Study On Models And Methods For Dynamic Optimal Power Flow Of Electrical Power System

Dynamic optimal power flow (OPF) of electrical power system is a valuable tool for unified power flow optimization in time scale, which has theoretical and practical siganificance for power system operating security and economically. In this thesis, classical dynamic OPF problem and dynamic reactive power optimization problem are deeply investigated, and new methods are proposed based on interior point method (IPM) and immune genetic algorithm (IGA). Moreover, voltage stability constrained dynamic OPF and dynamic OPF with discrete control variables are estabilished, and the proposed new methods are applied to solve these new models. The field of dynamic OPF is developed and main results are listed as follows:â‘ For a classical dynamic OPF problem, the feature of the problem is analysed. It is pointed out that the problem is composed of several sub-problems with weak interconnection. While the problem is solved by IPM, KKT (Karush-Kuhn-Tucker) matrix is a block diagonal bordered matrix. To solve this special kind of large-scale nonlinear programming problem, a predictor-corrector decomposed IPM (PCDIPM) is proposed. Blocked LU factorization and blocked triangular matrix computation are used to combine the predictor-corrector process with decomposed KKT system. Simulation results show that the proposed method is computationally attractive and encouraging than primal-dual decomposed interior point method. The effect of contracts and ramp-rate limits of generators are also studied.â‘¡With adjustment limits of generators and operation consatraints under normal load and given heavy load considered, a model for voltage stability constrained OPF (VSCOPF) is formulated. Moreover, VSCOPF is extended to the field of dynamic OPF. With voltage stability constraints in each time interval considered, a model for voltage stability constrained dynamic OPF is also estabilished. These two models both have the similar feature with classical dynamic OPF. Therefore, the KKT matrics are also block diagonal bordered matrices. The proposed PCDIPM is applied to solve these two models. The effect of adjustment limits of generators and voltage stability indictors are also discussed.â‘¢Considering a dynamic reactive power optimization problem, a hybrid method is proposed to solve, in which the IGA and IPM are combined and mutually complemented. With different combination ways of IPM and IGA, two hybrid strategies are proposed. In a hybrid strategy, the original problem is decomposed into two optimization sub-problems, one for discrete control variables and the other for continuous variables. The optimal solution is approached through alternate iteration of IPM and IGA. In the alternative hybrid strategy, IPM is entirely embedded within IGA to evaluate the fitness of antibodies, and the hybrid IGA approximating the optimal solution through evolution of population. Simulation results show that the dynamic reactive optimization problem, which is a large-scale mixed integer nonlinear programming problem, can be effectively solved by the hybrid method.â‘£Two dynamic OPF models with discrete variables are developed, and these two mixed integer nonlinear programming problems are solved by a hybrid method based on IGA and PCDIPM. One model is the dynamic OPF with Q-type discrete control variables, in which the limits on adjustment times of shunt capacitors (reactors) and on load tap changers (OLTC) are considered. This problem is solved by alternate iteration of IGA and PCDIPM. The other model is the dynamic OPF with unit up/down variables, in which the unit minimum up/dowm times are considered. This problem is solved by a hybrid IGA method with PCDIPM embedded. Simulation results verify these two models and methods.