Model experiments on piled embankments. Part II

This two-part study concerns a series of twelve 3D laboratory model tests on piled embankments. In the first part, the measured load distribution, deformation, and strains were presented and analysed. In this second part, the measurements are compared with calculations made using the EBGEO (2010) and CUR 226 (2010) analytical model, hereafter called ‘EBGEO’. Possible improvements to the analytical model are also suggested, and the resultant calculations are compared with the measurement results.EBGEO calculations consist of two steps: (step 1) load distribution in the fill i.e. arching behaviour, and (step 2) the load deflection behaviour of the geosynthetic reinforcement (GR). For the test conditions (static load, laboratory scale), it was found that the GR strains calculated using EBGEO are larger than the measured GR strains (approximately a factor of two for GR strains larger than 1%). The EBGEO calculations are therefore too conservative. The following reasons were found. In step 1, the response of arching to subsoil consolidation in the experiments is different from that assumed in EBGEO. In step 2, the distribution of loading on the ‘GR strips’ lying between adjacent piles is quite different from that assumed by EBGEO, and EBGEO only takes part of the subsoil support into account. Modifications are suggested for the second calculation step. It is shown that this modified step 2 model results in significantly closer agreement with the test measurements.