Understanding energy performance in drinking water treatment plants using the efficiency analysis tree approach

Abstract

Water treatment processes are known to consume substantial amounts of energy, making it crucial to understand their efficiency, drivers, and potential energy savings. In this study, we apply Efficiency Analysis Tree (EAT), which combines machine learning and linear programming techniques to assess the energy performance of 146 Chilean drinking water treatment plants (DWTPs) for 2020. Additionally, we utilize bootstrap regression techniques to examine the influence of operating characteristics on energy efficiency. The results indicate that the evaluated DWTPs exhibited poor energy performance, with an average energy efficiency score of 0.197. The estimated potential energy savings were found to be 0.005 kWh/m3. Several factors, such as the age of the facility, source of raw water, and treatment technology, were identified as significant drivers of energy efficiency in DWTPs. The insights gained from our study can be valuable for policymakers in making informed decisions regarding the adoption of practices that promote efficient and sustainable energy use within the water cycle