基于神经网络的中线工程高边坡土体渗透参数反演研究  

Seepage Parameter Inversion of High Slope Soil Body Based on Neural Network

作　　者：宋志宇[1] 李斌[1] 王新奇[1] 

机构地区：[1]黄河勘测规划设计有限公司,郑州45003

出　　处：《长江科学院院报》2009年第B10期29-32,共4页Journal of Changjiang River Scientific Research Institute

摘　　要：基于神经网络的非线性映射特性,在渗流有限元计算的基础上,结合水头和流量等实测资料提出了边坡渗透参数的反演方法。将该方法应用到南水北调中线工程高边坡渗流反演分析中,得到了能够反应工程实际的边坡土体的综合渗透参数。同时利用反演出的渗透参数进行渗流计算,结果表明,该方法对高边坡的渗透参数反演具有较高的识别精度,反演结果可靠。According to the nonlinear mapping characteristics of neural network, and in combination with the actual measured data of water head and flow rate, the authors suggest the seepage parameter inversion method for high slope based on the seepage finite element calculation, obtain the comprehensive seepage parameter of the actual slope soil and apply the method to the high slope seepage inversion analysis for South-to-North Water Diversion Middle Route Project. Meanwhile, seepage calculation is conducted with the inversive seepage parameter. The application of this method proves that it has better identification accuracy for the parameter inversion and the rationality of the results.

关 键 词：神经网络 渗透参数 反演 高边坡 

分 类 号：TV223.6[水利工程—水工结构工程]