摘要
工业过程中普遍存在噪声污染,本文在概率主元分析方法(PPCA)的基础上,把该方法推广到非高斯过程,提出一种新的基于概率独立成分分析(PICA)的过程监控方法。针对过程的非高斯和噪声信息,分别建立其对应的统计量I2和MR。通过对Tennessee Eastman(TE)过程的仿真研究,验证了该方法的可行性和有效性,较好地改善了过程的监控效果,从而更好地保证过程运行的安全、稳定性。
Noise corruption always exists in the industrial process. Based on the probabilistic principal component analysis (PPCA) method, a new process monitoring method based on probabilistic independent component analysis (PICA) was proposed, which extends PPCA to the non-Gaussian process, Two statistical quantities (I^2 and MR) were constructed for monitoring non-Gaussian and noise information of the process. A case study of the Tennessee Eastman (TE) process showed that the proposed method was feasible and efficient. The process monitoring performance was evidently improved, thus enhancing the reliability and stability of the TE process.
出处
《化工学报》
EI
CAS
CSCD
北大核心
2008年第7期1665-1670,共6页
CIESC Journal
基金
国家自然科学基金项目(60774067)~~
关键词
概率独立成分分析
非高斯
噪声污染
probabilistic independent component analysis
non-Gaussian
noise corruption