1. Introduction
Ground-coupled heat pumps (GCHPs) are among the general renewable technologies used for space heating, cooling and hot water supply for buildings, especially in cold climates. Significant reductions in energy consumption, 30–70% in heating mode and 20–50% in cooling mode, have been reported when compared with conventional air-conditioning systems [
1]. The ground heat exchanger (GHE), installed vertically or horizontally to exchange heat with the soil, is one of the main components of the GCHP system. Although GCHPs are more efficient than conventional air-to-water heat pumps [
2], the investment cost of the GHE is high and make GCHPs less competitive than the more widespread air source heat pumps (ASHPs), especially in mild climates. Thus, despite having great potential in the future for many countries [
3], the presence of GCHPs in countries with mild climates is currently not very widespread. Some novel and shallow horizontal GHE (HGHE) solutions take advantage of a better heat transfer from advanced shapes, achieving an energy performance similar to that obtained by the more thermally stable, though more expensive, vertical configurations. In this regard, new developments of very shallow geothermal systems combined with different natural and cheap backfilling materials are currently under assessment at several sites around Europe [
4]. However, it seems HGHEs hold some drawbacks regarding land use [
5] and payback still remains too long to justify the initial investment [
6]. Yet, given the existing potential and resources, a major development is expected on shallow geothermal for HVAC (heating, ventilation and air conditioning) and GCHP systems [
7,
8,
9]. In this context, the use of phase change materials (PCMs) may also assist GCHP technology in the search for making the systems more energetically efficient and economically viable. Thus, Spitler and Bernier [
10] stated that the application of PCMs inside boreholes could be a viable means to reduce borehole length at peak conditions. Due to its higher thermal capacity and latent heat release during phase change, PCM is able to improve the energy storage performance of the GHE and slow down the temperature change of soil. On that basis, Eslami-nejad and Bernier [
11] designed a ring with a mixture of PCM and soil around the borehole, allowing a reduction in borehole length of up to 9%. From laboratory-scale experiments and numerical models, several authors [
12,
13,
14] have shown that the thermal influence radius of the GHE, and consequently the land area needed for boreholes, can be decreased by utilising PCMs as backfill material. The effect of using PCMs as grout on the thermal performance of GHEs on their own or coupled with the heat pump have also been analysed. Results provided by several studies [
15,
16] showed that a proper mixture of PCM and soil as grouting material of the GHE was able to smooth the thermal wave generated by the heat pump on the ground, and to enhance its coefficient of performance (COP). Some authors [
17,
18] have pointed out that, in hybrid systems such as dual source heat pumps (DSHPs), which change between ground and air as heat sources, the use of an air heat exchanger allows a further reduction of the GHE size, lowering the total cost of the GCHP system. Furthermore, DSHPs can be optimised by switching to the more favourable source/sink between the air and ground according to their temperature, achieving higher efficiencies in comparison with ASHPs and GCHPs [
19,
20,
21]. This solution may also avoid frosting problems and save the system against extreme temperatures that affect air-mode by using the ground as an alternative heat source. Due to the spread of heat pump (HP) systems and the incontrovertible climate change, the resilience of HVAC systems to adverse weather conditions is a key factor in ensuring internal comfort.
Considering the potential of DSHPs and the advantages of using PCMs in GHEs, the present work numerically analyses the coupling of a DSHP and a novel flat-panel HGHE [
21] with a mixture of sand and PCMs as backfill material into the trench. This combination, used both for space heating and cooling, has not been considered or analysed previously in the scientific literature. Thus, one of the main objectives has been to check the suitability of this system from energy numerical simulations by comparison with different possible combinations. First, the methodology proposed to assess system thermal performance is described. Model hypotheses, boundary conditions and the study cases used for comparison are shown in this section. Secondly, the results from the numerical simulations are discussed. Finally, the main conclusions of the study are drawn.
3. Results and Discussion
All models were run for a two-year period by repeating weather data of the year 2015 and under the same boundary conditions, in order to set the typical initial conditions of this kind of facility and to ensure that all models reproduced their stationary trend according to their different exploitations. Thus,
Figure 8a shows flat-panel temperatures for the second simulation year cases DP00, DP01, DP10 and DP11, together with outdoor air and undisturbed ground temperature (at a depth of 1.7 m). Similarly,
Figure 8b shows cases DP11, DP11*, DP11# and DP11+. As it can be inferred from the series of undisturbed ground temperature, despite ground thermal exploitation and unlike deep geothermal systems, thermal drift is avoided by using this shallow HGHE.
