(587e) Cloud-Based Economic Model Predictive Control for Residential Heat Pump Water Heaters

Water heating is the second-largest energy end-use in U.S. households and accounts for 19% of total energy consumption [1]. Aligning with California’s goal of achieving a 100% carbon-free electricity grid by 2045, the state has proposed policies that promote renewable energy, energy efficiency, and energy storage technologies [2]. These policies incentivize energy-efficient heat pump use for domestic water heating [3]. In addition to energy efficiency, load flexibility is vital for achieving aggressive grid goals. Residential customers are incentivized to provide load flexibility through time-of-use tariffs, which feature a high-cost or peak period and a low-cost or off-peak period [4]. In this sense, load flexibility can lead to cost savings for the customer while reducing the grid emissions rate.

Heat pump water heaters (HPWHs) are water heaters with a hot water tank where the water is primarily heated through a heat pump. HPWHs usually have one or two electric resistive heating elements to heat water during high demand periods. Control approaches that provide HPWHs load flexibility, i.e., load shifting capability, shift electricity consumption away from peak periods to periods when renewable energy sources are abundant. HPWH load flexibility is enabled by the built-in tank, which can serve as thermal energy storage. HPWHs usually employ rule-based control (RBC), which uses a set of predefined rules to maintain the water temperature at a setpoint. While rules can be developed to provide some form of load flexibility [5], a drawback of RBC for providing load flexibility is that the trigger parameters of these rules are fixed. This may lead to suboptimal operation that gets worse over time as tariff pricing structures and demand patterns change [6].

Economic model predictive control (EMPC) is an advanced control method that integrates control and economic optimization in a predictive control setting [7]. EMPC is an ideal choice for providing load flexibility to HPWHs because it can account for time-varying costs, grid carbon emissions rate, availability of energy storage, and hot water demand to make optimal operating decisions. A few studies have demonstrated the benefits of using EMPC for direct control of HPWHs (e.g. [8], [9]). A few aspects need to be addressed to make EMPC for HPWHs a reality. First, the EMPC needs a model to forecast occupant hot water demand to make optimal operating decisions. However, a flow meter measuring hot water flow rate is rarely available on residential HPWHs. Second, limited work has been undertaken to implement supervisory EMPC for HPWHs, even though this implementation could be readily deployed to existing HPWHs.

In this work, a cloud-based supervisory EMPC for residential HPWHs is designed to minimize electricity cost and marginal GHG emissions rate. In this control architecture, the EMPC sends temperature setpoints to an existing RBC that decides how to operate the HPWH heating devices so that the tank water temperature reaches the setpoint. The multi-objective EMPC formulation considering electricity cost and predicted grid emissions rate is presented. The advantage of the cloud-based implementation over an on-premise implementation is that EMPC can be applied to any HPWHs that are cloud-connected with no modifications needed to the existing RBC. Compared to EMPC which directly controls the HPWHs, the supervisory implementation minimizes operational disruptions in the event of communication failures because the HPWH can continue to function properly with the RBC. A model-based estimation method for estimating hot water demand and a data-driven hot water demand model for forecasting future demand is developed. Field data collected from a recent study is used to validate the hot water estimation and prediction approach. The data-driven predictive hot water model is used to provide input data to the EMPC. To evaluate the performance of the resulting control structure, a detailed first-principles model of HPWH is developed. Laboratory data are used to validate the HPWH model. The proposed approach is compared to an RBC and EMPC which controls the HPWH directly to demonstrate the effectiveness of the proposed supervisory EMPC approach.

References

[1] C. Berry. Space heating and water heating account for nearly two thirds of u.s. home energy use. Technical report, U.S. Energy Information Administration, 2018.

[2] SB 100 joint agency report. Technical report, California Energy Commission, 2021.

[3] California Public Utilities Commission. Fact sheet: Heat pump water heater incentive programs, May 2020.

[4] PG&E. Residential rate plan pricing, Mar 2022.

[5] A. Kathirgamanathan, M. De Rosa, E. Mangina, and D. P. Finn. Data-driven predictive control for unlocking building energy flexibility: A review. Renewable and Sustainable Energy Reviews, 135:110120, 2021.

[6] T. Q. Péan, J. Salom, and R. Costa-Castelló. Review of control strategies for improving the energy flexibility provided by heat pump systems in buildings. Journal of Process Control, 74, 2018.

[7] M. Ellis, H. Durand, and P. D. Christofides. A tutorial review of economic model predictive control methods. Journal of Process Control, 24:1156–1178, 2014.

[8] M. D. Knudsen and S. Petersen. Model predictive control for demand response of domestic hot water preparation in ultra-low temperature district heating systems. Energy and Buildings, 146:55–64, 2017.

[9] E. M. Wanjiru, S. M. Sichilalu, and X. Xia. Model predictive control of heat pump water heater - instantaneous shower powered with integrated renewable-grid energy systems. Applied Energy. 204:1333-1346, 2017.