Optimal power flow algorithm based on self-adaptive filter-trust region method

The Optimal Power Flow (OPF) problem for power systems could be expressed by a Non-Linear Programming (NLP) model, which is difficult to be solved. This study presents a practical algorithm to solve OPF through a successive linear programming (SLP), in which conventional power flow and linearized Sub-Problem (LSP) are solved by alternating iteration. A Self-Adaptive Filter-Trust Region (SAIFT) method is proposed to solve the LSP, in which the acceptance of the current iteration point is decided by a filter, and the step size is controlled by the trust region. The Wolfe linear search method is used to provide the direction of optimization when the consequence is dominated by filter. Furthermore, the linearization strategy of transformer ratio and the dynamic adjustment strategy of step size are proposed, which simplifies the formulation and improves the convergence of the algorithm. Case studies based on Ward-Hale 6-bus system demonstrate the efficacy of the new method in solving the OPF problems.