Algorithm to detect growing gas leakage based on a statistical criterion

The current commercial production technologies often involve the generation and pipeline transportation of gaseous substances in utilities and premises. Airborne concentration is a quantitative gas leakage characteristic, which may increase up to threshold values causing danger to humans and the environment and creating an emergency. Herein, changing the airborne concentration of impurities is described by a time series model. The possibility of creating an algorithm to automatically search for a trend break and estimate a new trend slope as a sign of emerging gas leakage has been studied. The authors propose a new way of applying a trend break as a criterion for detecting abnormalities. This allows identifying an event that leads to growing gas leakage. The authors have tested an algorithm that allows detecting new trends in the time series and predicting rapidly approaching threshold gas impurity concentration.