数据驱动蒸馏塔的安全生产的研究  

Research of safety production of distillation tower based on the data driven

作　　者：李毅 常勇[2] 

机构地区：[1]中海石油(中国)有限公司天津分公司,天津塘沽300452 [2]天津理工大学自动化学院天津市复杂系统控制理论及应用重点实验室,天津300384

出　　处：《计算机与应用化学》2015年第10期1191-1196,共6页Computers and Applied Chemistry

基　　金：国家自然科学基金资助项目(天津市中青年骨干创新人才培养计划基金资助项目(20130830);基于轨迹灵敏度的随机网络控制系统不敏感控制(61403279))

摘　　要：流程工业过程的故障检测对于企业安全生产和提高产品质量是十分重要的,基于数据驱动的方法相比较传统故障诊断方法,不需要详细了解过程机制和模型,满足了流程工业复杂难于建立精确模型的特点而引起广泛关注。主元分析法是基于数据驱动方法中典型的代表,本文首先研究了主元分析法在田纳西伊斯曼仿真过程中的应用,并以工业过程中的蒸馏塔单元为平台,将主元分析法应用到蒸馏塔故障检测中,实验结果表明:主元分析法能够准确及时的检测出蒸馏塔的故障,对蒸馏塔的故障检测有显著的指导作用,实现了基于数据驱动蒸馏塔的故障诊断,保证了企业的安全生产。Fault detection of process industry in the process is very important for enterprise production safety and improving the quality of our products, compared with the traditional fault diagnosis methods based on the data driven approach do not need to understand the process mechanism and model in detail, Because of meeting the characteristics of process industry complex is difficult to establish accurate model caused wide public concern. Principal component analysis method is the typical representative of methods based on the data driven, this paper studied the application of the principal component analysis in Tennessee Eastman process, and taking the unit of industrial distillation tower in the process as a platform, applying the principal component analysis method to fault diagnosis of distillation tower, the experimental results showed that principal component analysis （PCA） method could accurately and timely detect the fault in the distillation tower, make an significant guidance to the fault detection of the tower, realize the fault diagnosis of distillation tower based on data driven, and ensure the enterprise＇s safety production.

关 键 词：安全生产 故障检测 数据驱动 主元分析 田纳西伊斯曼 蒸馏塔单元 

分 类 号：TQ051.8[化学工程]