Dynamic optimisation of unbalanced distribution network management by model predictive control with Markov reward processes

• DOI: 10.1016/j.heliyon.2024.e24760
• Publication Date: 2024-01-17T22:04:01Z
• Authors: César Álvarez-Arroyo; Salvatore Vergine; Guglielmo D'Amico; Juan Manuel Escaño; Lázaro Alvarado-B...
• Keywords: Model Predictive Control; Time horizon

In this work, a two-level control system is used to minimize the total active power losses of an distribution connected external grid and composed wind turbine, two photovoltaic sources, batteries. At first level, model-based predictive (MPC) run, using non-homogeneous Markov reward models for prediction homogeneous power. second algorithm run optimal management voltage assets, such as regulating transformers, losses. Different scenarios have been considered, highlighting advantages MPC framework. This results in optimization process that can be influenced by different time horizons depending on whether or not applied. The predictions allow considering long-horizon stepwise leads increasing number variables along with decrease When applied, short-horizon analysis performed both quality results. cases are considered which nominal unit battery capacity modified.