基于相量测量技术和模糊径向基网络的暂态稳定性预测

APPLICATION OF PMU AND FUZZY RADIAL BASIS FUNCTION NETWORK TO POWER SYSTEM TRANSIENT STABILITY PREDICTION

提出一种新的基于模糊聚类的径向基神经网络及其训练算法,利用同步相量测量装置获得的故障后短时间内各发电机的功角,经简单运算后作为神经网络的输入,其输出为多机电力系统稳定性的分类结果。对49机实际系统在不同接线方式和故障位置等条件下,进行了有无切机控制两种情况下的数值仿真实验,结果表明所提出的方法对系统的失稳预测和切机控制决策是有效的,神经网络训练时间短。

A new radial basis function network based on fuzzy clustering (FCRBFN) and its learning algorithm is proposed in this paper. The FCRBFN, whose inputs are simple function of generator rotor angles after fault measured by PMUs, is used to predict transient stability of multimachine power system. The numerical results of a real 49 machine power system demonstrate that the proposed method is effective to transient stability prediction with and without generator shedding, considering different operating conditions and fault locations. The learning process is considerable fast and the neural network have very high classification precision.