Validation Technique for Pipeline Leakage Localization (VTeLL)

Pipelines are essential parts of the infrastructure of modern society; Leak detection systems (LDS) are used to protect the public and the environment from the effects of pipeline failure as well as bringing down operational costs, LDS do this by alerting when a leak occur and solving the problems of leakages most often involves the application of two or more methods which largely depends on several factors. Identification and isolation are important steps in minimizing the consequences of any leak; LDS are used in order to do both. In most cases, it is impossible to pinpoint the exact location of a leak because no single method has become an industry standard due to the limitations of each. We present an alternative validation technique for pipeline leakage prediction through behavioral modeling of the data pattern associated with the various LDS predictions. Our tested data resulted in less than 0.3% localization error.