Detecting and locating leakages in a piping system of transport and/or gas distribution by using acoustic methods and artificial neural networks

Abstract

The purpose of increasing the operational reliability of the ducts is crucial because every year there is a growing number of pipelines employed in the transport and/or distribution of gas and liquid products. The control operation of these systems is a very complex task. The diversity of the network components, often coupled to the great distances involved, impose the need for ongoing monitoring of operating conditions. Considering the importance of monitoring of piping systems, this research project aims to develop a technique for detecting and locating leaks in an experimental piping system of transmission and/or distribution of gas, based on the acoustic method, transient pressure measurements and artificial neural networks. The signals from the sensors, generated by simulated leaks of different magnitudes, will be detected and analyzed in a 50-meter- long pipe, under various pressure conditions. The experimental data acquisition will be performed by the computer using microphones and pressure transmitters installed at various points of the pipeline. The dynamics of the noise generated by the microphones, together with the signals of pressure sensors, will be used as input to the neural model to detect and locate leaks (model outputs). Once determined the leak location methodology, valves will be implemented and controlled by PID strategy for the purpose of acting on the system, diverting the flow passage of the line where the leak occurred. (AU)