自适应遗传算法在变压器局部放电超高频模式识别中的应用

Application of adaptive genetic algorithm to ultra^high^frequency partial discharge pattern recognition in transformers

采用自适应遗传算法(AGA)作为神经网络的学习算法,对实验室中变压器局部放电超高频自动识别系统检测到的5种放电类型进行了模式识别。实验结果表明,AGA神经网络解决了BP神经网络对初始权值敏感、收敛速度慢和容易局部收敛的问题,具有较高的识别率和较强的推广能力,可以很好地应用于变压器局部放电的超高频模式识别中。

An automated recognition system of ultra^high^frequency (UHF) partial discharge (PD) designed by the authors has been put forward to study the discharge properties in transformers. This paper presents adaptive genetic algorithm (AGA) to train neural network (NN) to distinguish between basic types of defects in transformers.Test results show that AGA^NN,as compared with BP^NN,can overcome slow convergence and possibility of being trapped at locally minimum value.Thus,the convergence, discrimination and generalization ability of AGA^NN is improved remarkably.