Prediction of Faulty Residential Gas Meters Using an Artificial Intelligence Method

One of the most important daily challenges of the gas distributing companies is the accurate screening of the faulty gas meters. Failing of which has caused serious financial shortcomings for these companies. In the present study, an intelligent method based on Support Vector Machine concept is suggested to simplify the aforementioned screening target. More than 140 number of gas meters in Shiraz are tested based on this novel idea, and the primary results are promising. It is expected that the suggested approach can be extended to address large size data

Keywords:

Gas meter screening , Artificial intelligence , Clustering , SVM , Fault detection