A framework combining acoustic features extraction method and random forest algorithm for gas pipeline leak detection and classification

Acoustic feature-based leakage event detection in near real-time for large-scale water distribution networks

“…To mitigate the requirement of large acoustic datasets for effective model training, the need to identify useful acoustic data features for leak detection has since garnered some research interest. For example, Kothandaraman et al (2020) coupled conventional cross-correlation analysis with EMD to detect and localize pipe leakages underground, and Ning et al (2021) recently developed a useful framework that couples EEMD and random forest algorithm to classify the different types of leaks. While these studies have quantitatively demonstrated the usefulness of EMD or EEMD pre-processing methods to remove ambient noises embedded in the acoustic signals, we again highlight that they were mainly collected under controlled lab-scale experiments where the frequency of the environmental noises (e.g., blowing fans, pump noise) can be identified easily and removed/filtered to extract the most useful leakage acoustic frequency.…”

Condition monitoring and temporal‐spatial assessment of composite pipeline transporting potable water

“…RF is characterized by the capability of handling high‐dimensional problems with good accuracy and reliability, a few parameters to be adjusted, insensitivity to noise due to its special sampling and decision‐making mechanism, and the exemption of over‐fitting, leading to its suitability for online condition assessment. RF‐based approaches are increasingly popular for structural condition assessment 42–50 . Some of them focus on deterioration modeling and forecasting of sewer pipelines, 44–46 and in these investigations the input features are the suggested physical sewer characteristics (e.g., age, material, type of effluent, depth, diameter, slope, and type of road) and the output is deterioration level (e.g., four‐grade scale: C1–C4) obtained by using CCTV or GIS records.…”

“…Bagriacik et al examine the performance of RF and other three approaches for modeling earthquake damage to water pipelines and identify the important features for earthquake damage detection of water pipelines (e.g., pipe length, peak ground velocity [PGV], liquefaction resistance index [LRI], and pipe material) 48 . Ning et al combine acoustic signals and RF method for leak detection of gas pipeline, and prove its effectiveness through lab tests of an experimental system 49 . Shi et al integrate a RF‐recursive feature elimination and a super‐learning algorithm to characterize the relationships between pipe deformation and environmental factors 50 .…”

“…RF-based approaches are increasingly popular for structural condition assessment. [42][43][44][45][46][47][48][49][50] Some of them focus on deterioration modeling and forecasting of sewer pipelines, [44][45][46] and in these investigations the input features are the suggested physical sewer characteristics (e.g., age, material, type of effluent, depth, diameter, slope, and type of road) and the output is deterioration level (e.g., four-grade scale: C1-C4) obtained by using CCTV or GIS records. The results show the RF-based method can identify the critically deteriorating pipelines among all pipelines in a designated area and classify the pipeline conditions into different levels at a high accuracy.…”

“…48 Ning et al combine acoustic signals and RF method for leak detection of gas pipeline, and prove its effectiveness through lab tests of an experimental system. 49 Shi et al integrate a RF-recursive feature elimination and a super-learning algorithm to characterize the relationships between pipe deformation and environmental factors. 50 The performance of pipe deformation prediction is examined by using environmental data of soil and responses of a polyvinyl chloride pipe with 150-mm diameter in a field instrumentation program.…”