A BP neural network model for SWCC considering consolidation stress

Soil-Water Characteristic Curve (SWCC) plays an important role in theoretical research and practical application. At present, SWCC can be obtained from experiments. The experimental results are inconvenient for practical application. Some models are presented to fit experimental results, such as Gardner model, V-G model, etc. Recent experimental results show that besides water content, the consolidation stress has influence on SWCC either, while most of the models did not consider this factor. In this paper, based on the experimental results of SWCC under different consolidation stresses, a BP neural network model for SWCC considering consolidation stress is created. The input layer of BP model contains 2 neurons, namely soil suction and consolidation stress; output layer includes 1 neurons, namely mass soil water content. And there are two layers in middle layer, one layer contains 6 neuroses and another layer contains 13 neuroses. Compared with previous research and experimental results, the BP model presented in this paper shows higher accuracy and more convenience.