基于神经网络技术的桩基承载力预测模型及其应用  

Prediction Model for Bearing Capacity of Pile Foundation Based on Neural Network Technique and Its Application

作　　者：孙立新[1,2] 

机构地区：[1]西北工业大学力学与土木建筑学院,陕西西安710072 [2]陕西交通职业技术学院公路工程系,陕西西安710018

出　　处：《水利与建筑工程学报》2012年第5期120-123,共4页Journal of Water Resources and Architectural Engineering

摘　　要：根据长期的工程实测资料,在分析小波概率神经网络(WPNN)与数据融合技术在预测单桩竖向承载力中的应用原理的基础上,建立了基于小波概率神经网络和数据融合技术的预测模型。利用静载实验数据对模型进行了预测,并对预测结果进行了误差分析,结果表明,预测的结果和静载实验数据吻合较好,从而证实了WPNN预测方法具有较好的可靠性和工程应用价值。According to the long-term actual engineering data, the prediction model is set up based on the wavelet probability neural network （WPNN） and data-interfusion technique through analyzing their application principle in the prediction of vertical single-pile bearing capacity. Then, the model is predicted by using dead-load experiment data, and the error analysis is made for the prediction results. The analysis results show that the prediction results tally better with the dead-load experiment data, which would prove that the WPNN prediction method has the satisfied reliability and engineering application value.

关 键 词：桩基承载力 小波概率神经网络 数据融合技术 承载力预测 

分 类 号：TU473.11[建筑科学—结构工程]