Engineering Structures 29 (2007) 2901–2907 www.elsevier.com/locate/engstruct FEM modelling of a large piled raft for settlement control in weak rock M. Tariq Amin Chaudhary ∗ 3942 Acorn Crescent, Windsor, ON, N8W 5R4, Canada Received 22 December 2005; received in revised form 2 February 2007; accepted 5 February 2007 Available online 21 March 2007 Abstract Use of piles for transferring structural load to suitable load bearing strata and settlement control in alluvial/diluvial deposits is widespread. This paper investigates the application of a group of 1072 piles for controlling settlement of an important and sensitive structure founded on weak rock. 2D axi-symmetrical and 3D finite element models are employed to model the 9.0 m thick and 76 m diameter circular raft, underlying soil/rock and a mix of annular and orthogonal pile layout. Effectiveness of pile foundation in reducing settlement is assessed by comparing the results with the case of raft foundation alone. Results of 2D axi-symmetrical and 3D FEM models are also compared and it is concluded that the numerically efficient 2D axi-symmetrical FEM model is capable of simulating the behaviour of this complicated 3D problem with reasonable accuracy. c  2007 Elsevier Ltd. All rights reserved. Keywords: Piled raft; Weak rock; Finite element modelling; Settlement control; Storage tank 1. Introduction Piled raft foundations provide an economical foundation option for circumstances where performance of the raft alone does not satisfy the design requirements. Addition of piles, in such a situation, helps to improve the ultimate load capacity, the settlement and differential settlement performance [1]. Zeevaert [2] was perhaps the first to report the use of such foundation system for buildings built on compressible clays. Over the past four decades, a number of researchers and practising engineers have contributed to the understanding of the behaviour of this type of foundation system by analytical modelling techniques [3–8]. Approximate analytical methods for piled rafts have also been proposed [9,10]. Understanding of the behaviour of piled-raft foundations is furthered advanced by examining their performance through field studies and measurements [11–13]. Liew et al. [13] demonstrated the unsuitability of simplified methods to accurately predict settlement of piled rafts with a large (100) number of piles. On the other hand, Mandolini and Viggiani [12] have concluded that the sophisticated non- linear pile foundation analysis, which consists of adding nonlinear component of settlement in the pile group, does not ∗ Fax: +1 519 966 8637. E-mail address: mtariqch@hotmail.com. significantly improve the analytical predictions as compared to linear elastic analysis when adequate factor of safety is included in the design. They found that the analytical settlement response of 19 structures supported on piled raft foundations with linear and nonlinear methods was within 20% of the measured values. Randolph [14] claimed that accuracy of current analytical methods for large pile groups is not more than 80% as compared to field studies. In view of these studies and a factor of safety of two for pile design, it was decided to employ a linear elastic analytical method for analysing the problem presented herein. In this paper effectiveness of a group of 1072 piles for controlling settlement of a 76 m diameter cylindrical raft founded on weak rock is investigated by finite element modelling. The problem was investigated by 2D axi- symmetrical and 3D FEM. Conclusions are drawn regarding the effectiveness of piles in reducing settlement and suitability of 2D axi-symmetrical FEM for this complex problem. 2. Overview of current application In the following sections, case study of a very large piled raft foundation for a storage tank structure is presented. The features that distinguish the current application from the ones reported in literature are: number of piles (1072), size of raft (76 m diameter), thickness of raft (9.0 m) and the supporting medium (weathered and soft rock with SPT N > 100). 0141-0296/$ - see front matter c  2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.engstruct.2007.02.001