A Second-Order Dual Update Approach for the Decentralized Optimal Scheduling of Polygenerative Microgrids

This article develops a decentralized energy management system (EMS) to optimally allocate electric and thermal power production in a polygeneration microgrid (PMG). The EMS problem is formulated as a multiperiod convex optimization problem solved using a model predictive control (MPC) framework. The MPC optimization problem is solved using the augmented Lagrangian with second-order dual update (AL-SODU) algorithm, a new decentralized algorithm based on the augmented Lagrangian method with SODUs. The proposed AL-SODU algorithm significantly outperforms the state-of-the-art algorithms employing first-order dual updates and their accelerated version. The developed EMS based on AL-SODU is tested on the real case study smart PMG located on the Savona Campus of the University of Genoa (Italy).