Study Of The Reactive Power Optimization For Regional Grid Based On Dynamic Division

Reactive power optimization is one of important means of economic operation to ensure security and stability of power system. Reactive power optimization, not only can improve the system voltage level, but also reduce the power loss to improve the economy of the grid.With the deep study on mathematical model of reactive power optimization, optimization algorithm and the control strategy, to be aimed at the shortcomings of the current local grid reactive power optimization which easily occurs optimization results invalid caused by inaccurate mathematical model, particle swarm optimization algorithm is easy to prone to get a local best answer and too many controlled variables need operation. The reactive power optimization scheme based on a control variable dynamic partitioning is proposed.First, an objective function that involves the least number of nodes whose voltage beyond limits, the least average of the voltage offset, the minimum power loss and the least number of equipment which need to be operated is proposed. It is handled by using the methods of an improved stratified sequence method combined with particle swarm optimization. Second, a hybrid method which involves particle swarm optimization, discrete quantum particle swarm optimization algorithm and chaos optimization strategy is proposed in this paper to prevent particles fall into local optimum.In the initialization of the particle swarm, part of particles randomly generated, some particles was used extreme value in theory, mixed encoding was used in control variables. Then, for control variables are much more, in order to reduce the target search space dimension and the computation time, this paper was according to the actual operation of the grid and based on the node reactive power changes on the sensitivity of the system active network losses, and more limited to electrical wiring of the adjacent nodes to control the variable dynamic partitioning. The control variables which belong to this partition is set up the candidate reactive power compensation point, if only by optimizing the control variables of this region cannot meet the voltage requirements, in accordance with the electrical wiring, to extend the range of partitions until a solution to meet the termination condition is found. Based on this partition method, the population and the inertia weight factor are set up a function which change along with number of control variables to improve the performance of the algorithm.In this paper VC++6.0prepare reactive power optimization software is used, The simulation results of a IEEE-30systems and an actual power network,comparing to particle swarm optimization from the nodal voltage, the average offset voltage and active power net loss, etc. In this paper the simulation results was proved that the reactive power optimization scheme was Superior, effective and feasible.