基于人工神经网络的电力系统暂态安全域估计

ARTIFICIAL NEURAL NETWORK BASED ESTIMATION OF POWER SYSTEM TRANSIENT SECURITY REGION

本文提出了一种利用人工神经网络来描述和拟合电力系统暂态安全性能的方法,介绍了一种将集总学习规则和最小二乘法结合起来的快速学习算法。利用某一简化的电力系统,对本方法进行了验证,并将所提出的快速学习方法和传统的B—P学习法进行了比较。

This paper presents an artificial neural network based method for power system transient security assessment. The basic idea is to synthesize the complex mapping-transient security region by a multi - layer perceptron. Once the learning phase demonstrates convergence, the transient security of power system can be fastly assessed. To accelerate the training process, a new learning algorithm, which combines the modified B-P method and the least square method, has been developped. The proposed method is tested on a simplified but real power system, and some important and illustrative simulation results are included.