Author(s): Edwin Gamboa; Mathias Giessler; Paul Engelke
Linked Author(s):
Keywords: Aeration; Predictive maintenance; Multi-label classification; LightGBM; Time series; Wastewater treatment
Abstract: Aeration represents the most energy-intensive process in wastewater treatment plants (WWTPs), accounting for 30–70% of total electricity consumption. Fouling, scaling, and aging of diffusers and blowers reduce oxygen transfer efficiency, increasing operational costs [1]. Predictive Maintenance (PdM) offers a data-driven approach to optimize maintenance schedules, reduce downtime, and extend equipment lifespan [2]. Although PdM has been widely studied in WWTPs [3,4], its application to aeration systems, especially for multi-parameter, full-scale monitoring, remains limited.
Year: 2026