Reducing Autocorrelation Effect in the Control Chart for PM10 Curves

ABDUL, AZIZ JEMAIN and NORSHAHIDA, SHAADAN and SAYANG, MOHD DENI (2014) Reducing Autocorrelation Effect in the Control Chart for PM10 Curves. In: International Conference on Advances In Applied Science and Environmental Engineering - ASEE 2014, 02 - 03 August, 2014, Kuala Lumpur, Malaysia.

[img]
Preview
Text
20141118_094605.pdf - Published Version

Download (763kB) | Preview
Official URL: https://www.seekdl.org/conferences/paper/details/3...

Abstract

The traditional control chart for industry often requires identical and independent (i.i.d) assumptions. Thus, some modification on the control chart used in the industry need to be done before it can be used in the environmental application due to the difference in the nature of the data. The recorded environmental data are usually multivariate, correlated, and non-stationary and can be expressed as a function of time. In this present study, a control chart for monitoring curves data is to be applied where the hourly recorded data within a one day period is treated as multivariate point data and the occurrence is treated as a function of time. The control chart is constructed using the Functional Principal Component Analysis (FPCA) model. In the monitoring of quality indices overtime, autocorrelated data often negatively affects the performance of a control chart resulting into increase in the number of false alarm. Data pre-whitening approach is proposed during the monitoring phase of analysis to tackle the problem of data autocorrelation. Based on the seven years (2004-2010) recorded data from the Shah Alam air quality monitoring station which is located in the Selangor state of Peninsular Malaysia, after data pre-whitening has been conducted, the analysis results have shown that the false alarm rate in the control chart for the PM10 indices has reduced and the existence of lag 1 autocorrelation AR(1) effect is found insignificant. Thus, it is indicated that the employment of data pre-whitening approach able to improve the control chart performance.

Item Type: Conference or Workshop Item (Paper)
Uncontrolled Keywords: control chart, PM10, monitoring, autocorrelation effect
Depositing User: Mr. John Steve
Date Deposited: 27 May 2019 07:58
Last Modified: 27 May 2019 07:58
URI: http://publications.theired.org/id/eprint/2737

Actions (login required)

View Item View Item