基于混沌-动态递归神经网络的地下厂房高边墙围岩变形预报  

Deformation Prediction of Country Rock of High Wall in Underground House Based on Chaos-Dynamic Recurrent Neural Network

作　　者：吴刚 郑睿 朱登军[3] 

机构地区：[1]洛阳市水利工程局,河南洛阳471001 [2]河南省冶金规划设计研究院有限责任公司,河南郑州450051 [3]河南省人民医院,河南郑州450003

出　　处：《华北水利水电学院学报》2009年第4期83-86,共4页North China Institute of Water Conservancy and Hydroelectric Power

摘　　要：预测地下厂房高边墙围岩变形是大型水电站设计和施工中重要的研究课题.引入混沌理论,对神经网络进行优化,建立变形预报的动态-递归神经网络模型,通过计算最大Lyapunov指数获得预报最大时间天数,运用混沌特性力学参数优化神经网络结构,通过递归神经网络映射混沌相空间相点演化的非线性关系,提高了预测精度和稳定性.某大型水电站实例表明,预报值与实测位移之间误差都小于10%,预测精度高,实时可靠,对开挖结束后的位移进行了预报,结果合理.It s an important research project to forecast the deformation of country rock of high wall of underground house during designing and constructing.The chaos theory was introduced,and neural network was optimized,then chaos-dynamic recurrent neural network was built to predict the deformation.The maximum predictable time was gained by calculating the maximum Lyapunov exponent,and the structure of neural network was optimized through chaotic characteristics.The practical results of large hydropower station show that the errors between prediction values and measuring ones are all no more than 10% ,so the characteristics of precision and stability are improved, and it is real time. The predicting results of displacement after construction are reasonable.

关 键 词：动态-递归神经网络 混沌时间序列 高边墙 变形预报 

分 类 号：TV731.6[水利工程—水利水电工程]