人工神经网络及在线电压无功静态安全状态分类

ANN AND ON-LINE CLASSIFICATION OF VOLTAGE/REACTIVE POWER STATIC SECURITY STATES

本文简要说明了几种在电力系统领域较有应用前途的人工神经网络(ANN)的应用特点。应用Kohonen自组织影射ANN模型对电压无功静态安全分类进行了研究,结果表明Kohonen模型的概括能力不仅能对在训练集中包括的状态,而且能够容许对在训练集中未包括的状态进行正确分类。经系统测试,结果证明本方法可行,适用于在线应用。 安全分析,Kohonen人工神经网络。

The appling Features of several artificial neural nets (ANN) which are of good prospects in power systems are concisely explained in this paper. Using Kohonen self-organizing feature map, thc classification of voltage/reactive power static security states is studied, and the results demonstrated that the gencralization capability of the Kohonen network permits the correct classification of power system states whether are included in the training sets or not. By systematic test, the results show the effectiveness ofthe proposed method which suits to the on-line application.