神经网络在变压器微机保护中的应用

Application of nerve network for transformer's PC protection

为了研究变压器的保护方法,建立了变压器的电气特征线性模型,该模型无需涉及变压器铁心的非线性关系和磁滞效应,并在此基础上提出一种基于人工神经网络的变压器状态模式识别的方法。由于已经考虑了变压器铁心的非线性关系和磁滞效应。因此,在识别变压器故障状态无需附加涌流识别。仿真结果表明了该方法具有可靠、快速识别特性。

The electrical characteristics linear mode of transformer was made to research its protection method. A transformer national format identification method was expounded based on artificial nerve network. Because the method had considered the nonlinear relation and hysteresis of transformer, the upwelling need not be added to identify transformer fault. The simulation result proved its reliability and quick.