Localization Of Multiple Leaks In Pipelines Using Decision Trees And Support Vector Machines

Pipeline transport is widely used in industrial production and daily life. To reduce the waste of resources and economic losses caused by pipeline leaks, leak detection, localization and estimation systems are implemented in liquid pipelines. To minimize leak interpretation errors, leak detection algorithms based on artificial intelligence (AI) and data analysis have been developed. This study proposes a scheme for the detection and localization of multiple sequential leaks, based on the combination of two techniques such as decision trees and support vector machines. The results show that the proposed models have high accuracy, precision, recall and F1 score of 99.9%, 99.7%, respectively, which are better than the traditional classification model.