Gypseous soil exhibits problematic geotechnical engineering properties as they expand, collapse, disperse, undergo excessive settlement, owns a distinct lack of strength, and it is soluble. Gypseous soil has a metastable structure, with dissolvable minerals with a minimal quantity of clay binding the particles together. When gypseous soil unsaturated, they are quite potent. When they are subjected to increased wetness, however, the excess water weakens or damages the bonds, resulting in shear failure and subsequent settlement. Estimating the settlement of shallow foundations on gypseous soils is a difficult topic that is still not fully understood. It is concluded that artificial neural network (ANN) is appeared to be viable solution since it has been successfully used in numerous prognosis applications in geotechnical engineering. In this research, the precipitation values of gypsum soil were predicted under the influence of the applied load using an artificial neural network. The study found that this model is very good in predicting precipitation and found a convergence between the real values and the predict values.
Read full abstract