摘要
基于人工神经网络方法 ,根据云峰大坝坝顶水平位移观测资料能够识别大坝混凝土和岩石基础的弹性模量 .采用 BP学习算法 ,并通过增加阻尼项以及对观测数据的归一化处理 ,避免了迭代过程中的振荡性 ,提高了参数识别精度 .将弹性模量识别结果代入到有限元模型中 ,计算所得到的坝顶水平位移与坝顶观测水平位移水压分量的最大误差小于 0 .15mm.工程实际应用表明 ,用神经网络方法识别材料参数具有识别精度高和收敛速度快等特性 .
Based upon the artificial neural networks,the elastic moduli of the concrete dam and the rock foundation are identified according to the horizontal displacements of the concrete dam in the Yunfeng Project.The oscillation property in the iteration process is overcome and the precision of estimated parameters is enhanced by adding damping item and handling in advance for the observed data.The maximum error of horizontal displacement between the calculated value and separated value is less than 0.15 mm.The practical application in the engineering facts that the neural networks algorithm for the parameters identification poses the properties of highly identification precision and fast convergence speed.
出处
《水利水电技术》
CSCD
北大核心
2000年第8期51-55,共5页
Water Resources and Hydropower Engineering
基金
国家自然科学基金!资助项目 (5 97790 0 1)
工业装备结构分析国家重点实验室开放基金
关键词
参数识别
神经网络
弹性参数
位移
混凝土大坝
parameter identification,neural networks,elastic parameters,displacements,Yunfeng Concrete Dam
