基于故障树的天然气流量计量仪表故障自动化识别研究  

Research on Automatic Fault Identification of Natural Gas Flow Measurement Instruments Based on Fault Tree

作　　者：王晨煜 范劲松[2] 程伟[2] 彭云[2] WANG Chenyu;FAN Jinsong;CHENG Wei;PENG Yun(School of Petroleum Engineering,Chongqing University of Science&Technology,Chongqing 401331,China;Chongqing Gas Mine,Southwest Oil&Gasfield Company,Chongqing 401120,China)

机构地区：[1]重庆科技学院石油与天然气工程学院,重庆401331 [2]西南油气田分公司重庆气矿,重庆401120

出　　处：《自动化与仪表》2023年第12期74-78,90,共6页Automation & Instrumentation

摘　　要：为实现故障识别效果的改善,为天然气流量的稳定计量提供保障,该文提出了基于故障树的天然气流量计量仪表故障自动化识别方法。通过故障树分析法获得天然气流量计量仪表的故障映射,确定故障产生的因果关系,采用KCPA特征集成算法提取天然气流量计量仪表故障特征,将故障因果关系与故障特征作为基于LS-SVM的多分类器组的输入,通过自整定权值的决策模板法(SWDT)评判每个LS-SVM分类器的故障识别性能,将初始故障识别结果作为依据,为各LS-SVM分配决策权值,实现天然气流量计量仪表故障的自动识别。实验结果表明,该方法故障识别精度达到94%左右。A fault tree based automatic fault identification method for natural gas flow measurement instruments is studied to improve the effectiveness of fault identification and provide guarantee for stable measurement of natural gas flow.The fault mapping of natural gas flow meters is obtained through fault tree analysis,and the causal relationship of the fault is determined.The KCPA feature integration algorithm is used to extract the fault characteristics of natural gas flow meters.The fault causal relationship and fault characteristics are taken as the input of LS-SVM based multi classifier group,and the fault recognition performance of each LS-SVM classifier is evaluated through the self-tuning weight decision template method(SWDT),based on the initial fault identification results,decision weights are allocated to each LS-SVM to achieve automatic identification of faults in natural gas flow metering instruments.The experimental results show that the fault identification accuracy of this method is about 94%.

关 键 词：故障树 天然气 流量计量仪表 故障识别 KCPA特征集成 多分类器 

分 类 号：TP274[自动化与计算机技术—检测技术与自动化装置] TM76[自动化与计算机技术—控制科学与工程]