摘要
针对传统仪表由于资源有限、成本控制等原因导致无法使用复杂过程趋势诊断方法的问题,分析现有基于差分值和峭度值的两种过程趋势诊断方法原理及实现方式,比较两者优缺点,提出一种结合两者优点、具有统计特性自学习特点的过程趋势诊断方法,并将该方法在压力变送器上进行验证。结果表明:该方法计算量小,占用资源少,能够通过实时采集的过程变量,分析过程变量的变化趋势,并根据预设判据实现过程趋势的准确诊断。
Due to limited resources and cost,common instruments can not use complicated diagnosis methods for process trend.This paper analyzes the principle and implementation of two existing process trend diagnosis methods based on difference algorithm and kurtosis,compares their advantages and disadvantages and proposes a process trend diagnosis method with statistical characteristics and self-learning characteristics,then the method is verified on the pressure transmitter.The actual verification results show that the method has the advantages of small amount of calculation,less resource occupation,real-time acquisition of process pressure value,analysis of change trend and accurate diagnosis according to the preset criteria.
作者
包伟华
徐华东
邱云周
赵勇
贾根团
BAO Weihua;XU Huadong;QIU Yunzhou;ZHAO Yong;JIA Gentuan(Shanghai Automation Instrumentations Co.,Ltd.,Shanghai 200072,China;Shanghai Institute of Microsystem and Information Technology,Chinese Academy of Sciences,Shanghai 200050,China)
出处
《流体测量与控制》
2021年第4期1-5,共5页
Fluid Measurement & Control
基金
上海市人工智能创新发展资助项目(2019-RGZN-01037)。
关键词
统计特性
自学习
过程诊断
仪表
statistics-characteristics
self-learning
process diagnosis
instrument
