基于软测量的设备异常识别技术研究  

Abnormality Diagnosis Technology Based on Soft Sensor

作　　者：王刚[1] 贺革[1] 邱旭蒙 卫柯丞 刘文[1] 吉旭[1] 

机构地区：[1]四川大学化学工程学院,四川成都610065

出　　处：《山东化工》2015年第20期88-91,93,共5页Shandong Chemical Industry

摘　　要：针用各类传感器监测系统运行过程中的异常是企业常用方法,但由于技术障碍或成本原因已无法满足工厂在复杂环境下对设备运行稳定性、可靠性更高的要求。设备异常识别技术作为智能制造的重要技术组成部分,已成为设备异常识别的研究热点。本文通过引入软测量的技术,基于企业工艺数据、设备运行机电数据、过程实时数据等,结合数据管理、软测量模型建立和异常结果发布等相关技术,设计了设备异常识别系统模型。在实例分析中,采用GA-BP算法实现了设备状态的实时监测以及异常识别,验证了软测量技术在设备异常识别系统中算法实现的可行性。In intelligent manufacturing,equipment abnormality diagnosis is an important part and makes much sense to process system. However,due to technical obstacles or economic restrictions,it has been unable to meet the requirements of factories to develop and apply current monitoring sensors. Based on data of electrical performance,process and real-time monitor,a system architecture of soft sensing technology is designed to solve this problem by introducing soft sensor technology. Besides,the relative data management technology,soft sensor modeling technology and abnormal results publishing technology are studied here. As an example,a BP neural network optimized by genetic algorithm is proposed to deal with abnormality diagnosis and real-time monitoring of equipment status in the concrete enterprises,and the feasibility of the soft sensor system is verified.

关 键 词：智能制造 软测量 异常识别 

分 类 号：TP274[自动化与计算机技术—检测技术与自动化装置]