基于粒子群支持向量机的高心墙堆石坝渗透系数反演

Back analysis of permeability coefficient of high core rockfill dam based on particle swarm optimization and support vector machine

渗流场参数的获取是研究运行期高心墙堆石坝渗流特性的难点之一。针对糯扎渡高心墙堆石坝,利用饱和–非饱和渗流场有限元程序生成学习样本,借助支持向量机的高度非线性映射能力,建立了渗透系数与水头之间的映射关系。再以识别误差目标函数为适应值,采用粒子群优化算法反馈搜索以建立大坝渗透系数反演模型。以大坝最大横剖面典型渗压计测点为实测点,采用一维固结理论推导了大坝心墙超静孔隙水压力消散计算公式,并对心墙水头实测值进行修正。通过对运行期库水位稳定时段渗流场的反演得到大坝待反演分区的渗透系数,再利用水位上升期对应的渗流场进行验证。结果表明,渗透系数反演结果是合理的。

It is important to determine the seepage field parameters of high earth-rock-fill dam using the observed seepage data during operation period. For Nuozhadu core rockfill dam, the training samples are produced for saturated seepage field which is calculated by the finite element program. The nonlinear relationship between seepage parameters and water heads is established using the SVM mapping. Then taking the error objective function as the fitness value of particle swarm optimization（PSO）, the seepage parameters should be identified by PSO. Based on the one-dimension consolidation theory, the dissipation formulae for the excess pore water pressure in the core wall are derived, and they are used to correct the measured seepage pressure values in the core wall. The recorded osmotic pressure curves of osmometers, which are distributed in the maximum section, are used for this back analysis. The permeability coefficients of the dam materials are retrieved using the corrected measured seepage pressure values under steady state of seepage condition, i.e., the water level remains unchanged. Meanwhile, the parameters are verified by the unstable saturated-unsaturated seepage field while the water lever rises. The results show that the permeability coefficients are reasonable.