质量相关的回转窑故障诊断

Quality relevant fault diagnosis of rotary kiln

针对水泥烧制过程工艺复杂,各变量数据之间存在强耦合关系的特点,提出了偏最小二乘(partial least squares,PLS)算法结合贡献图分析的方法,应用到回转窑系统中进行故障诊断。该文选取质量变量,建立过程变量和质量变量之间的回归关系,构建PLS模型实现通过过程变量数据对质量变量数据的预测。设置一定置信度的控制限,对超过控制限的故障进行报警。通过相对贡献图方法找出对故障贡献最大的变量,分析故障原因。仿真结果表明,该方法可以及时检测到影响水泥品质的故障并确定故障原因。

In view of the complex process of cement firing and the strong coupling between the variable data,the partial least squares(PLS)algorithm combined with the contribution plot analysis method is proposed to the rotary kiln system for fault diagnosis.The quality variables are selected and the regression relationship between process variables and quality variables is established in this paper.PLS model is built to predict the quality variable data from the process variable data.By setting acontrol limit of a confidence coefficient,the fault exceeding the control limit will be alarm.The variables contributing the most to the fault are found out by the contribution plot method and the fault causes will beanalyzed.The simulation result shows that this method can detect the faults affecting the quality of cement in time and determine the cause of the faults.