Experimental Assessment of an Air-to-Water Heat Pump Driven by a Demand Response Strategy

A relevant part of the global electricity demand is used for the HVAC (Heating, Ventilation, and Air-Conditioning) sector, demand that is expected to rise up to 4000 TWh in 2050 [1]. Space heating and cooling of buildings are responsible for about 30–45% of the energy demand [2], and in the European residential sector, where the main consumption of energy includes hot water production and space heating, heat production is largely based on non-renewable sources, which are responsible of large emissions of greenhouse gases [3]. Since this practice is not sustainable, the European Performance of Building Directive (EPBD) [4] required a nearly-zero energy building (nZEB) standard for new buildings by 2021. This standard has led to a widespread diffusion of high-efficient energy systems in the household market, in particular heat pumps. The use of heat pumps is growing rapidly: for example, with a 12% increase reached in 2018, the European heat pump market has achieved double-digit growth for the fourth year in a row, and a doubling of the European heat pump market by 2024 is expected [5].

Heat pumps, as well as other energy systems such as refrigerators, air-conditioners, etc., can be considered as thermostatically controlled loads in buildings. On the other hand, the energy flexibility provided by buildings is paramount to mitigate the upcoming challenges of future power systems, as outlined by the working group of Annex 67 about energy flexible buildings [6]. For this reason, if managed together in a smart way, heat pumps and buildings can be used to modify the electric load without affecting the quality of the energy service [7].

Considered the aforementioned points, space heating and cooling in buildings is promising for the application of demand side management (DSM) strategies aimed at altering the electrical energy demand of the final user. DSM includes all the policies used to influence the customer’s energy curve, focusing on changing the shape of the load and thereby helping to optimize the whole power system from generation to delivery, to end use [7, 8]. DSM strategies can have different objectives, e.g.: peak shaving, valley filling, load shifting, energy conservation, and strategic load growth [7]. One of the most important aspects of DSM is related to the growing share of renewable energy sources (RESs) in the generation mix and consequently to the necessity of managing them due to their intermittent and unpredictable production.

A possible DSM strategy is represented by demand response (DR), which refers to changes in electricity use of end customers from their normal consumption patterns in response to changes in the electricity price over time [8]. Usually, two types of tariffs can be found in the market: time of use (TOU) tariffs, varying on the period of the day, and real-time pricing tariffs, that change with the electricity market price [9]. Based on the electricity price, the heat pump demand can be shifted with the aim of reducing electricity costs, improving the system performance, increasing RESs integration, or improving power system benefits.

Given this premise, it is clear that the study of the energy performance of heat pumps used to meet the hot water and space heating demands of buildings, when the latter are able to unlock their energy flexibility thanks to a DR strategy, is paramount. Thus, in this work a commercial air-to-water heat pump was experimentally characterized to evaluate its energy performance when managed with a DR program. A test rig based on a closed-loop system allowed to reproduce a thermal load in the condenser side. The thermal load was associated to the space heating demand of a virtual building, which was modeled with a resistance-capacitance thermal network. The DR program, based on a real-time pricing control strategy, managed the interaction between the closed-loop system and the virtual building, which was able to unlock its energy flexibility thanks to a variable indoor air setpoint temperature. The results obtained with the DR program, in terms of energy and cost saving as well as thermal comfort, were compared with those obtained with a conventional control based on a fixed setpoint temperature (baseline).

The paper is organized as follows. The experimental setup, that includes the heat pump, the thermal load and the acquisition system, is described in Section 2. Section 3 reports the details of the case study, and how the baseline and the DR program were defined. Section 4 reports the results of the study, and their discussion. The conclusions of the work can be found in Section 5.

