数据驱动的闭环系统传感器故障诊断方法  

Fault Diagnosis Method Based on Data Driven for Closed-loop System's Sensor

作　　者：李明 那文波 刘甜甜 高宇 LI Ming;NA Wen-bo;LIU Tian-tian;GAO Yu(College of Mechanical and Electrical Enginnering,China Jiliang University,Hangzhou 310018,China)

机构地区：[1]中国计量大学机电工程学院,浙江杭州310018

出　　处：《仪表技术与传感器》2020年第3期89-94,99,共7页Instrument Technique and Sensor

摘　　要：为解决传感器微弱故障难以诊断的问题,提出了一种基于数据驱动的一阶定值闭环系统中传感器微弱故障实时诊断方法。将传感器故障分类为加性故障和乘性故障,通过对系统实时工作数据的残差分析以及曲线拟合,提出了基于正常工作窗口残差分析故障检测方法,并确定故障检测阈值;基于故障动态响应过程分析的故障估计方法;基于两次一元曲线拟合分析的故障分离方法。给出了系统在线标定及故障诊断流程。在单容液位控制系统上验证了所提诊断方法的有效性及先进性。In order to solve the problem that the weak fault of the sensor is difficult to diagnose,a real-time diagnostic method for sensor weak fault in the first-order fixed-value closed-loop system based on data drive was proposed.The sensor faults were classified into additive and multiplicative faults.Through the residual analysis and curve fitting of the real-time working data of the system,the fault detection method based on normal working window residual analysis was proposed,and the fault detection threshold was determined.The fault estimation method based on fault dynamic response process analysis was proposed,the fault isolation method based on two-dimensional curve fitting analysis was proposed.The online calibration and fault diagnosis process of the system were given.The effectiveness and advancement of the proposed diagnostic method were verified on a single-capacity level control system.

关 键 词：一阶定值系统 传感器 故障诊断 数据驱动 残差分析 曲线拟合 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置]