Prediction of Bearing Capacity of the Soil using Artificial Neural Networks

The bearing capacity of soils stands as one of the most important parameters that determine the vehicles’ off-road mobility. Soil bearing capacity can be determined either experimentally or by calculation using analytical and or empirical formulas. One of the most famous formulas is the Bekker's. Recently, Artificial Neural Networks (ANNs) technique became a powerful tool that can be used for predicting systems’ behavior and performance. The main objective of this paper is to predict the bearing capacity of the soil (plate-sinkage relationships) by using Artificial Neural Networks and to compare the actual results of soil bearing capacity (collected data from Ph.D. Thesis) with results obtained from neural network model and Bekker's formula.

The comparison showed clear superiority and accuracy of neural network technique.

Another objective is to check the generalization ability of the neural network model in predicting the plate-sinkage relationships by using the hypothetical plate.

Results also proved that neural network technique possesses high generalization ability.

In addition, the present study presents a trial towards correlating bearing capacity and soil structure (percentage of sand, silt or clay in soil texture); such correlation has not been studied before.