基于改进DE算法的负荷建模参数辨识

Parameters identification of load modeling based on improved differential evolution algorithm

负荷建模对电力系统运行及其控制起着重要的影响,主要有基于元件和基于量测两种建模方法。基于量测的负荷建模已广泛用于实践中,当考虑负荷的时变性后,基于量测的负荷建模即成为数学优化问题。因此,寻找到有效可靠的优化方法用于负荷建模参数辨识具有重要的现实意义。微分进化(DE)算法具有全局寻优能力,对初值不敏感,经改进后还可以加快收敛速度并防止出现早熟现象,因此可用于负荷建模参数辨识实践中。将DE算法用于负荷建模参数辨识实践中,在现有改进算法的基础上,借鉴遗传算法引入移民策略以防止早熟现象,通过对两个实测建模参数辨识实例的分析,表明经改进后的DE算法具有比改进遗传算法、蚁群算法和粒子群算法更好的性能。

Load modeling plays an important impact on the power system operation and control, and there are two ways to form the model, components-based modeling and measurement-based modeling. Load modeling based on measurements has been widely used in practice, when considering the load variability, it＇s a mathematical optimization problem. Therefore, to find effective and reliable method for optimizing load modeling parameter is important and of realistic significance. Differential Evolution （DE） algorithm has global optimization capabilities, which is not sensitive to the initial value .The improved one can speed up the convergence rate and prevent premature, therefore, it can be used to load modeling to identify parameters in practice. Through the analysis of two measured identification results, improved differential evolution has better convergence, preventing pre-mature and precocious performance than improved genetic algorithm.