Prediction of composite foundation settlement process based on a modified Poisson-superposition wavelet model

The model and method used to predict a composite foundation settlement process were studied. The characteristics of the modified Poisson model and its applicability were analyzed and a modified Poisson-superposition wavelet neural net model was proposed. Combined with practical observation data, the CFG pile composite foundation settlement process was analyzed and predicted. A comparison of the obtained theoretical results with those from the modified Poisson model was made. It is shown that the suggested model has better applicability and enables to predict with a higher accuracy, whose absolute error is less than 1mm.