GAS LEAK DETECTION SYSTEM AND METHOD, METHOD FOR DETERMINING THE IMPORTANCE AND LOCATION OF A GAS LEAK BY MEANS OF NEURAL NETWORKS, AND USE IN RIGID AND/OR FLEXIBLE PIPES

摘要

The present invention relates to a gas leak detection system and method, to a method for determining the importance and location of a gas leak by means of neural networks, and the use thereof in rigid and/or flexible pipes. The method for determining the importance and location of a gas leak involves the use of neural networks for data analysis, which offers the advantage of allowing large quantities of data to be efficiently handled, and of having excellent generalisation capabilities. Moreover, the use of neural modelling dispenses with human monitoring, by means of a computer that analyses graphs generated by the acoustic signals, since through neural modelling, warnings that give information about the occurrence, the importance and the location of a leak are generated. Fast detection and location of leaks by means of an efficient monitoring system are essential to reduce and prevent large-scale accidents.