基于粒子群支持向量机的高心墙堆石坝渗透系数反演

发布时间：2018-11-21 11:56

【摘要】：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。
[Abstract]:The acquisition of seepage field parameters is one of the difficulties in studying seepage characteristics of high core rockfill dam. For Nuozhadu high core rockfill dam, the learning samples are generated by using the finite element program of saturated unsaturated seepage field, and the mapping relationship between permeability coefficient and water head is established with the help of the highly nonlinear mapping ability of support vector machine (SVM). Based on the recognition error objective function, the particle swarm optimization (PSO) feedback search is used to establish the inversion model of dam permeability coefficient. Taking the typical seepage pressure measurement point of the maximum cross section of the dam as the measured point, the calculation formula of excess pore water pressure dissipation of the dam core wall is derived by using the one-dimensional consolidation theory, and the measured value of the core wall head is corrected. The seepage coefficient of the dam to be retrieved is obtained by inversion of seepage field during the steady period of reservoir water level during the running period, and the seepage field corresponding to the rising water level is used to verify the seepage field. The results show that the inversion results of permeability coefficient are reasonable.
【作者单位】：大连理工大学建设工程学部水利工程学院;
【基金】：国家自然科学基金项目(51379029)
【分类号】：TV641.4

【相似文献】