基于ACS-SA文化基因算法的BP神经网络变压器故障诊断

Fault Diagnosis of Transformer Based on BP Neural Network and ACS-SA

针对BP神经网络在变压器故障诊断上存在的不足,提出基于ACS-SA文化基因算法的BP神经网络变压器故障诊断方法。在实际系统中,针对缺乏准确的变量参数估计,将边界变异策略和自适应步长策略引入标准布谷鸟算法中;提出一种在改进的布谷鸟算法中结合局部搜索策略的文化基因算法;建立BP神经网络变压器故障诊断模型,并用文化基因布谷鸟算法优化BP神经网络的权值和阈值。仿真实验及对比研究结果表明,该算法能准确有效地识别变压器的故障类型,较其他算法(CS-BP神经网络算法和POS-BP神经网络算法)有更高的准确率,为变压器故障诊断提供一种新思路。

Aiming at the deficiency of BP neural network in the fault diagnosis of transformer,to tackle this prob-lem,a BP neural network transformer fault diagnosis method based on ACS-SA culture genetic algorithm is pro-posed. In the practical system,in view of the lack of accurate variable parameter estimation,the boundary mutationstrategy and adaptive step strategy are introduced to the standard cuckoo algorithm.Puts forward a cultural genetic al-gorithm combining improved cuckoo algorithm and local search strategy. Establish the fault diagnosis model of trans-former based on BP neural network,and cultural gene cuckoo algorithm is used to optimize the weights and thresh-olds of BP neural network. The final findings show that,this method is accurate and effective in identifying the faulttype of the transformer,and compare with other algorithms（CS-BP neural network algorithm and POS-BP neural net-work algorithm）,the method has a higher accuracy,as well as provides a new way of thinking for transformer fault di-agnosis.