THEORETICAL AND EXPERIMENTAL INVESTIGATION OF ESTIMATING CHANGE POINT IN MULTIVARIATE PROCESSES VIA SIMULTANEOUS COVARIANCE MATRIX AND MEAN VECTOR

Authors

  • Alireza Firouzi School of Mechanical Engineering, Faculty of Engineering, Universiti Teknologi Malaysia, 81310 UTM Johor Bahru, Johor, Malaysia https://orcid.org/0000-0003-3149-7032
  • Noordin Mohd Yusof School of Mechanical Engineering, Faculty of Engineering, Universiti Teknologi Malaysia, 81310 UTM Johor Bahru, Johor, Malaysia
  • Muhammad Hisyam Lee Department of Mathematical Sciences, Faculty of Science, Universiti Teknologi Malaysia, 81310 UTM Johor Bahru, Johor, Malaysia
  • Robabeh Bashiri Department of Environmental Technology, Faculty of Chemistry, University of Gdańskul. Wita Stwosza 63, 80-308 Gdańsk, Poland

DOI:

https://doi.org/10.11113/jurnalteknologi.v84.17419

Keywords:

Multivariate normal process, simultaneous covariance matrix and mean vector, artificial neural networks (ANN), support vector machine (SVM), change point

Abstract

The identification of change points in statistical process control (SPC) data is the critical criterion for multivariate techniques when output is out-of-control condition. Therefore, monitoring all independent variables is essential and demands targeted attention to avoid errors at the systems control stage. However, estimating change-point in multivariate control charts is the main problem when these correlated quality characteristics monitor together. Therefore, we proposed a combination of an ensemble learning-based model of artificial neural networks with support vector machines to monitor process mean vector and covariance matrix shifts simultaneously to estimate the change point in a multivariable system. The performance of the final model indicated an estimated changing point with one sample over 6,000 simulated cases with a probability of 98 percent, which is a significantly high accuracy rating. Finding suggests the outcome of the project confirms that the proposed model can provide a precise estimating the change point by monitoring the mean vector and the covariance matrix simultaneously and, helps to identify those variable(s) responsible for an out-of-control condition. For further validation of the model, the performance of the proposed model has been compared with previous reported which confirms a better performance of the proposed model. Finally, the model was applied to monitor the performance of the solar hydrogen production system and the model identify the variables which have negative effects on the performance of the system.

Published

2021-11-21

How to Cite

Firouzi, A. ., Mohd Yusof, N. ., Lee, M. H. ., & Bashiri, R. . (2021). THEORETICAL AND EXPERIMENTAL INVESTIGATION OF ESTIMATING CHANGE POINT IN MULTIVARIATE PROCESSES VIA SIMULTANEOUS COVARIANCE MATRIX AND MEAN VECTOR. Jurnal Teknologi, 84(1). https://doi.org/10.11113/jurnalteknologi.v84.17419

Issue

Section

Science and Engineering