A novel improved sequential quadratic programming algorithm to solve DG dispatch in distribution system

Optimal power flow (OPF) problem for distribution networks is investigated with the integration of distributed generation (DG) in this paper. By considering the objectives of minimal line loss, minimal voltage deviation and maximum DG active power output, the proposed OPF formulation is a multi-object optimization problem. Through normalization of each objective function, the multi-objective optimization is transformed to single objective optimization. To solve such a nonconvex problem, the trust region sequential quadratic programming (TRSQP) method is proposed which iteratively approximates the OPF by a quadratic programming with the trust region guidance. The TRSQP utilizes the sensitivity analysis to approximate all the constraints with linear ones, which will reduce the optimization scale. Active set method is utilized in TRSQP to solve quadratic programming sub-problem. Numerical tests on IEEE 33-, PG&E 69- and actual 292-bus systems show the applicability of the proposed method, and comparisons with the primal-dual interior point method (PDIPM) and sequential linear programming method (SLPM) are provided, which indicates that the proposed algorithm for distributed generation control optimization in distribution system is feasible and effective.