基于内点法和邻域搜索解耦动态规划法的区域电网动态无功优化方法

动态规划法可高效、准确求解小规模区域电网动态无功优化问题，但随着电网规模的增大，存在组合爆炸而导致求解时间急剧增长的问题。为此，提出了基于内点法和邻域搜索解耦动态规划法的区域电网动态无功优化两阶段混合方法。第1阶段，采用Sigmoid函数实现原模型的连续化，然后采用内点法求连续最优解；第2阶段，在连续解基础上，采用启发式邻域搜索策略确定解耦动态规划法搜索空间，通过站间解耦、调压和无功补偿设备的解耦协调以及站内的动态规划求解区域电网动态无功优化问题。该两阶段方法既保证了优化解的质量又可以有效避免离散变量求解状态组合爆炸问题，大幅度提高了动态规划法的计算效率。以某220 kV控制分区的仿真分析，验证了所提方法的有效性。

Algorithm for Dynamic Reactive Power Optimization of Regional Power Grid Based on Interior Point Method and Neighborhood Search Decoupling Dynamic Programming Method

The dynamic programming method can efficiently and accurately solve the dynamic reactive power optimization problem of the small-scale regional power grid. However, with the power grid scale increment, the optimization problem would have the combination explosion issue, which would lead to the rapid growth of solution time. Therefore, a two-stage hybrid dynamic reactive power optimization method based on the interior point method and the neighborhood search decoupling dynamic programming method is proposed. In the first stage, Sigmoid function is used to achieve the continuity of the original model, and then the interior point method is adopted to find the continuous optimal solution; while in the second stage, on the basis of the continuous solution, the heuristic neighborhood search strategy is employed to determine the search space of the decoupling dynamic programming method. The regional power grid dynamic reactive power optimization problem is solved through the decoupling between stations, the decoupled coordination of voltage regulation and reactive power compensation equipment, and the dynamic programming within the stations. The two-stage method could not only ensure the quality of the optimal solution, but also avoid solving the state combination explosion problem with discrete variables, which greatly improves the computational efficiency. The effectiveness of the proposed method is verified by the simulations on a 220 kV control zone.