Optimal Energy Management in Microgrids with Integrated Multi-zone Heating/Cooling Control

Abstract

There is a growing need for technologies that can help integrate distributed renewable energy and storage capacity into the electricity grid and increase the efficiency of its operation. Microgrids provide an effective framework for achieving these objectives. The energy consumed in buildings accounts for a significant part of the total energy demand. Space conditioning, which includes heating, ventilation and air conditioning (HVAC), is the largest contributor to the building energy consumption. Therefore, combined management of heat and power in microgrids can potentially yield substantial energy and cost savings. This thesis presents an integrated approach to energy/power management and control of a HVAC system in a grid-connected microgrid with energy storage. The control strategy is based on an on-line optimal model predictive control philosophy. An optimization problem is formulated and solved over a rolling horizon using a model of the system and predicted microgrid load and outside temperature to obtain optimal control decisions, i.e. the energy storage charge/discharge power and HVAC control commands. The overall problem is formulated as a constrained optimization in the form of a mixed integer linear program (MILP), with the objective of minimizing energy cost subject to model and operational constraints. First, a single temperature zone configuration is presented where only one temperature variable is controlled throughout the building. The controller maintains the building temperature within user-defined upper and lower limits, which may also be determined based on occupancy. By employing auxiliary on-off controllable fans in the temperature zones, a multi-zone control strategy is also proposed to independently regulate the zone temperatures within their corresponding comfort limits. Several strategies for reducing real-time computations of the controller are also proposed. Simulations have been carried out under different scenarios including single-zone and multi-zone cases that demonstrate significant efficiency gains from the application of the proposed controller for energy management and HVAC control in a microgrid system. Sensitivity of the system performance with respect to modeling errors and uncertainty is examined in a number of simulations. The results demonstrate a robust behavior due to an inherent feedback mechanism in the proposed rolling horizon model predictive controller.