Model Predictive Energy Management for Building Microgrids with IoT-based Controllable Loads

Abstract

This paper develops an economic scheduling framework for a building microgrid with internet of things (IoT) based flexible loads to synchronize the buildings' controllable components, with occupant behavior and environmental conditions. We employ model predictive control (MPC) methods to minimize building operating costs, while maximizing the utilization of the on-site resources. The main research thrusts are 1) Developing the building microgrid model; 2) Defining different building operation strategies; 3) Minimizing the building's daily operating costs. Simulation results show that the proposed approach provides superior energy cost savings and peak load reduction in comparison with other operation controls, such as offline Mixed Integer Linear Programming (MILP), All from Utility (AFU), and MPC-MILP with non-controllable loads.