A comparison of various residual-based control charts for detecting and monitoring abnormal river water levels
Malaysia experiences recurring floods with numerous issues that have surfaced annually. The rapid overflow of river water necessitates monitoring abnormal levels with a visualization tool as an alarm system. This study utilized a control chart approach to monitor the Klang River's water level i...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Penerbit Universiti Kebangsaan Malaysia
2025
|
| Online Access: | http://journalarticle.ukm.my/25781/ http://journalarticle.ukm.my/25781/1/253-265%20-.pdf |
| Summary: | Malaysia experiences recurring floods with numerous issues that have surfaced annually. The rapid overflow of river water necessitates monitoring abnormal levels with a visualization tool as an alarm system. This study utilized a control chart approach to monitor the Klang River's water level in Taman Sri Muda for a signal of potential flood occurrence. Aiming to identify a suitable method, this study compares several residual-based control charts to achieve the purpose. Due to the autocorrelation issue that might cause false alarms, this study emphasizes the implementation of control charts based on residuals. The control chats were developed based on the residual observations obtained from the best time series model, ARIMA (1,1,1), involving the Mean Shewhart (I-Shewhart), Moving Average (MA), and Exponential Weighted Moving Average (EWMA). The results show that EWMA Residual with λ=0.05 is found to have the best performance compared to other control charts; followed by EWMA Residual with λ=0.1 and the I-Shewhart Residual with a detection score of 7.1, 3.84, and 3.56 percent respectively. The results revealed that the detected water level anomalies occurred dominantly in April, October, and November. The study emphasizes the efficiency of early signals produced by the control charts in detecting abnormal water levels by visualizing their occurrence and behaviour in the Klang River at the study location. |
|---|