基于小波概率神经网络的单桩竖向承载力预测模型及应用研究

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,建立基于小波概率神经网络和数据融合技术的预测模型。根据长期的工程实测资料,利用高层建筑物静载试验数据对模型进行检验,并选取典型的样本进行预测值的误差分析。结果表明,预测的结果与静载试验数据吻合较好,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。

STUDY ON THE PREDICTION MODEL OF THE SINGLE PILE VERTICAL BEARING CAPACITY BASED ON WAVELET PROBABILISTIC NEURAL NETWORK AND ITS USE

It was analyzed that the applied principles of wavelet probabilistic neural network(WPNN) and data fusion technique in the prediction of single pile vertical bearing capacity,and a prediction model based on WPNN and data fusion technique was set up.This model was examined by the static load test data of tall buildings,and according to the measured data of long-term projects.The error analysis of the predicted values was also carried out by selecting typical specimens,the results showed that the predicted data agreed well with those of the static load test,which verified the better reliability and applied value of WPNN prediction method.