Pump Efficiency Analysis of Waste Water Treatment Plants: A Data Mining Approach Using Signal Decomposition for Decision Making

In Waste Water Treatment Plants (WWTPs), the pump systems are one of the most energy intensive processes. An efficient energy management of pumps should produce environmental and economic benefits. In this paper, we propose a daily data-driven approach for a detailed pump efficiency analysis that reduces the time gap between an inefficiency and its detection, provides detailed information for decision making by using new Key Performance Indicators (KPIs), and detects inefficient pump set-ups and designs. The proposed approach based on signal decomposition relies on sensors generally available in WWTPs, e.g. daily pump inflow and energy consumption. Moreover, it allows decomposing the data signal in an automatic way into a long-term trend and short-term fluctuations. This information can then be used to support plant managers more effectively.