Data Center Energy Optimization Under Different Dynamic Pricing Schemes

The increasing power demands in data centers (DCs), driven by an expanding variety of digital applications and a growing number of Internet users, necessitate alternative strategies to curtail DC energy consumption expenses. In the realm of the DC facility, opportunities lie in the availability of renewable energy sources, uninterruptible power supply battery systems, and the varying nature of IT workloads. In this paper, a mixed-integer linear programming energy management system is proposed to exploit on-site renewables, existing energy storage, and IT workload flexibility to optimize electricity expenditure for a hypothetical small-scale DC in Istanbul, Turkey. The proposed EMS model is evaluated under four pricing schemes: real-time pricing (RTP), time-of-use (TOU), critical peak pricing (CPP), and flat rate pricing (FRP). An annual cost savings analysis demonstrates considerable cost reduction advantages by the existing TOU pricing in Turkey (15.7%) and artificially generated tariffs of CPP and RTP (23.1% and 22.7%, respectively) to DC operators in comparison to the FRP.