2.png

Figure 2. 2R2C resistance-capacitance thermal network of the virtual building

The case study considers that the heat pump is used to meet the space heating demand of a virtual building, which in the experimental setup described in Section 2 is simulated with the heat exchanger fed with cold water collected from the dedicated tank. In particular, two cases were evaluated: in the first case (referred to as baseline) the heat pump works with a fixed indoor air setpoint temperature, while in the second case (referred to as DR) the heat pump is allowed to work with a variable indoor air setpoint temperature, determined by means of a DR program based on a real-time pricing control strategy.

To simulate the thermal behavior of the virtual building, a simple 2R2C resistance-capacitance thermal network was used. As can be seen in Figure 2, this model assigns a capacitance (C_i) to the mass of air in the building, a capacitance (C_w) to the opaque walls of the envelope, a resistance (R_i) to the internal part of the opaque walls (all the layers from the internal surface up to the insulating material, which is not considered), and a resistance (R_e) to the external part of the opaque walls (all the layers from the insulating material, which is included, up to the external surface). The model does not take into account the thermal behavior of transparent surfaces and it is assumed that the design space heating load is only by transmission through the walls. With these assumptions, the thermal behavior of the virtual building is described by the following set of two differential equations:

$\left\{\begin{array}{c}Q_{\mathrm{hp}}=\frac{T_{\mathrm{air}}-T_{\mathrm{w}}}{R_{\mathrm{i}}}+C_{\mathrm{i}} \frac{\mathrm{d} T_{\text {air }}}{\mathrm{d} t} \\ \frac{T_{\mathrm{air}}-T_{\mathrm{w}}}{R_{\mathrm{i}}}=\frac{T_{\mathrm{w}}-T_{\mathrm{amb}}}{R_{\mathrm{a}}}+C_{\mathrm{w}} \frac{\mathrm{d} T_{\mathrm{w}}}{\mathrm{d} t}\end{array}\right.$               (1)

where:

• Q_hp is the thermal power transferred by the heat pump to the building, to balance its transmission heating load;

• T_air is the indoor air temperature;

• T_w is the temperature of the wall;

• T_amb is the outdoor air temperature.

The set of equations (1) has two unknown variables, represented by the variations of T_air and T_w with time; the system can be solved with a finite-difference method, imposing a time interval of 3 s, equal to the acquisition time of the test rig. In this way it is possible to know, for each time interval, the temperatures assumed by both the indoor air and the walls.

The values of the parameters C_i, C_w, R_i and R_e were determined by assuming a design heating load of 2 kW_t, equal to the nominal power of the heat pump. Considering a building located in Ancona, Italy, a design outdoor temperature of 0°C was assumed. The indoor air comfort temperature is 20°C. With these assumptions, the total thermal resistance of the building (R_tot) is equal to 0.01 K/W. If the area of the dispersion surfaces is 129 m², then it is possible to evaluate the average transmittance of the opaque walls and a stratigraphy that guarantees such transmittance. In this work, we referred to a typical stratigraphy reported in the standard UNI/TR 11552, and in this way it was possible to determine the values of the parameters C_i, C_w, R_i and R_e, which are provided in Table 1.

Table 1. Values of the parameters used in the 2R2C network

As introduced above, the DR program proposed in this study is based on a real-time pricing control strategy. In particular, the DR algorithm imposes that the indoor air setpoint temperature is allowed to vary by following the Italian PUN (Prezzo Unico Nazionale, Italian Single Price), which was considered as real-time price applied to the costumer. The PUN is defined by a day-ahead electricity market and has a resolution of one hour. The DR program can be described by the following set of equations:

$T_{\mathrm{sp}, i}=\left\{\begin{array}{l}T_{\mathrm{sp}}+\Delta T\left(\frac{\overline{P U N}-P U N_{i}}{P U N_{\max }-\overline{P U N}}\right), \text { if } P U N_{i} \geq \overline{P U N} \\ T_{\mathrm{sp}}+\Delta T\left(\frac{\overline{P U N}-P U N_{i}}{\overline{P U N}-P U N_{\min }}\right), \text { if } P U N_{i}<\overline{P U N}\end{array}\right.$     (2)

where:

• T_sp,i is the indoor air setpoint temperature evaluated at the i-th hour;

• T_sp is the indoor air setpoint temperature of the baseline, equal to 20°C;

• ΔT is the maximum temperature band, equal to 2°C;

• PUN_i is value of the PUN at the i-th hour;

• $\overline{P U N}$ is the average value of the PUN, calculated for the 24 hours of the reference day;

• PUN_max is the maximum value of the PUN in the reference day;

• PUN_min is the minimum value of the PUN in the reference day.

