Prediction of subgrade reaction modulus of clayey soil using group method of data handling

Abstract

Settlement-based designs for foundations, using subgrade reaction modulus (K_s), is an important technique in geotechnical engineering. Plate load test (PLT) is one of the commonly applied methods to directly determine K_s. As the determination of the K_s from PLT—especially at depths—is relatively costly and time-consuming, it is necessary to develop models that can handle simply determinable properties. In the present study, the suitability of the Group Method of Data Handling (GMDH)-type neural network (NN) to estimate the subgrade reaction modulus of clayey soils has been investigated. In order to derive GMDH models, a database containing 123 datasets compiled from geotechnical investigation sites in Qazvin, Iran, has been used. The performance of the GMDH models has been compared with other available correlations for clayey soils, and it has been demonstrated that an improvement in estimating the K_s has been achieved. Finally, a sensitivity analysis has been conducted on the proposed models, showing that the proposed K_s is considerably influenced by changing the LL value.