Improved leak localization in water distribution systems using SGTCC: Comparative analysis with EMD-HT and EEMD-HT under transient conditions
Leak detection in water pipeline systems remains a critical challenge, particularly for small-sized leaks that generate weak transient signals and are often masked by noise. This study proposes and evaluates a novel method, Squeezed Gammatone Cepstral Coefficients (SGTCC), for detecting and localizi...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier Ltd
2025
|
| Subjects: | |
| Online Access: | http://umpir.ump.edu.my/id/eprint/45053/ http://umpir.ump.edu.my/id/eprint/45053/1/Improved%20leak%20localization%20in%20water%20distribution%20systems%20using%20SGTCC.pdf |
| Summary: | Leak detection in water pipeline systems remains a critical challenge, particularly for small-sized leaks that generate weak transient signals and are often masked by noise. This study proposes and evaluates a novel method, Squeezed Gammatone Cepstral Coefficients (SGTCC), for detecting and localizing pipeline leaks under varying pressure conditions. The SGTCC approach enhances transient features through cepstral compression and gammatone filtering, improving signal-to-noise ratio (SNR) and peak resolution. Its performance is benchmarked against two conventional methods: Empirical Mode Decomposition with Hilbert Transform (EMD-HT) and Ensemble EMD with HT (EEMD-HT). Experimental tests were conducted using a controlled pipeline network at 1 bar and 2 bar pressures, each with five leak sizes (1 mm–5 mm). Results show that the SGTCC method consistently achieved the lowest mean error, ranging from 1.92 % to 2.69 % at 1 bar and 1.06 % to 2.36 % at 2 bar. In contrast, EMD-HT and EEMD-HT recorded higher error ranges, up to 7.37 % and 9.16 %, respectively. Furthermore, statistical analysis using ANOVA and t-tests confirmed that SGTCC significantly outperforms the other methods (p < 0.01). The technique also demonstrated real-time capability with acceptable computational time and robustness across all test cases. The findings highlight SGTCC’s potential as a reliable, accurate, and noise-resilient leak detection technique, particularly effective for identifying small leaks in noisy environments. These results establish a foundation for deploying SGTCC in innovative water monitoring systems and future field-scale applications. |
|---|