基于改进泊松-复合小波模型的复合地基全过程沉降预测

对复合地基全过程沉降预测模型与方法进行了研究.分析了改进泊松模型的特点和适用性,提出了改进泊松-复合小波神经网络修正模型.结合实际工程数据对CFG桩复合地基全过程沉降进行了分析和预测,并与改进的泊松模型进行了对比分析.结果表明,提出的模型适用性强,具有更高的预测精度,其绝对误差在1mm以内.

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.