【摘要】 基于人工神经网络方法,采用修正的BP算法,根据坝基扬压力和渗漏量的观测数据建立了识别混凝土大坝渗透系数的反演方法.工程实践表明,该法具有识别精度高和收敛速度快等特点.由反演结果进行大坝渗流正分析,所得水头预报值的误差小于0 6%.更多还原

【Abstract】 Based on artificial neural network,a modified BP method is used to build an inversion method to identify the permeability coefficient of concrete dam with data of uplift pressure of the dam foundation and leakage.The practical application shows that this method has higher accuracy and faster convergence rate.The percentage error of the forecast water head is less than 06 percent with this method. 更多还原