The Research Of Partition Computation For Reactive Power Optimization Based On Improved BCC Algorithm

Reactive power optimization is not only an effective mean to ensure power system security and economical operation, but also one of significant measures to improve system's voltage quality. In view of the growing of the power scale and the improving of the online analytical requirements, in this article, the large system is partitioned and the global reactive power optimization is decomposed into multiple sub-parallel network optimizations from the perspective of reducing the size of the calculation.Firstly, introduce the new way of power system partition based on D-S evidence fusion theory. Calculate partition support degrees of basic probability assignment and fuzzy sample similarity degree based on electrical distance, and then fuse these two support degrees using D-S evidences fusion which form objective evidence. If a node objective partition has enough evidence, the partition results are obtained directly. Otherwise, the objective evidence is comprehensive evaluated considering the expert group's advice. Thus conclusive result of network partitioning for reactive power/voltage control which fuses subjective and objective evidence is obtained. The simulation shows that the proposed method is feasible and effective.Then, each sub-network reactive power optimization using the mathematical model without penalty factor strategy and improved BCC optimize algorithm. Introduce the mathematical model without penalty factor strategy. The model uses the possible bacterial proportion of bacteria to guide bacteria to search the best possible value quickly. Dynamic adjustments of speed and perception range, Gaussian mutation mechanism are introduced to improve the accuracy; mapping factor is introduced to improve the performance of BCC algorithm to solve the problem of discrete domain. The simulation results show that the improved BCC algorithm has good optimize performance, and the mathematical model is simple and exercisable, combination can quickly come to a reasonable reactive power optimization strategy. Based on that, combined with characters of the offshore wind farm system, the model and method are extended to solve reactive power optimization problem of offshore wind farm system.Finally, the theoretical framework of global reactive power optimization based on MAS is proposed. Details of different Agents'functional modules, the data need to be exchanged between the regional optimization Agent and co-ordination mechanism, and so on, are expatiated. Through the partition of large-scale system, complex global reactive power optimization problem is decomposed into multiple sub-network parallel optimization problems, can increase the computing speed, and promote the development of the global for reactive power optimization scheduling.