Abstract

Drinking water treatment systems (DWTSs) are energy intensive facilities, and are an example of the water-energynexus. Benchmarking energy efficiency is a valuable tool for improving the economic and environmental performanceof such facilities. Data envelopment analysis (DEA) is typically used to assess efficiency, allocatingflexible weights(FSW) to variables that maximise energy efficiency scores for each DWTS (DEA-FSW). It means that different condi-tions are applied to each DWTS. Moreover, the DEA-FSW approach hasfinite discriminatory power which limitscross-unit comparison of energy efficiency hindering the benchmarking of DWTSs. To overcome these limitations,our study explored the effect of estimating the energy efficiency scores of DWTSs by allocating common sets of weights(CSW) within DEA (DEA-CSW). This approach was applied empirically on a sample of 146 DWTSs. Evaluated DWTSshad poor energetic performance based on both DEA-FSW and DEA-CSW estimates (low energy efficiency scores: 0.329and 0.163, respectively). Even in the optimistic scenario, the average energy efficiency score was low (0.220), withpotential electricity savings of 78 % by DWTPs when energy efficient. Unlike DEA-FSW, DEA-CSW allowed energyefficient DWTSs to be distinguished from the 146 facilities. Significant differences in the weights allocated to electric-ity and pollutants removed from raw water were reported for both approaches, and contributed to diverging energyefficiency scores. In conclusion, this study demonstrated the relevance of using suitable methods to generate compa-rable results for water companies, allowing the energy performance of DWTSs to be objectively evaluated forbenchmarking purposes