Unlocking Spatio-Temporal Flexibility of Data Centers in Multiple Regional Peer-to-Peer Energy Transaction Markets

Recently, the growth of distributed energy resources (DERs) has aroused peer-to-peer (P2P) energy transaction. With abundant flexibility of shifting computational tasks, data centers could participate in the P2P energy transaction market. However, the revenue that one data center receives through participating in regional P2P markets may conflict with that of the whole company owning several geographical data centers. Furthermore, the market clearing would be more complex after considering the latency characteristics of data centers and privacy protection in P2P transactions. This paper proposes an integrated multiple regional P2P energy transaction (MRPET) model, unlocking both the spatial and temporal flexibility of data centers in P2P trading. The detailed models of data centers, especially considering the total delay in processing tasks, are constructed, as well as several operational constraints of prosumers and P2P energy transaction. A two-stage algorithm based on Alternating Direction Method of Multipliers (ADMM) is proposed for joint clearing of the MRPET markets. The proposed models and algorithm are examined in test systems and the results show that more renewable generation could be locally consumed with the re-dispatch of multiple data centers. Reduced costs are achieved for both the data center company and prosumers. The case study also analyzes the impact of task variation and delay cost on the schedule of data centers and market clearing.