Performance of geogrid-reinforced pile-supported embankments over decomposed granite soil

This paper presents a full-scale test of high-speed railway embankments over completely decomposed granite soil foundations in order to investigate the performance of geosynthetic-reinforced and pile-supported (GRPS) embankments. The emphasis is placed on the study of the load-transfer mechanisms in those GRPS embankments and on verifying the existing design approaches, taking into account soil arching effects. To do so, four fully instrumented embankment sections were studied, with two sections of geogrid-reinforced and cement-mixing pile-supported embankments and the other two of geogrid reinforcement only. Six commonly used existing design methods for GRPS embankments were tested to show their limitations and applicability. Experimental data from field monitoring for nearly 2 years in these sections were obtained. Results show that all of the six existing design methods tested significantly over-predict the pile efficiency at the end of full embankment, thus leading to a conservative design. BS 8006 and modified BS 8006 yield an overestimation of geogrid strains and thus a conservative estimation. However, all the other methods tested for geogrid strain calculation may lead to an unsafe design. Therefore, it is highly recommended to compare the design results using different approaches in order to optimise the design.