BP神经网络在预测路基沉降中的应用

Application of BP Neural Networks in Forecast the Final Settlement of Highway Roadbed

为了预测高速公路路基最终沉降量,首先依据影响软土路基沉降的因素选取参数建立了BP神经网络预测最终沉降量模型.结合成都-南充高速公路沉降实测资料及其它文献中大量路基沉降资料,利用BP神经网络预测了其各自最终沉降量.通过检验样本验证,预测精度较高,能够满足实际需要.并对BP神经网络在公路建设中的应用提出了一些注意事项.

In order to forecast the final settlement of highway roadbed, BP neural networks predicting final settlement is established. The parameters are fetched based on the settlement influence factors of soft soil roadbed. Combined with the field data of settlement from the expressway between Chengdu and Nanchong and other highways, the final settlements are predicted with BP neural networks. The results show that precision of prediction of BP neural networks is high and can be use in projects. The attention points for using BP neural networks in highway construction are suggested.