一种混合遗传算法在HVDC定电流控制器参数优化中的应用

为了改进遗传算法的局部收敛能力,提出了一种新的融合遗传算法和单纯形法的混合遗传算法。该算法兼具遗传算法的全局收敛性和单纯形法良好的局部收敛能力,当遗传算法陷入停滞后,由单纯形法引导进化方向。针对直流输电定电流控制器参数优化的具体问题,设计了一套系统化混合遗传算法寻优方案,并详尽叙述了该方案的流程及具体实现步骤。该方案选取ITAE为目标函数,直流输电准稳态模型为目标函数的计算模型;PI参数的可行域由基于直流输电PSCAD/EMTDC模型仿真的直流电压稳定性计算确定。最后通过CIGRE HVDC Benchmark模型的PSCAD/EMTDC仿真验证所设计参数优化方案的有效性。

An application of hybrid genetic algorithm in the parameters optimization of HVDC constant current controller

First of all, in order to improve the local search ability of genetic algorithm, a new hybrid genetic algorithm combining genetic algorithm and simplex method is presented; in which both the global convergence of genetic algorithm and local search ability of simplex method are in full use; when the evolution of genetic algorithm is bogged down, the simplex method is involved to lead the direction. Then aiming at optimizing the parameters of HVDC constant current controller, a systematic scheme using the hybrid genetic algorithm is designed. After an expatiation upon the basic flow and main steps of the hybrid genetic algorithm, the implementation of parameters optimization is detailed. The ITAE is selected as objective function in the scheme. By building the HVDC Quasi-Steady-State (QSS) models, a reasonable calculation models are achieved. The DC voltage stability in the process of PSCAD/EMTDC simulation is judged to search the PI parameters boundary. At last the effectiveness of the proposed scheme is demonstrated by PSCAD/EMTDC simulation results on the CIGRE HVDC Benchmark model.