Practical strategies for managing resistance heating in heat pump water heater predictive control

Abstract

As the U.S. grid transitions to 100% carbon-free electricity, adopting electric heat pump water heaters (HPWHs) is key for decarbonizing homes and reducing operating costs for end users. However, their widespread adoption could strain the electric grid. Load-shifting control strategies for HPWHs are needed to shift demand from peak hours to periods with low-cost renewable energy, while ensuring occupant comfort. Economic model predictive control (MPC) can optimize HPWH operation by accounting for physical constraints, tank thermal dynamics, and time-varying factors like electricity prices and hot water demand. A key aspect of the MPC for HPWHs is the use of a thermal energy storage tank model as its prediction model. While various techniques exist for modeling tank thermal stratification, they typically have nonlinear dynamics. Conversely, studies have shown improved performance of HPWHs under MPC with a simplified, low-order tank thermal model compared to the performance under conventional control strategies. This study investigates the adverse effects of enabling resistance heating in MPC with a low-order tank thermal model for heat pump water heaters with two resistance elements, including overheating and unnecessary tank heating. To address these issues, practical strategies, in the form of logic-based constraints, are incorporated into the MPC formulation to manage resistance heating activation and the selection between the two resistance elements. Extensive simulation results are presented to examine the effectiveness of the logic-based constraints in mitigating overheating and unnecessary tank heating in HPWHs under the MPC with a low-order tank thermal model. Additionally, the closed-loop results under the MPC are compared against those under a typical HPWH rule-based control strategy to assess its ability to minimize electricity costs while maintaining occupant comfort.