Operating Performance Assessment of Non-Gaussian Processes Based on Performance-Relevant Independent Component Analysis

Abstract

In this study, a novel operating performance assessment method based on performance-relevant independent component analysis (PR-ICA) is proposed for non-Gaussian industrial processes. Compared with the original ICA, the proposed PR-ICA takes the comprehensive economic index (CEI) into account, and the objective is to maximize the non-Gaussianity of performance-relevant independent components (PR-ICs) and statistical dependency between PR-ICs and the CEI simultaneously. When applying PR-ICA to online assessment, it shows strong robustness to the interference of performance-irrelevant variation information and high sensitivity to the change of performance-relevant variation information. Finally, the efficiency of the proposed method is illustrated with the case of cyanide leaching gold.