Author:

Keywords:

Demand response, Heat pump, Residential buildings, Curtailment, Renewable energy sources, Optimal control problem

Abstract:

Renewable energy sources (RES) will play a vital role in reducing the impact of climate change. Solar and wind energy, captured by PV-panels and wind turbines respectively, are two major RES that pose enormous potential but have two important disadvantages: a limited predictability and intermittency. Demand response (DR) is often put forward as part of the solution for intermittency, by shifting electricity demand away or towards times of shortages or abundances of RES respectively. One of the major technologies that pose significant potential for DR are electrical heating and cooling systems. Within these systems, this work focuses on the heating of residential buildings by means of a heat pump. Residential buildings with heat pumps show potential for DR as the building structure and domestic hot water tank can be used as thermal energy storage. This allows a decoupling in time of the delivery of thermal comfort and the heat pump electricity demand. One of the factors hampering a widespread implementation of DR for heat pumps is a thorough understanding of the potential benefits. To this aim, this work presents an integrated modeling approach that captures both the incentives for DR by explicitly modeling the electricity generation system as well as the flexibility potential of residential buildings with heat pumps. In contrast to the literature, a bottom-up representation of this flexibility potential is developed, resulting in a linear optimal control problem (OCP). This linear OCP of buildings with heat pumps is combined with a state of the art unit commitment and economic dispatch model of the electricity generation system. The added value of this integrated modeling approach is shown with respect to typical other approaches in the literature. It correctly captures the maximal potential for DR, weighs the thermal losses against the supply side incentives and includes the feedback with the electricity generation system. This integrated model is employed in a number of case studies to explore the potential benefits of DR for residential heat pumps. In a first case study for a Belgian context, it is shown that DR with residential heat pumps can shift electricity demand towards moments of curtailment in order to avoid electricity demand later on when fuel fired power plants are running. In this case study, half of the shifted electricity demand was directly lost due to thermal losses in the residential buildings. In this manner, applying DR to residential heat pumps allows reducing the CO2 emission further by an extra 15 \% on top of the CO2 emission reduction of installing a heat pump. Furthermore, the contribution of the electricity demand of residential heat pumps to the peak electricity demand is almost one on one in case no DR is applied. With DR, well insulated buildings can significantly shift their electricity demand away from the peak. Buildings which are not well insulated are shown to be unattractive for installing heat pumps, even with a DR implementation. In a second case study, the monetary benefits of applying DR for residential heat pumps are identified to go up to 150 EUR per participant per year from operational savings and up to 300 EUR per participant per year by avoiding peak electricity demand. These cost savings are shown to diminish as the participation in DR rises. This discourages extreme configurations in the residential buildings: high spreads on temperature set points or high domestic hot water tank sizes pose little added value in case of higher DR participation. The integrated model has potential for a practical implementation of DR with residential heat pumps. It can be employed to anticipate, in a day ahead setting, the reaction of residential heat pumps to incentives from the electricity generation system. Residential heat pumps controlled by MPC are shown to be very greedy to these incentives. In this manner, sending an electricity price profile leads to poor performance when a high number of buildings, 100,000 in this work, join in on DR. Sending an electricity demand profile for the residential heat pumps to follow attains superior performance in this context. Finally, this work combines the integrated model with a heating system design optimization in order to investigate whether heating systems should be designed differently in the light of high RES shares in the electricity generation system. Two modeling additions towards this aim, temperature level and electricity generation side modeling, show limited added value to the heating system optimization but does allow a correct quantification of the CO2 emissions. From the results it follows that a residential building is best equipped with a single main heat production system. The combination of heat pump and PV panel reduce the CO2 emission of the building with up to 2 ton per building per year compared to typical fuel fired options for a limited extra cost. A storage tank for space heating and solar thermal panels appear to be unattractive technologies. A large scale combined heat and power unit is attractive but the CO2 emission strongly depends on the CO2 emission calculation method.