库区河道洪水演进稀疏矩阵求解技术及应用研究  

Research on Solving Technique for River Flood Routing Model with Sparse Matrices and Application

作　　者：席秋义 徐建光 张洪波[2] 

机构地区：[1]国网陕西省电力公司电力科学研究院,陕西西安710054 [2]长安大学,陕西西安710054

出　　处：《陕西电力》2013年第10期61-65,共5页Shanxi Electric Power

基　　金：国家自然科学基金资助项目(51009008)

摘　　要：鉴于河道洪水演进计算模型大型方程组直接求解具有耗时长、占用计算机运算空间大以及精度没有保证的问题,采用系数矩阵压缩存贮方式,可大大减小洪水演进模型求解计算机存贮量和内存占用量,通过选取含有最少非零元素行作为主元行,该行绝对值最大的元素作为主元元素,有效克服了系数矩阵元素相差数量级时计算结果不准确的问题,且消去运算不占用多余的空间,回代也很简单。在安康水库的应用表明,该方法具有计算效率高和精度高的优点。Whereas the problem of long time consuming, larger memory occupancy and poor precision in solving the fiver flood routing model equations, the paper uses the compressed storage technique of coefficient matrix, so the occupied computation memory is decreased greatly. By selecting the row with least nonzero elements as the pivotal row and the maximum element of absolute value in the pivotal row as pivotal element, the problem of inaccuracy is avoided when the magnitude difference existing in coefficient matrix elements, and elimination operation and back substitution is simple. The application result shows that the method is of high efficiency and precision in Ankang reservoir.

关 键 词：安康水库 洪水演进 稀疏矩阵 非恒定流 

分 类 号：TV122[水利工程—水文学及水资源]