基于滤波器-信赖域方法的最优潮流算法

在电力市场环境下，求解诸多问题都需要最优潮流作为理想的工具。该文基于滤波器和信赖域的思想提出了求解最优潮流的新算法：由信赖域决定线性化步长，线性规划子问题由多步中心校正原一对偶内点法进行求解，并采用了考虑电网拓扑的物理策略和动态调整线性规划子问题的收敛判据策略来改善最优潮流算法的稳定性和收敛性。该算法通过逐次求解线性规划子问题，在滤波器中利用多目标规划的优超(Domin Rllce)概念决定是否接受新的点，算法本身具有非单调的性质。通过与预测一校正方法的比较，进一步验证了多步中心校正方法的求解效率。对系统规模从14节点到662节点的7个电力系统作了全面的数值计算，计算结果表明，该算法具有较高的稳定性和快速收敛性，具有实用意义。

OPTIMAL POWER FLOW ALGORITHM BASED ON FILTER-TRUST REGION METHOD

A novel algorithm for the optimal power flow (OPF) is presented based on a filter-trust region method to solve linear programming subproblems. The linear step size is controled by the trust region. A primal-dual interior point method with multiple centrality corrections is applied to solving the linear programming subproblems. Two strategies are employed to improve the stability and the convergence of the OPF algorithm. One is a physical strategy based on the network topology and the other is the heuristic of dynamically adjusting linear programming tolerance. Whether or not to accept a new point is decided by the filter using the dominance concept of multi-objective programming. The algorithm has the non-monotonic properties is demonstrated compared with the predictor-corrector method. The superiority of the multiple centrality corrections technique. Numerical results of several test systems ranging in the size of 14 to 662 buses are presented. The results show that the algorithm is fast, robust and suitable to real-time applications.