Development of a Global Classification Index for Drinking Water Quality Monitoring

This study provides an automated classification tool for assessing the quality of drinking water distributed through supply networks. The methodology is based on the construction of a global quality index, using the multicriteria analytic hierarchy process method, which enables the objective weighting of 23 parameters: physico-chemical and bacteriological. The classification process, divided into five quality classes, is fully automated through a python algorithm, ensuring an evaluation that is both rapid and precise. Application to an extensive database of 1718 samples from the water service of the city of Bejaia (Algeria) revealed that 97.56% of cases fall within the good and very good quality classes, thereby confirming the effectiveness of the public distribution service. Sensitivity analysis using the Sobol method highlighted the decisive importance of certain parameters: specifically, total coliforms, manganese, calcium, and conductivity in defining the final quality. Flexible and operational, the tool allows managers to quickly identify at-risk situations and to target corrective interventions at critical indicators. Thus, this model constitutes an innovative and effective approach to strengthening monitoring and ensuring intelligent management of drinking water quality.