A Second-Order Cone Programming Approach to Nonlinear Model Predictive Control of Optimal Active and Reactive Power Dispatch for Microgrid System

Abstract

This paper aims to enhance Energy Management System (EMS) in a microgrid by integrating Nonlinear Model Predictive Control (NMPC) with Optimal Active and Reactive Power Dispatch in accordance with Optimal Power Flow techniques. In power system architecture, we consider the setting of tap-changing transformers and reactive power compensators as additional decision variables in the microgrid. This approach enables efficient and reliable system operation by minimizing Total Power Loss in transmission lines while considering Total Operating Cost (TOC) and Total Carbon Dioxide Emissions (TCOE). We then formulate the multi-objective function by combining these three objective functions and name the new function Combined Economic, Emission, and Power Loss. The proposed work leverages NMPC to optimize power dispatch from various energy sources within the microgrid on a time-varying basis. The NMPC problem can be cast as a Second-Order Cone Program. This proposed formulation proves that reactive power dispatch can be applied successfully with real power dispatch in the NMPC for better energy management in a microgrid. It reduces power loss in transmission lines and improves overall energy efficiency. Additionally, because of lower power loss, it leads to a decrease in both TOC and TCOE compared to the previous EMS without reactive power dispatch. Moreover, the proposed approach transforms the nonconvex problem into the convex SOCP which can be solved with an efficient solver, such as the interior-point method, and global optimality can be guaranteed. To prove the theoretical result, the proposed EMS is validated through simulations on the Mae Hong Son microgrid.