Energy Efficiency Optimization of a Wastewater Pumping Station Through IoT and AI: A Real-World Application of Digital Twins

One of the challenges associated with the green transition involves advances in digitization and the integration of intelligent demand-response services to improve energy effi-ciency and unlock on-the-fly adaptability of electrical consumers. This paper presents a cloud-based optimization framework that relies on synergies between the Internet of Things (loT) and Digital Twin (DT) technology to provide energy-saving services in the wastewater pump station located in Bornholm, Denmark. The real-time collected data from the pump station establish the backbone of the DT, enabling high-fidelity simulations of system behaviors under different conditions and assessing the effectiveness of various energy -saving strategies. Moreover, the proposed methodology incorporates a Model Predictive Control (MPC) combined with the inflow forecasts to unlock energy optimization potential. Finally, the implementation of a pruned Neural Network Imitator (NNI) executed on an IoT gateway shows significant improvement in operating the pumps in the most efficient zone by emulating the functionality of MPC, tailoring control strategies for real-time efficiency improvement.