A Comparative Framework for Environmental Performance Assessment of Wastewater Treatment Processes Using Multicriteria and Statistical Methods

Wastewater treatment plants are central environmental processes because their performance directly controls pollutant removal, effluent quality and the ecological impact of discharged water. However, many data-driven studies focus on prediction or fault detection, while fewer provide transparent daily indicators for environmental performance assessment. This study develops and compares four index-based methods for monitoring wastewater treatment performance using open operational sensor data and Python. The public UCI Water Treatment Plant dataset was used, comprising 527 daily records collected between January 1990 and October 1991. The dataset includes variables related to influent conditions, effluent quality and removal efficiencies. Sixteen normalised criteria were derived and grouped into three dimensions: Effluent Quality Index, Removal Efficiency Index and Process Stability Index. Four methods were implemented and compared: a simple weighted composite index, a hierarchical Analytic Hierarchy Process index, a Principal Component Analysis index and an AHP-TOPSIS index. The hierarchical AHP method ranked best because it combined transparent environmental priorities with coherent pairwise matrices, with consistency ratios equal to 0. It produced 520 valid scores, a mean of 72.04/100 and a median of 73.93/100. AHP-TOPSIS ranked second, with 527 valid scores, a mean of 63.56/100 and a strong correlation with AHP (r = 0.956), while providing a stricter distance-to-ideal assessment. The simple weighted index ranked third, with a mean of 73.20/100 but fewer low-performance days detected. PCA ranked fourth: although correlated with AHP (r = 0.909), PC1 explained only 30.39% of total variance and class agreement with AHP was 45.38%. The framework demonstrates that open sensor data can support low-cost, transparent and reproducible monitoring for Environmental and Green Processes.