摘要
污水生化处理过程是一个典型的多变量、非线性、具有强外部干扰的复杂工业过程,基于主元分析(PCA)的方法在污水处理过程的故障诊断中已经得到了广泛的应用和发展。但是,主元分析存在着噪声和不确定信息描述能力不足的问题。针对污水过程的特点提出了因子分析(FA)故障诊断方法,同时拓展了因子分析监控指标以及解决了故障辨识贡献图的控制问题,为污水处理过程的有效监控提供了可行性。所提出方法在国际水协会BSM1模型上得到了有效的验证,与PCA的对比研究表明FA提供了更低的故障误报警率和更完整的不确定信息描述能力。
Wastewater treatment is a complex process with multiple variables, nonlinearity, significant extemal disturbance, and fault diagnosis based on Principal Component Analysis (PCA) has been widely used and developed. However, the fact limiting its application is noise and its weak ability of uncertainty description. Given the characteristics of wastewater process, this paper proposes a Factor Analysis (FA)-based fault diagnosis methodology, extends the FA fault detection index and overcomes the control limit problems of contribution plot, which facilitates the efficient monitoring of the wastewater BSM1 platform efficiently, and the results illustrate that the ability to describe the uncertain information. process. The proposed method is validated in the FA method offers lower false alarm rate and better
出处
《控制工程》
CSCD
北大核心
2015年第3期447-451,共5页
Control Engineering of China
基金
国家自然科学基金-青年基金(61403142)
2012年度高等学校博士学科点专项科研基金(20120172110026)
2014中央高校基本科研业务资助项目-博士启动(2014ZB0028)
2014中央高校基本科研业务资助项目-重点项目(2014ZB0043)
关键词
因子分析
污水处理
故障诊断
Factor analysis
wastewater treatment
fault diagnosis