The set of equations (2) allows the heat pump to consume less electrical energy when its price is high, and to absorb more electrical energy when its price is low. In the first case, the indoor air setpoint temperature is reduced up to -2°C, while in the second case the setpoint can be increased up to +2°C. Thanks to the thermal capacity of the building, it should be possible to maintain comfort conditions even when there is a peak of the thermal load.

A typical trend for the PUN is represented in Figure 3, along with the outdoor air temperature and the indoor air setpoint temperature determined according to the DR strategy provided by the set of equations (2). In order to have an acceptable time for the experimentations, the time interval considered was not referred to an entire day, but only to 8 hours (from 04:00 AM to 12:00 AM, as highlighted in Figure 3).

3.png

Figure 3. PUN, indoor air setpoint temperature for the DR program and outdoor temperature

The delivery temperature of water in the condenser side of the heat pump was set in the range 50-55°C, to simulate a real application where fan coils are used as ambient terminals. The interaction between the DR program and the test rig implies that when the actual value of the indoor air temperature (T_air), determined with the thermal balance (1), is lower than the setpoint value (T_sp,i) determined with the algorithm (2), the electric pump in the circuit of cold water is activated and the fluid is allowed to absorb heat from the condenser side. This condition simulates the ON cycle of the heat pump, where the compressor is activated and there is absorption of electrical energy from the grid. During the ON cycle, the rotational speed of the electric pump is managed by the inverter, whose working frequency is controlled by means of a PID (proportional-integral-derivative) control defined in LabVIEW. On the other hand, when T_air > T_sp,i, the electric pump is deactivated, condition that simulates the OFF cycle of the heat pump.

In this work, the energy performance of a commercial air-to-water heat pump was experimentally evaluated when managed with a demand response (DR) program. The thermal load was associated to the space heating demand of a virtual building, which was modeled with a resistance-capacitance thermal network. The DR program, based on a real-time pricing control strategy, managed the interaction between the closed-loop system and the virtual building, which was able to unlock its energy flexibility thanks to a variable indoor air setpoint temperature. The results obtained with the DR program were compared with those of a baseline condition, obtained with a conventional control based on a fixed setpoint temperature.

The results of the study showed that the use of the DR program allowed to reduce electricity consumption and cost, without altering thermal comfort in a relevant way. Specifically, for the baseline condition the compressor of the heat pump resulted to be active for almost the entire duration of the test (85% of the time), with an absorption of 800-900 W_e. For the DR condition, instead, the compressor was on only when the strategy required a variable setpoint temperature of the indoor air higher than that of the baseline condition (20°C), or when the supply temperature of water was too low; this allowed to reduce the ON cycle of the compressor to 63% of the entire time of the test. Respect to the baseline, the DR program was able to reduce the overall absorption of electrical energy by -18.1%, and the total cost by -25.7%. While in the baseline the internal air temperature was constantly kept at 20°C, with the DR program the actual indoor air temperature was allowed to vary in the range 18-22°C. This variation led to a more variable predicted mean vote (PMV) for the DR case; however, it was verified that PMV was always in the range between -0.5 and +0.5, condition that can be considered acceptable for the indoor thermal comfort.

Based on the results of the study, it is possible to conclude that heat pumps, thanks to their thermal and demand-side flexibility potential, will play a key role in future energy systems and will be fundamental to achieve the 2050 climate targets defined by the EU.