Jet grouting mechanicals properties prediction using data mining techniques

Jet Grouting (JG) is one of the best technologies to improve soils and particularly soft soils. Its main characteristic is its versatility. However, while widely applied (e.g. in important JG works), its design is essentially based on empirical methods that are often too conservative. Hence, is very important to develop rational models to estimate the effect of the different parameters involved in JG process. For this purpose, in previous works we applied Data Mining (DM) techniques to estimate mechanical properties of JG laboratory formulations. In this paper is originally proposed and compared three DM techniques in order to estimate Uniaxial Compressive Strength (UCS) of samples collected from JG columns. In particular, we analyze and discuss the predictive capabilities of Artificial Neural Networks, Support Vector Machines and Functional Networks. The achieved results are compared to Eurocode 2 analytical model adapted to JG mixtures. Furthermore, explanatory knowledge is given in terms of a sensitivity analysis procedure. Such procedure confirms that the relation between the mixture porosity and the volumetric content of cement, age of the mixture, jet grouting system and soil properties play an important role in UCS estimation. The effect of these key parameters is also measured and discussed.The authors wish to thank to Portuguese Foundation for Science and Technology (FCT) the support given through the doctoral grant SFRH/BD/45781/2008Tecnasol-FG