摘要
针对绝缘状态预测前未能解调变压器的差分频谱,存在绝缘电阻拟合误差大、绝缘子检测精度低和预测时间长等问题,提出基于模糊神经网络的电力变压器绝缘状态预测方法。基于变压器原理,对微分谱线进行解调处理,分析变压器特征量,构建模糊神经网络,引入鸡群优化算法改进模糊神经网络,将获取的变压器特征量输入优化后的模糊神经网络中,完成电力变压器绝缘状态的动态预测。实验结果表明,运用该方法进行绝缘状态预测时,绝缘电阻拟合误差小、绝缘子检测正确率高,以及预测时间短。
Aiming at the problems that the differential spectrum of transformer is not demodulated before insulation state prediction,such as large insulation resistance fitting error,low insulator detection accuracy and long prediction time,a power transformer insulation state prediction method based on fuzzy neural network is proposed.Based on the principle of transformer,the differential spectral line is demodulated,the transformer characteristic quantity is analyzed,a fuzzy neural network is constructed,the optimized chicken swarm algorithm is introduced to improve the fuzzy neural network,and the obtained transformer characteristic quantity is input into the optimized fuzzy neural network to complete the dynamic prediction of power transformer insulation state.The experimental results show that when using this method to predict the insulation state,the fitting error of insulation resistance is small,the accuracy of insulator detection is high,and the prediction time is short.
作者
王峰
WANG Feng(State Grid Jiangsu Integrated Energy Services Co.,Ltd.,Nanjing 210019,China)
出处
《机械与电子》
2022年第3期50-53,57,共5页
Machinery & Electronics
关键词
模糊神经网络
电力变压器
绝缘状态
预测方法
鸡群算法
fuzzy neural network
power transformer
insulation state
prediction method
chicken swarm algorithm
