Monitoring of Mineral Processing Operations based on Multivariate Similarity Indices

Multivariate process monitoring through covariance control charts considers changes in the relationships among process variables, but is limited by linearity assumptions. In this paper two nonlinear indicators of multivariate structure are considered, viz. mutual information and random forest proxim...

Full description

Saved in:
Bibliographic Details
Published inIFAC Proceedings Volumes Vol. 44; no. 1; pp. 9923 - 9928
Main Authors Auret, L., Aldrich, C.
Format Journal Article
LanguageEnglish
Published 01.01.2011
Subjects
Fault detection
machine learning
mineral processing
random forests
random forests
mineral processing
Fault detection
machine learning
Online AccessGet full text

Cover

More Information
Summary:Multivariate process monitoring through covariance control charts considers changes in the relationships among process variables, but is limited by linearity assumptions. In this paper two nonlinear indicators of multivariate structure are considered, viz. mutual information and random forest proximity measures. Similarity matrices are constructed from data encapsulated by sliding windows of different sizes across the time series data associated with process operations. Diagnostic metrics reflect the differences between stationary base windows representative of normal operating conditions and test windows containing new process data. A case study in mineral processing shows that better results can be obtained with these nonlinear methods.
ISSN:1474-6670
DOI:10.3182/20110828-6-IT-1002.01928