Digital Twins for Data Centre Cooling Optimisation and Waste Heat Recovery

The paper explores the development of a Digital Twin (DT) for the management and optimisation of heat recovery in data centres, using HYCOOL-IT project demo site. The project features the development of a “Building Digital Twin Environment” as a “Platform as a Service”, integrating SIMBOT-based interactive simulators and Web API microservices to support processes from planning to performance evaluation. The article describes the integration of a waste heat recovery system within the data centre Z3 to heat the adjacent university building BL26. The DT architecture incorporates real-time data processing and a Software-in-the-Loop Model Predictive Control system to optimise control actions. Two DTs are under development: the data centre DT and the active heat recovery system DT. The former will optimise performance by continuously monitoring and controlling data centre conditions, cooling system settings, and power supply system. Simulations of the server room, chillers, and power systems will enable testing of management strategies for use cases like free cooling and emission subsystem operation optimisation. The latter will monitor the integration of a water-water heat pump (WW-HP) with the BL26 heating system, tracking electricity use, water temperatures, and flow rates. Control strategies will adjust setpoints, operation modes, and flow or temperature differences to improve system efficiency. Dynamic simulations could assess BL26 heating demand and WW-HP. Use cases include integrating WW-HP into the current plant, optimising mid-season operation, and implementing free cooling strategies to enhance energy efficiency.