论文部分内容阅读
针对仅以水头观测资料进行渗透系数反分析存在结果不唯一的问题,引入渗流量观测资料建立以水头和渗流量的误差加权平方和作为目标的函数,提出了基于层次分析法(AHP)和遗传算法(GA)的某混凝土面板堆石坝渗透参数反分析方法。先通过BP神经网络建立介质渗透系数及边界水位与计算水头和渗流量之间的隐式映射关系,将水头和渗流量等数据的各项误差归一化后构造目标函数;然后应用AHP计算目标函数中水头和渗流量测点误差项的权系数,利用GA进行全局寻优,求解带有约束条件的多目标优化问题;最后,应用提出的方法反演分析了某水电站面板坝的渗流场,得到了坝体和坝基材料的渗透系数。结果表明,该渗流反分析方法收敛快、计算效率高且拟合的测点水头和渗流量与实测值吻合好,渗透系数符合实际。
Aiming at the problem that the reverse analysis of the permeability coefficient is based on the head observation data, the result is not unique. By introducing the observation data of the seepage, a weighted sum of squares of the error of the head and the seepage volume is established as a function of the target. Based on the analytic hierarchy process (AHP) ALGORITHM (GA) BACK ANALYSIS METHOD FOR SEEPAGE PARAMETERS OF A CONCRETE FACED ROCKFILL DAM. The BP neural network is used to establish the implicit mapping relationship between the medium permeability coefficient and the boundary water level and the calculated head and seepage volume. The errors of the head and the seepage data are normalized to construct the objective function. Then, the target is calculated using AHP In this paper, we use GA to solve the multi-objective optimization problems with constraints, and then use the proposed method to inverse analyze the seepage field of a dam of a hydropower station. Get the permeability coefficient of dam and dam foundation material. The results show that the method of seepage back analysis has the advantages of fast convergence, high computational efficiency, good agreement between the measured head and seepage and the measured values, and the permeability coefficient in line with the actual situation.