基于遗传算法优化BP神经网络的面板堆石坝变形预测  

Prediction on deformation of concrete face rockfill dam based on BP neural network optimized by genetic algorithm

作　　者：董霄峰[1] 王海军[1] 杨敏[1] 

机构地区：[1]天津大学水利工程仿真与安全国家重点实验室,天津300072

出　　处：《水利水电技术》2012年第12期22-25,共4页Water Resources and Hydropower Engineering

基　　金：国家自然科学基金创新研究群体科学基金资助项目(51021004)

摘　　要：结合实测数据建立了面板堆石坝坝体变形量的BP神经网络预测模型,并引入遗传算法对其进行优化,结果表明经遗传算法优化后的模型预测结果要优于未优化模型的预测结果,优化模型具有更高的预测精度和更强的预测能力。基于某在建工程实例验证了本方法的可行性与实效性,预测结果不仅满足工程安全要求,而且具有较好的可信度和工程参考价值。在上述优化预测模型基础上,实现了引入施工沉降作为输入量对面板挠度进行精确预测,证明了应用这种方法进行面板挠度预测的合理性和优越性。A BP neural network predicting model for the deformation of concrete face rockfill dam is established with the meas- ured data, which is then optimized by the genetic algorithm. The result indicates that the prediction result from the genetic algorithm optimized model is better than that from the model without the optimization and has higher prediction accuracy and greater prediction capacity. Simultaneously, the feasibility and effectiveness of this method are proved by taking a project under construc- tion as the actual case; from which the prediction result can not only meet the relevant engineering safety requirement, but also has better reliability and engineering reference value. Moreover, an accurate prediction made on the deflection of the concrete face is realized by taking the construction settlement as the input value, thereby, the rationality and superiority of this method are demonstrated as well.

关 键 词：遗传算法 BP神经网络 面板堆石坝 变形量预测 施工沉降 面板挠度 

分 类 号：TV641.43[水利工程—水利水电工程]