摘要
提出了一种利用人工神经网络进行电力系统暂态稳定分析的方法。该神经网络取故障后系统暂态量为特征量,采用改进BP算法进行训练。将样本空间进行模式分类,并对不同类样本作不同处理。以实际系统为例,将选用暂态特征与选稳态特征进行比较。
In the paper, a kind of ANN method is used to attack the difficulty in analyzing the transient stability of a power system. Compared with the other analyzing theories, the method selects the postfault transient features as the inputs of the neural network, classifys the samples into different classes according to their stability and treats them respectively. Several points are proposed to improve the training speed of BP algorithm. This method is then employed to analyze the transient stability of a real system and compared with the method using stable features in the meantime. The results show that the new method is more effective.
出处
《电力系统自动化》
EI
CSCD
北大核心
1996年第11期23-26,共4页
Automation of Electric Power Systems
关键词
神经网络
电力系统
稳定性
artificial neural network (ANN) BP algorithm transient stability analysis of power system pattern classification