大电网动态无功优化问题的快速解耦算法研究

Research on Fast Decomposition Algorithm for Dynamic Reactive Power Optimization on Large Scale Power System

动态无功优化模型是计及设备动作次数约束并考虑变量离散化特性的一种非线性混合整数规划问题。根据修正方程系数矩阵具有可分块结构的特点,利用块矩阵解耦算法求解,指出其关键点是对修正方程系数矩阵的计算存储以及对系数矩阵进行三角分解结果的计算存储,提出对系数矩阵进行2次三角分解,即以时间换取空间的处理方法,实现对大规模电网的动态无功优化,有效地解决了大规模电网的动态无功优化"维数灾"问题,缺点是2次三角分解计算时间稍长。

An integrated dynamic reactive-power optimization （DRPO） model is represented as mixed-integer/ non-linear programming problem, which takes the control devices action times and variables discretization into account. In accordance that coefficient matrix of the reduced linear correction equation has an arrow-shaped block structure and by use of decomposition algorithm of block matrix for solution, the paper points out that the most crucial part of solving reduced linear correction equation with block-matrix decoupling is computing and storage of coefficient matrix and its triangular factorization resuits. Moreover, it indicates twice triangular decomposition of coefficient matrix of the reduced linear correction equation can reduce the data storage requirements to realize dynamic reactive power optimization on large scale power system, which effectively solves curse of dimensionality in the process. However, the disadvantage is long calculation time for twice triangular decomposition.