回转窑烧成系统故障诊断方法研究  

Research on Fault Diagnosis Methods for Rotary Kiln Sintering System

作　　者：艾红[1] 丁俊龙 刘云龙 AI Hong;DING Jun-long;LIU Yun-long(School of Automation,Beijing Information Science and Technology University,Beijing 100192,China)

机构地区：[1]北京信息科技大学自动化学院,北京100192

出　　处：《控制工程》2022年第2期223-230,共8页Control Engineering of China

基　　金：国家自然科学基金资助项目(61903043);北京市自然科学基金资助项目(4162025)。

摘　　要：针对水泥烧成系统过程变量繁多、变量间静态关系耦合强等特点,采用因子分析方法建立静态过程监控模型。针对系统时序相关问题,结合经典动态主元分析DPCA方法和典型变量分析CVA方法,提出典型变量动态主元分析CVDPCA过程监控方法,有效解决了DPCA方法扩展后的数据矩阵维度大等不足之处。将算法用于水泥烧成系统故障检测,结果表明该算法能准确识别故障和更早检测到微小渐变故障。将CVA和DPCA算法相结合,可以同时监控动态过程和静态关系,且不需要大量的故障数据建立故障模型池,具有一定研究价值。Aiming at the cement sintering system featured by diversified variables, strong coupling between static relations of variables and so forth, a static process monitoring model is established with factor analysis(FA). Aiming at problems related to system time sequences, a process monitoring method of canonical variate dynamic principle component analysis(CVDPCA) is proposed in combination with dynamic principal component analysis(DPCA) and canonical variate analysis(CVA). With this approach, defects of the DPCA method, such as large data matric dimensions after expansion, are solved effectively. The algorithm is applied to fault detection of the cement sintering system. The results indicate that with this algorithm, faults can be recognized accurately and minor gradually-changing faults can be detected more quickly. Combining the CVA algorithm which has high efficiency for dynamic process fault detection with DPCA algorithm which can effectively monitor the change of static process relationship, dynamic process and static relation can be monitored at the same time. This fault separation method does not need a large number of fault data to establish a fault model pool, which has certain research value.

关 键 词：水泥烧成系统 故障诊断 因子分析 典型变量分析 温度 回转窑 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置]