Optimal standalone data center renewable power supply using an offline optimization approach

Because of the increasing energy consumption of data centers and their CO2 emissions, the ANR DATAZERO2 project aims to design autonomous data centers running solely on local renewable energy coupled with storage devices to overcome the intermittency issue. In order to optimize the use of renewable energy and storage devices, a MILP solver is usually in charge of assigning the power to be supplied to the data center. However, in order to reduce the computation time and make the approach scalable, it would be more appropriate to use a polynomial time algorithm. This paper aims at showing and proving that it is possible to provide an optimal power profile via a deterministic algorithm using a binary search approach. Considering the main constraints of the initial problem, numerous experimental results show similar results to those given by the MILP. These promising results encourage us to continue in this direction for proposing an efficient management of the data center power supply that takes uncertainty into account.