Multivariate Exponentially Weighted Moving Average (MEWMA) and Multivariate Exponentially Weighted Moving Variance (MEWMV) Chart Based on Residual XGBoost Regression for Monitoring Water Quality

Nur Sulistiawanti, Muhammad Ahsan, Hidayatul Khusna

Research output: Contribution to journalArticlepeer-review

2 Citations (Scopus)

Abstract

Water Treatment Plant (WTP) in Palopo City, Indonesia, is one of the national companies engaged in the provision of clean water services for the people of Palopo City. There are 5 WTP managed in Palopo City. Testing the quality of production water at each WTP uses several interrelated quality characteristics, including turbidity, pH, and residual chlorine. The data used is obtained from the results of testing the quality of production water at each WTP in July-December 2021 for a daily period. In this study, the method used to control the quality of production water at each WTP is the Multivariate Exponentially Weighted Moving Average (MEWMA) control chart to control the process mean and the Multivariate Exponentially Weighted Moving Variance (MEWMV) control chart to control process variability based on residual XGBoost regression to overcome autocorrelation. In phase I, the process has been statistically controlled for both the variance and the process mean. The weighting values for the process variability are ω = 0.4 and λ = 0.4, and for the process mean is λ = 0.3. In phase 2, based on the data analyzed only WTP V in Batupapan whose process variability has been statistically controlled. The other WTPs in phase 2 have not been statistically controlled for both the variability and the process mean.

Original languageEnglish
Pages (from-to)1001-1008
Number of pages8
JournalEngineering Letters
Volume31
Issue number3
Publication statusPublished - 2023

Keywords

  • Autocorrelation
  • MEWMA
  • MEWMV
  • Water Quality
  • XGBoost Regression

Fingerprint

Dive into the research topics of 'Multivariate Exponentially Weighted Moving Average (MEWMA) and Multivariate Exponentially Weighted Moving Variance (MEWMV) Chart Based on Residual XGBoost Regression for Monitoring Water Quality'. Together they form a unique fingerprint.

Cite this