One of the most important areas of the water utilities is the water quality management. This area is responsible of guaranteeing safety in the water supply to the citizens. The strategy to guarantee the safety is based on two principal elements: disinfection and monitoring. Disinfection techniques, such as chlorination, allow to prevent the growing of microorganisms present in the water. Moreover, in order to guarantee this safety in the whole water network and avoid any unexpected event, on-line sensors are required to monitor a set of quality parameters. The whole process is based on the assumption that the information retrieved from quality sensors is totally reliable. But due to the complexity of the calibration and maintenance of these chemical sensors, several factors affect the accuracy of the raw data collected. Consequently, any decision based on this raw data might be based on a non solid base. Therefore, this work presents a data analytics approach consisting in two modules: fault diagnosis and prognosis. The fault diagnosis module first discerns if a sensor is detecting a real change on water quality parameters or actually is providing inconsistent information due to some malfunction. The prognosis module aims to predict the fault instant due to a slow degradation, which is very common in chlorine sensors. This approach allows to apply a predictive maintenance strategy reducing corrective actions. The proposed methodology has been satisfactorily tested on the Barcelona Drinking Water Network.

Peer Reviewed