基于主元分析的挖掘技术在过程监控中的应用

Application of data tapping technique in monitoring based on principal element analysis

在连续重整加热炉等设备的生产运行中,集散控制系统的过程监控功能存在诸多不足之处,为此应用数据挖掘技术中的主元分析方法,通过建立数学模型,从集散控制系统自身所带的大量数据中提取有用信息,实现对过程数据的有效压缩和准确的恢复。随后,利用多变量统计分析方法对该加热炉运行过程实施监控分析,通过绘制T2图和SPE控制量图判断过程的运行状态,从而监控整个生产过程,及时发现和预防故障的产生,实现对所有可能故障最大灵敏度和鲁棒性的量度。

In the process of continuous and reforming heating stove operation, the method of principal element analysis in data tapping is used to extract useful information in a distributed control system to implement an effective compression and accurate restoration for process data by establishing math model. And a method of multi-variable statistical analysis is deployed to detect the stove operation. T2 diagram and SPE diagram are developed to judge the status of its operation, by which the production process is monitored to find out and prevent the failures in production, thus the measurement for maximum sensibility and robustness of possible failures is implemented.