PERLE

Project description

Building on the previous work in the OASEE project, in which the indoor climate in an aquaculture hall was improved using an energy-optimized model predictive controller (MPC), a further control concept based on deep reinforcement learning is to be developed. This state-of-the-art approach is intended to provide an alternative to the complex development of the implemented MPC. Both control concepts are to be tested and evaluated in terms of performance, development effort and transferability to other buildings. This project is being carried out in collaboration with the company Fieles Dithmarscher Kältetechnik.