Advanced predictive modelling applied to a fibre-reinforced cement stabilised soft soil

One of the solutions to mitigate the brittle behaviour of a cement stabilised soil is to add natural or synthetic fibres. However, there are many experimental parameters (type and length of the fibre, fibre and/or binder content) that may influence the mechanical behaviour of a fibre-reinforced cement stabilised soil. In this paper, a novel statistical methodology of design of experiments is applied to study a fibre-reinforced cement stabilised soil through an integrated analysis. It was observed that the model may not only predict the behaviour of the composite material but also provide the effect of the input parameters on the response. This confirms experimental results that suggest that the binder content is the most important parameter when analysing the unconfined compressive strength and failure mode. It also showed that polypropylene fibres are more efficient than sisal fibres for the conditions studied. In addition, the brittleness index is a good criterion to study the failure mode, showing that the composite material exhibits a brittle failure mode for values higher than 40%, while for values lower than 40% a ductile failure mode is observed.