基于粗糙集-BP神经网络的变压器故障快速诊断研究
Research on Rapid Diagnosis of Transformer Faults Based on Rough Set and BP Neural Network
周雪莹 1潘赟颖2
作者信息
- 1. 国网上海市电力公司浦东供电公司,上海 200122
- 2. 国网上海金山供电公司,上海 200540
- 折叠
摘要
由于现有故障诊断方法的可靠性较差,达不到准确提取的要求,为此,基于粗糙集-BP神经网络进行变压器故障的快速诊断研究.根据故障数据,运用粗糙集理论建立故障诊断决策表,获得了变压器的故障类型,分别为正常、低温过热、低能放电、高温过热和高能放电,并得到了最优的决策规则.运用并行化的BP神经网络算法获得故障信息预测值.将预测值和决策规则表中的决策值进行比较,通过输出神经网络比值判断具体的故障类型,从而完成故障诊断.结果表明,所提故障诊断方法具有较好的神经网络性能,能够准确地诊断出变压器的故障.
Abstract
Due to the poor reliability of existing fault diagnosis methods,they cannot meet the requirements of accurate extraction.Therefore,research on fast diagnosis of transformer faults based on rough set and BP neural network is conducted.Based on the fault data,a fault diagnosis decision table was established using rough set theory,and the fault types of the transformer were obtained,including normal,low-temperature overheating,low-energy discharge,high-temperature overheating,and high-energy discharge,and the optimal decision rules were obtained.Using parallelized BP neural network algorithm to obtain fault information prediction values.Compare the predicted value with the decision value in the decision rule table,and determine the specific fault type by outputting the neural network ratio to complete the fault diagnosis.The results show that the proposed fault diagnosis method has good neural network performance and can accurately diagnose transformer faults.
关键词
粗糙集/BP神经网络/变压器/故障诊断Key words
rough set/BP neural network/transformer/fault diagnosis引用本文复制引用
出版年
2024
NETL
NSTL
万方数据