Rule-Based Predictive Control for Battery Scheduling in Microgrids Under Power Generation and Load Uncertainties
Mojtaba Kaheni, Jiali Fu, Alessandro V. Papadopoulos
Abstract
This paper addresses the control of the state of charge (SoC) of a Battery Energy Storage System (BESS) in a microgrid, considering uncertainties in load and Renewable Energy Sources (RES) generated power estimations. To achieve this objective, we propose RubPC, a novel rule-based Model Predictive Control (MPC). We partition the feasible operation space of the microgrid into two subzones, referred to as the white and yellow zones. The yellow zone represents the boundary space between the feasible and unfeasible operation spaces. In RubPC, we initially implement MPC on a predefined optimization window to determine the optimal SoC of the BESS, aiming to keep the microgrid within the white zone. Noting that mismatches between estimated and actual load and generated power may lead to constraint violations, we introduce a rule-based controller as a supervisory control. This controller monitors the microgrid’s state, and if the microgrid enters the yellow zone, it adjusts the control to maintain the microgrid within the white zone. We validate our proposed method by simulating it using data from an electrified quarry site in Sweden. Note to Practitioners—Optimizing the charge and discharge schedule of BESSs in microgrids offers a promising avenue for substantial economic and technical benefits. However, the successful realization of these benefits hinges on accurately estimating and aligning the values of load and RES-generated power. In industries, the consequences of mismatches between these estimates and actual values can translate into unexpected costs that may outweigh the anticipated economic benefits of BESS’s optimal charge and discharge schedule. This paper underscores the critical importance of addressing this concern to ensure the viability of BESS applications in various industries. To tackle this challenge, we present RubPC, an innovative Rule-Based MPC framework. Unlike conventional approaches, RubPC is specifically designed to effectively handle discrepancies between estimated and actual values, thereby preventing potential constraint violations. Our aim is to offer practitioners a robust solution that not only brings economic benefits but also ensures their safe and reliable operation.