6. The difference is insignificant in the early stages of the excavation due to a mainly elastic response of the soil, but becomes more pronounced as the excavation goes deeper, and a plastic flow of the soil becomes more pronounced. However, it was also found that the effect of the soil tension capacity may be dependent on excavation support conditions, because a different behaviour has been observed for an unsupported excavation, as discussed by Poulos & Chen (1995b).

5.3 Pile Group Effect Both experimental and numerical work (e.g. Chen, 1994) has shown that the pile ultimate response, in particular the maximum bending moment, due to lateral soil movements may be affected significantly by pile-soil-pile interaction for closely spaced piles. Chen (1994) has shown that the group effect in this case is mainly due to the change in the limiting pile-soil contact pressure, p,. A larger p, generally tends to increase the maximum bending moment if the soil which is in contact with pile has reached an ultimate state. It was found in the present study that a large soil movement (for example, corresponding to a N, value greater than 6) is necessary to cause the pile-soil lateral pressure to reach its ultimate value of 9cu. Therefore, for small soil movements there is effectively no group effect on the pile response.

Bending moment (kN.m) -100 0 100 200 300 400

@) Bending moment profile

D Figure 6. Effect of soil tension capacity on pile response

To illustrate the effect of p, value on the maximum bending moment, a case was studied where three different p, values were used and the soil movement at X=lm corresponded to a N, value of 7. The maximum bending moment is plotted against the p, value in dimensionless form in Figure 7, where p, = p, due to group effectJp,, for a single isolated pile (e.g. 9c3; and m, = M- due to group effecthi- for a single isolated pile. p, and ma may be taken as the group factor for p, and M-, respectively. It can be seen that m, increases with increasing p,, reflecting a similar effect of the p, value on the maximum bending moment.

Group factor forb, pa

Figure 7. Relationship between p, and m,

To obtain the p, value for closely spaced piles, the finite element method described by Chen & Poulos (1994) may be used. In this method, the pile-soil system is simplified to a two-dimensional system in the horizontal plane, and a lateral soil movement is applied incrementally until the system has reached an ultimate plastic state. The p, value for each pile in the group can thus be obtained and used to compute bending moments for each pile.

In the present study, m, values have been computed for piles in one row and two rows for specified pile centre-to-centre spacings, and are shown in Tables 1 and 2. As shown in Figure 8, for piles in two rows there are two different arrangements, i.e. "I" type and "Z" type arrangements. From these values it appears that the group effect may either increase or decrease the maximum pile bending moment, depending on the pile arrangement and spacing. These m, values may be used approximately as correction factors in conjunction with results for single isolated piles, such as the design charts presented by Poulos & Chen (1995a), to estimate the maximum bending moments for closely spaced piles.

6. CONCLUSIONS In this paper, a two-stage analysis involving the finite element method and the boundary element method has been used to study the pile behaviour due to excavation-induced lateral soil movemenl, focusing on braced excavations in undrained clay layers. Both pile deflection and bending moment are shown to increase with increasing excavation depth. The induced additional moment may be detrimental to the structural integrity of existing piles, especially when it is combined with moments which may have existed before the excavation. Factors which have been found to have some influence on the pile response include pile head condition, soil tension capacity and the group effect. It is shown that the bending moment at the pile head level tends to increase greatly when the head is restrained against translation andlor fixed against rotation. Care should be taken of such a case. It is also shown that a larger soil tension capacity generally leads to a larger pile response, particularly for a large depth of excavation. The

Table 1 m, for piles in one infinitely long row

Table 2 m, for piles in two infinitely long rows

Case

I- 1 1-2 1-3 z- 1 2-2 2-3

Sdd

3 3 6 3 3 6

Sdd

3 6 3 3 6 3

mP row 1

1.1 1.4 0.9 1.2 1.4 1.1

row 2 0.9 1.2 0.7 1.1 1.2 1.0

row 1 c9

row 2

e3 @-

row 1 .... @..

row 2 .a @....

"I" ;mangement "2" arrangement

Figure 8 Piles in two infinitely long rows

group effect may either increase or decrease the maximum bending moment in the pile, depending on the pile arrangement anld pile spacing. Group factors for some cases have been presented, and these may be used in. conjunction with results for single isolated piles to es,tirnate the maximum bending moment for closely spaced piles. However, due to a complex nature of the pile-soil-pile interaction, the presented group factors should be treated with caution.

7. ACKNOWLEDGICMENr The work described in this paper forms part of a project on "The effects o:F construction-induced movements on existirig pile foundations" which is supported by a grant from Tihe Australian Research Council.

8. REFERENCES Amirsoleymani, T. (1 991). Elimination of Excessive Differential Settlement by Different Methods, Proc. Ninth Asian Regional Conf. on Soil Mechanics and Foundation Engrg., Vol. 2, pp. 35 1- 354. Chen, L. and Poulos, H.G. (:1993). Analysis of Pile- Soil Interaction Under Lateral Loading Using Infinite and Finite Elements, Computers and Geotechnics, Vo1.15, pp. 189,-220. Chen, L. and Poulos, H.G. (1994). A Method of Pile-Soil Interaction llnalysis for Piles Subjected to Lateral Soil Movement, Proc. 8th Int. Conf. on Computer Methotls and Advances in Geomechanics, pp.23 1 1-23 16. Chen, L. (1994). Effect of Lateral Soil Movements on Pile Foundations, :PhD Thesis, The University of Sydney. Chu, Y.K. (1994). A Failur~e Case Study of Island Method Excavation irl Soft Clay, Roc. Int. Conf.

on Design and Construction of Deep Foundations, VO~. m, pp.1216-1230. Finno, RJ., Lawrence, S.A., AUawh, N.F. and Harahap, I.S. (1991). Analysis of Performance of Pile Groups Adjacent to Deep Excavation, J. Geotech. Engrg. Div., ASCE, Vo1.117, No.6, pp.934-955. Hull, T.S. (1992). User's Manual for Program PALLAS. The University of Sydney. Poulos, H.G. (1973). Analysis of piles in soil undergoing lateral movement, Journal of Soil Mechanics and Foundation Engineering, ASCE, Vol. 99, pp. 391-406. Poulos, H.G. and Chen, L.T. (1995a). Pile Response Due to Excavation - Induced Lateral Soil Movement, Submitted for publication to Journal of Geotechnical Engineering, ASCE. Poulos, H.G. and Chen, L.T. (1995b). Pile Response Due to Unsupported Excavation - Induced Lateral Soil Movement, Submitted for publication to Canadian Geotechnical Journal.