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

The load demand of power system always changes with time. To keep power system operating safely and economically, control variables should be frequently adjusted to follow the change of load demand. However, the values of control variables in each time interval are correlated by dynamic constraints. Therefore, the optimal power flow in a time horizon should be formulated and solved as a unified problem, dynamic optimal power flow (OPF) problem. As an expansion of OPF in time scale, dynamic OPF is significant important for coordinating control of power system. In this thesis, based on interior point method (IPM) and immune genetic algorithm (IGA), the following issues are investigated: method for dynamic OPF, voltage stability constrained OPF, method for dynamic reactive power optimization, and dynamic OPF with discrete control variables. The main achievements are as follows:1) For a dynamic OPF problem with ramp-rate limits and contract of generators, the KKT matrix of primal-dual IPM is a blocked diagonal and band-edge matrix. Based on the decomposed primal-dual IPM, a decomposed predictor-corrector IPM is proposed. In the proposed method, the efficiency of predictor-corrector process is improved by using blocked LU decomposition.2) Assumed that generators could be properly adjusted in the process of load increasing, a new model for voltage stability constrained OPF is established. In the model, if power system is static security under heavy load, then the voltage stability of power system is acceptable. Moreover, the model is extended to the field of dynamic optimal power flow. With voltage stability constraint of each time interval considered and dynamic constraint taken into account, a model for voltage stability constrained dynamic OPF is developed.3) Based on IGA and IPM, a hybrid method is proposed for dynamic reactive power optimization. In the hybrid method, IGA is applied to solve discrete variables while IPM is employed to optimize continuous variables. By the mutual complement of IGA and IPM, the efficiency of the proposed method is improved. With different combination styles, there are alternately iterative hybrid method and entirely nested hybrid method. In the alternately iterative style, discrete control variables and continuous variables are decomposed, and the optimal solution is approached through alternately iterative of IPM and IGA. In the entirely nested style, IPM is entirely nested within IGA to evaluate the fitness of antibody, and the optimal solution is approached by laminate optimization problems where the design variables are the discrete variables, and the auxiliary variables are continuous variables.4) Considering the action number constraints of switchable capacitor (reactor) groups and on load tap changers (OLTC), as well as the ramp-rate limits and contracts of generators, a dynamic OPF with reactive power discrete variables is formulated. An alternately iterative hybrid method based on decomposed predictor-corrector IPM and IGA are proposed to solve the problem. Moreover, considering the minimum start-up time and minimum shut-down time of units, as well as the ramp-rate limits and contracts of generators, a dynamic OPF with unit commitment is formulated. An entirely nested hybrid method based on decomposed predictor-corrector IPM and IGA are proposed to solve the problem.