Modeling soil correlations using neural networks

In geotechnical engineering, empirical relationships are often used to estimate certain engineering properties of soils. Using data from extensive laboratory or field testing, these correlations are usually derived with the aid of statistical (regression) methods. This paper demonstrates the use of back-propagation neural networks to capture multivariate nonlinear interactions between various soil parameters in a system. Two examples are presented to demonstrate the potential of this approach. The first example involves the analysis of data obtained from calibration chamber tests on sand. The other example relates to the determination of the hydraulic conductivity of clay liners. Actual laboratory test data were used in training the neural network. The performance of the neural-network and multiple-regression models were assessed by evaluating the scatter between the predicted and measured values, via the coefficient of regression and the average sum squared error. The neural-network models were found to be more reliable than the multiple-regression models.