基于LMDI分解法的火电厂热力设备故障自动化诊断技术  

Automatic Fault Diagnosis Technology of Thermal Equipment in Thermal Power Plant Based on LMDI Decomposition Method

作　　者：马战南 蒋俊英 MA Zhannan;JIANG Junying(Guoneng Shanxi Hequ Power Generation Co.,Ltd.,Xinzhou 036599,China)

机构地区：[1]国能山西河曲发电有限公司,忻州036599

出　　处：《自动化与仪表》2025年第2期89-92,98,共5页Automation & Instrumentation

摘　　要：针对已有故障诊断技术存在的数据特征提取性能差、故障诊断精度低等问题,提出基于LMDI分解法的火电厂热力设备故障自动化诊断技术研究。采集并预处理热力设备运行数据,基于改进LMDI分解法构建热力设备运行数据分解模型,计算运行数据对于热力设备故障的贡献度,确定热力设备关键运行数据,提取关键运行数据特征,计算其与已知故障类型特征的相关系数,依据制定规则实现热力设备故障的自动化诊断。测试结果显示,应用该技术提取的关键运行数据特征数值与实际特征数值最大误差为0.04,故障诊断正确率达到了93.75%。Aiming at the problems of poor data feature extraction performance and low fault diagnosis accuracy of existing fault diagnosis technologies,an automatic fault diagnosis technology for thermal equipment of thermal power plant based on LMDI decomposition method is proposed.Collect and pre-process the operation data of thermal equipment,build the thermal equipment operation data decomposition model based on the improved LMDI decomposition method,calculate the contribution degree of operation data to the thermal equipment failure,determine the key operation data of thermal equipment,extract the characteristics of the key operation data,and calculate the correlation coefficient between them and the characteristics of known fault types.Automatic fault diagnosis of thermal equipment is realized according to rules.The test results show that the maximum error between the feature values of key operation data extracted by the proposed technology and the actual feature values is 0.04,and the correct diagnosis rate reaches 93.75%.

关 键 词：故障诊断 火电厂热力设备 自动化 运行数据分解 改进LMDI分解法 运行数据预处理 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置]