Hardware-in-the-Loop Simulation of Data-Driven Economic Dispatch Problem for Integrated Water–Energy Systems

Integrating smart water and energy systems introduces substantial complexities, prompting the development of more advanced algorithms and techniques for their economic and efficient operation. However, applying these algorithms in real-world scenarios, particularly for (near-) real-time economic dispatch problems, presents notable computational challenges. This underscores the need for validated solutions that minimize computational complexity while optimizing power scheduling for water distribution systems. To address this, the present work employs machine learning (ML) models to solve integrated water–energy economic dispatch problems in the shortest possible time. A hardware-in-the-loop (HIL) water distribution system (WDS) testbed is used for data generation and model validation, while a conventional deterministic optimization algorithm is used to obtain the final training data for the ML model. The testbed’s integration of WDS hardware and software components facilitates iterative testing and validation of developed models and algorithms. The findings include reliable real-time solutions to the economic dispatch problem with reduced computational time, showcasing its potential for real-world implementation.