Lateral load capacity of single piles socketed into Melbourne mudstone

Industrial practice often adopts analytical methods developed for soils based on the continuum approach when designing laterally-loaded piles socketed into rock mass. High safety considerations and the uncertainties arising from the assumption of continuum rock mass may produce inefficient and unsafe design solutions to foundation problems subjected to lateral loads. To date, the research in the area of laterally-loaded rock-socketed piles has not been advanced sufficiently. The available analytical methods assume rocks to be intact and homogenous materials, and the physical effect of joints is not taken into account. Therefore, this study aims to develop an analytical method for laterally-loaded piles socketed into rock mass incorporating the physical effect of joints. A three-dimensional distinct element code, 3DEC, was utilised to simulate the behaviour of laterally-loaded piles socketed into mudstone rock. The capability of 3DEC to model jointed rock mass was first examined by simulating the confined (triaxial) and unconfined (uniaxial compressive test) loading conditions. The results obtained from this modelling were compared with existing empirical relationships and field test data. 3DEC was then employed to simulate laboratory-scale model pile-load tests. A comprehensive parametric study was carried out based on the calibrated laboratory 3DEC model. It was found that the most influential parameters impacting the lateral behaviour of piles socketed into jointed mudstone are rock modulus, pile diameter, pile bending stiffness, pile socket length, joint dip angle and spacing. Based on the understanding of the key influential parameters affecting lateral load behaviour, 3DEC modelling was extended further to simulate full-scale pile load tests conducted in Melbourne mudstone. Based on extensive numerical modelling, the p-y behaviour of piles socketed into mudstone was derived. It was found that a single p-y curve for mudstone accurately predicts pile-head load-deflection behaviour. Subsequently, a thorough parametric study based on full-scale models was carried out to study the effect of joints on the lateral load capacity of piles. This study found that joint sets in different directions have a significant impact on the p-y and the corresponding pile-head load-deflection behaviour. The worst-case condition of four joint sets reduces the load-carrying capacity of the pile by more than 90% compared to a homogenous mudstone. This is due to the deformation mechanisms of the pile-rock system such as the formation of weak wedges/pyramid blocks around the pile. The results obtained from the extensive parametric study were integrated in the derivation of new p-y criteria for laterally-loaded piles socketed into homogenous and jointed mudstone rocks. The p-y criteria developed require basic rock properties which can be conveniently obtained in the laboratory or using the well-established empirical relationships for mudstone based on water content. The p-y criterion proposed for homogenous mudstone condition was validated using the results of another two field pile-load tests. For the p-y criterion of jointed mudstone, a field load test of a pile with an apparent steep joint in the vicinity of the pile was employed to validate its accuracy. It was found that the p-y criterion for jointed mudstone gives a slightly more conservative prediction of the pile-head load-deflection response. It is suggested that the p-y criterion for homogenous and jointed mudstone conditions will provide the upper and lower bound solutions, respectively. This study has successfully developed an analytical method for laterally-loaded piles socketed into mudstone. The proposed design approach pioneers the integration of the physical effect of joints on the p-y behaviour of piles socketed into jointed mudstone. The proposed p-y criteria were validated using field pile load tests and very close predictions were achieved.