RBFNN模型在渗透系数反演中的应用

针对经典的BP网络存在的一些缺陷，采用了径向基函数神经网络（RBFNN）。在相同的收敛条件下，用RBFNN和经典算法的BP网络进行了比较，表明前者具有优越性。在工程实例中，基于人工神经网络的非线性特点，在三维渗流有限元的基础上，利用RBFNN反演了大坝的渗透系数。并利用反演结果进行渗流场分析，水头预报值也有很高的精度，说明反演结果是正确的，从而，验证了RBFNN应用于反演分析中的可靠性。

Application of RBFNN model to inversion of percolation parameter

Aimed at some limitation of the traditional BP neural network, the radial basic function neural network (RBFNN) is brought forward. On the same convergence condition, RBFNN is compared with the neural network based on the traditional algorithm and the excellence of the former is explained, As an example, on the basis of the nonlinear characteristic of ANN and 3-D seepage flow FEM computation , inversion results of the dam percolation parameter are obtained. The seepage flow field is analyzed by using the inversion results; the forecasted water heads approach to the observed values, which illuminates the result is accurate; RBFNN applied to the back analysis is reliable.