Towards the Decentralized Energy Optimization of Active Unbalanced Distribution Systems Integrated with Natural Gas Systems

This article presents the ADMM-based decentralized multi-period energy optimization of an active unbalanced power distribution system integrated with the natural gas distribution network. The second-order cone programming (SOCP) based relaxations are utilized to convexify the nonconvexities involved in the multi-energy system. The relaxation error in each subsystem problem is iteratively reduced to predefined accepted levels. The power distribution subsystem problem is modeled as a bi-level problem. Level 1 solves the linearized version, and level 2 solves the nonlinear version of the subsystem problem. The sequential bound tightening algorithm and solution recovery procedures are proposed to recover meaningful solutions to the original nonconvex subsystems problem from their relaxed counterparts. The proposed solution procedure minimizes the multi-period operation cost of the multi-energy system. The simulations are conducted on an active unbalanced IEEE-13 bus distribution network and a 6-node natural gas distribution network. The results obtained corroborate the proposed solution strategy.