Studies on Leak Detection in Process Pipelines Using Artificial Neural Networks/Machine Learning

Abstract

In process industries, leakage in pipelines is common and an environmental, health and economic issue to be addressed without fail. To detect the presence of leaks in pipelines, there are several conventional leak detection techniques that are available. From the literature, conventional leak detection techniques can be mainly classified into: hardware-based, visual and software-based methods. Researchers have focused more on software-based methods due to simple and reliable operation. Under software-based methods, using machine learning algorithms that is a data driven approach for leak detection and localization is becoming popular because of learning capabilities. Therefore, in this review, several recent conventional leak detection methods and artificial neural networks/machine learning based leak detection methods are described. It is an attempt made to identify leak detection techniques along with the common methodology followed for building a leak detection system using artificial neural networks/machine learning.