论文部分内容阅读
针对变压器故障的特征,结合变压器油中气体分析法以及三比值法,提出了基于遗传算法改进极限学习机的故障诊断方法。由于输入层与隐含层的权值和阈值是随机产生,传统的极限学习机可能会使隐含层节点过多,训练过程中容易产生过拟合现象。该方法运用遗传算法对极限学习机的输入层与隐含层的权值与阈值进行优化,从而提高模型的稳定性和预测精度。将诊断结果与传统的基于极限学习机故障诊断进行对比,结果表明,基于遗传算法改进极限学习机变压器故障诊断的精度更高。
Aiming at the characteristics of transformer fault, combined with gas analysis method in transformer oil and triple ratio method, a fault diagnosis method based on genetic algorithm to improve extreme learning machine is proposed. Because the weights and thresholds of input layer and hidden layer are randomly generated, the traditional extreme learning machine may make the nodes of hidden layer too much, and the over-fitting phenomenon easily occurs in the training process. The method uses genetic algorithm to optimize the weights and thresholds of the input layer and the hidden layer of the extreme learning machine, so as to improve the stability and prediction accuracy of the model. Comparing the diagnostic results with the traditional fault diagnosis based on extreme learning machine, the results show that the improved fault diagnosis based on genetic algorithm is more accurate.