Professor Oliver Müller and Matthew Caron attended the kick-off of the research project "Data Center Heat Recovery with AI-Technologies (DC2HEAT)" in Frankfurt on Friday, 15.12.2023 and were able to exchange ideas, network and gain insights into current developments with around 70 participants.
The DC2HEAT project aims to create an AI lighthouse for the use of waste heat from data centers using Frankfurt am Main as an example. DC2HEAT is intended to help ensure that the dynamically growing data center infrastructure can be operated in a climate-friendly manner and that the use of fossil fuels for heat supply in Frankfurt and other regions can be avoided. As a result, CO2 emissions from the heating requirements of private households can be reduced by up to 15%.
As part of the project, a process model for intelligent digital twins for waste heat recovery from data centers is to be developed and tested for two waste heat recovery projects in Frankfurt. The intelligent digital twin developed in this way is dynamic and is constantly updated with real-time data from the subsystems so that it can use machine learning models and simulations to make predictions about future states and developments of the system and make recommendations for its optimization.
To this end, predictive and prescriptive models are developed for (1) the cooling of the data center, (2) the central heat pump including downstream network components and (3) the decentralized heat pumps of the consumers. As a result, the intelligent digital twin enables real-time monitoring and situational decision support along the life cycle of waste heat utilization projects.As part of the project, an intelligent digital twin for a project with a water-cooled data center and an intelligent digital twin for a project with an air-cooled data center will be created.