基于人工神经网络与遗传算法的暂态稳定评估

Transient stability assessment method based on back propagation algorithm andgenetic algorithm

提出了ＢＰ算法和遗传算法相结合的电力系统暂态稳定性评估新方法,选取了以稳态特征量、故障初始时刻特征量和暂态特征量构成的综合特征量作为稳定性评估系统的输入特征量。以０至１之间的小数作为评估系统的输出值,根据输出值的大小判定系统是否稳定。理论分析及仿真结果表明,本方法解决了ＢＰ网络易陷入局部最小的问题和遗传算法应用于复杂的多变量优化系统时速度很慢的问题,并且对不同电力系统暂态稳定性评估具有一定的通用性。

BP Algorithm is combined with Genetic Algorithm to generate a new scheme to assess power system transient stability. BP-GA Algorithm can avoid both the local minimum of BP network and also the low speed of Genetic Algorithm when it is used in complex system. A set of comprehensive features, composed of steady state features, fault initiation moment features and transient state features, are defined as inputs in the paper based on analysis and simulations. Transient stability assessment using these selected features can pay attention to both speed and accuracy.