Abstract:Pipeline monitoring and leak detection are critical for safe and economical operation as well as preventative maintenance in water and oil industry. They also provide environmental protection from crude oil emission or theft. Leak detection and localization play a key role in the overall integrity monitoring of a pipeline system especially for long residential water pipelines and midstream crude pipelines. Operational data such as fluid flow and pressure data are often utilized in many physics and computation-based leak detection and localization methods. However, noise in both flow and pressure affects recognition of actual leak pressure drop and results in false leak alarm and inaccurate leakage localization. In this paper a total variation regularized numerical differentiation (TVRegDiff) algorithm is used to estimate derivatives from noisy fluid flow and pressure data. Then possible leak events from both fluid flow data and pressure data are obtained by examining these derivatives against corresponding control limit lines. Customized parameter criteria of flow and operation activities are used to effectively filter flow and operation events from possible leak events. Therefore, earliest leak occurrence time and pressure meters can be identified. Finally, negative pressure wave (NPW) analysis is combined with the leak recognition to compute and locate leakage. Negative pressure wave methods are based on the principle that a leak will cause a sudden pressure alteration as well as a decrease in fluid flow speed which will result in an instantaneous pressure drop and speed variation along the pipeline. As the instantaneous pressure drop occurs, it generates a negative pressure wave starting at the leakage position and propagates with certain speeds towards both the upstream and downstream ends of the pipeline section. Depending on the leakage position from the two ends and actual negative pressure wave velocities towards upstream and downstream there are time sequence of the pressure wave arriving at both ends. The lapse of time arrival and negative pressure wave velocities are then used to calculate the actual leakage position within the pipeline section. A field leak case study indicates that actual leak pressure drop is correctly captured and pipeline leakage can be accurately located. The accuracy of leak localization is theoretically dependent on the pressure and flow meter frequency. With widely accepted industry standards for pressure devices used in industry applications it is proved to be a practical, efficient and accurate leak detection and localization method. This method can also be applied to not only residential water pipelines but crude oil pipelines.

Application of support vector machine learning to leak detection and location in pipelines

In-Situ Leak Detection in Urban Water Supply Pipelines Using Flexible Vibration Sensors

Leak Detection in Water Distribution Systems by Classifying Vibration Signals

An Angle-Based Leak Detection Method Using Pressure Sensors in Water Distribution Networks

Leakage Localization Using Pressure Sensors and Spatial Clustering in Water Distribution Systems

A BiLSTM Based Pipeline Leak Detection and Disturbance Assisted Localization Method

Negative pressure wave based pipeline Leak Detection: Challenges and algorithms

Pipeline Leak Detection and Localization Based on Advanced Signal Processing and Negative Pressure Wave Analysis

Oil Pipeline Leak Detection and Location Based on Image Processing

Leak location of pipelines based on transient model and PSO-SVM

A SVM-based pipeline leakage detection and pre-warning system

Leakage detection of low-pressure gas distribution pipeline system based on linear fitting and extreme learning machine

Leak Location of Pipelines Based on Characteristic Entropy

An Intelligent Water Supply Pipeline Leakage Detection Method Based on SV-WTBSVM

Application of wavelet transform to leak detection and location in transport pipelines

PIPELINE DEFECT PREDICTION USING SUPPORT VECTOR MACHINES

Application of time-frequency image feature extraction method based on gray level co-occurrence matrix in coal mine pipeline leakage detection

An improved variational mode decomposition method based on particle swarm optimization for leak detection of liquid pipelines

Hierarchical Leak Detection and Localization Method in Natural Gas Pipeline Monitoring Sensor Networks

A New Leak Detection Method for Subsea Pipelines

An SAE-based resampling SVM ensemble learning paradigm for pipeline leakage detection