Distribution Network Congestion Management with Internet Data Centers Using Bi-level Programming

Large-scale demand-side resources (DSRs) like geo-distributed internet data centers (IDCs) access to the distribution network may cause security problems, e.g. line congestion. The demand response (DR) method is wildly used according to price signals to motivate users to schedule the loads. This paper proposes a bi-level programming optimization framework for distribution network congestion management considering power network and workload allocation. The upper level is the workload scheduling strategy of an internet service company (ISC) with multiple IDCs to minimize the total electric power cost. The lower level is the strategy of distributional locational marginal pricing (DLMP) based on direct current optimal power flow (DCOPF) to minimize the total generating cost. Three scenarios with a modified IEEE-30 bus system are studied and analyzed to prove the effectiveness of the proposed strategy both in cost reduction and congestion management.