Flat-panel temperature for DSHP cases (DP10 and DP11) is higher than for GCHP cases (DP00 and DP01), since the dual system is able to switch to the air when ground temperature is lower in winter (
Figure 8a). In this way, the lowest flat-panel temperatures for DP10 and DP11 cases are 2.1 °C and 1.7 °C, respectively. This issue may avoid the use of anti-freeze (e.g., propylene glycol) in the secondary loop of the system and reduce or exclude all frosting problems at the air heat exchanger. In contrast, the lowest flat-panel temperatures for GCHP cases DP00 and DP01 are −1.3 °C and −2.4 °C, such that anti-freeze usage is needed. As a consequence, dual strategy leaves a warmer ground at the beginning of summer which is less advantageous for this period. On the contrary, the lower temperature of GCHP systems makes the condition disadvantageous in wintertime, but more profitable in summer. When PCMs are used in the backfill material (DP01 and DP11) a similar behaviour is found, although the performance of the system is clearly lower in wintertime (minimum temperatures of −2.4 °C and 1.7 °C). These results are mainly due to the low thermal conductivity of the mixture in comparison with that of the sand. However, despite this drawback, the performance is improved in summertime, according to the higher cold energy stored in the ground, as maximum temperatures of 24.9 °C and 25.7 °C are achieved for DP01 and DP11, respectively. In
Figure 8b, DP11* and DP11# minimum temperatures are −0.3 °C and −0.6 °C in wintertime, respectively, whilst 1.3 °C is that of DP11+. In summertime, the maximum temperatures are 26.1 °C, 26.2 °C and 24.7 °C for DP11*, DP11# and DP11+, respectively.
Overall, dual systems are able to perform well during extreme weather conditions (very low or very high outdoor air temperatures) for which a sole ASHP system would be unable either to work or perform efficiently. This makes the dual system a robust alternative in Southern European countries in which weather conditions are expected to become more severe in the future, with higher inter-annual increase in summer temperatures and low variability in current winter temperatures [
28].
Details of the annual trend are shown at the beginning of the year in
Figure 9, and by the middle of the year in
Figure 10; that is, in winter and summer, respectively. When comparing, it should be noted that case DP11* has a higher
lf factor (30 m
3/m), DP11# also includes a narrow trench and consequently a smaller quantity of PCMs, whilst DP11+ has a standard trench (60 cm wide) and load factor (20 m
3/m), but uses PCMs with higher thermal conductivity.
In
Figure 9a, the steady decrease of flat-panel temperature in January due to a continuous ground exploitation carried out by the GCHP (DP00 and DP01) contrasts with the nearly constant trend shown by the DSHP (DP10, DP11, DP11*). DP11+ shows the promptest celerity in recovering the exploitation in comparison with all other DSHPs (
Figure 9b). Cases with PCMs show better performance than those without (
Figure 10a), whilst the aforementioned behaviour of DP11+ is confirmed also in summertime (
Figure 10b). Therefore, this last property seems to be the key factor in using PCMs: not to penalise the improvement of energy storage with a lower thermal conductivity of the backfill material.
In
Figure 11, the resulting performance of the GCHP (DP00, DP01) and DSHP (DP10, DP11) are depicted in terms of COP (wintertime,
Figure 11a) and EER (summertime,
Figure 11b), according to the previous Equations (1) and (2). GCHP cases show better COP values than DSHP cases at the beginning of winter, when the ground temperature is very high due to the previous heating up in summer. However, the reverse happens late in winter, when the ground has been deeply exploited by the GCHP operation, and partially saved by the DSHP mode. This inversion does not happen in summertime (EER), since the lowest ground temperature performs better during the whole summer.
A summary of the energy exchange and performance of the different systems in winter, summer and the whole year is shown in
Table 3,
Table 4 and
Table 5, respectively. The values of thermal (
Qt) and electrical (
Qe) energy exchanged per unit length of trench (kWh/m) are given too.
In wintertime (
Table 3), DP10 and DP11 perform similarly, and better than DP00 and DP01. Specifically, DP11 shows a higher COP than DP01 (3.44 against 3.24), and also DP10 (COP value of 3.43) compared to DP00 (COP value of 3.26). Furthermore, DP11* and DP11# achieve very similar performance if compared with DP11, but with a significant reduction of the HGHE size (
lf of 30m
3/m against 20 m
3/m), and a narrower trench for case DP11# (40 cm against 60 cm).
Due to the much shorter use of the ground heat exchanger carried out by the DSHP cases (DP10, DP11, DP11*, DP11# and DP11+), very high values of heat flux per metre of trench are obtained, that range from more than three times for case DP11+ up to almost five times for case DP11*, when compared with GCHP. This is mainly related to the shorter ground exploitation of the DSHP cases, for which heat transfer takes from 500 h (DP11#) up to 720 h (DP11+).
This seems to reflect a good performance of the dual system even if coupled with a shorter HGHE, and that PCMs can improve the system performance only if their thermal conductivities are higher. Therefore, the common low thermal conductivity of PCMs attenuates their potential good performance in wintertime related to their underground thermal energy storage.
In comparison with the cheapest ASHP which shows an overall COP value around 3.31, the best value of a DSHP, around 3.46 (DP11#) does not seem to justify the investment of PCMs and GHE. However, it should be noted that the air temperature drops below 3 °C for 747 h over 5228 h of heating time, that is, 14.2% of the working period, in which frosting issues are very common at this location. As a consequence, the ASHP performance given here should be considered overestimated up to a value around 15%, as reported by Dongellini and colleagues [
29].
In summertime, DP01 shows the best performance (EER value of 4.37), as justified by the ground cooling occurring for long thermal exploitation in wintertime, and the presence of PCMs. The best DSHP case is DP11# with an EER value of 4.02, resulting from the 4.24 and 3.80 values in ground and air exploitation, respectively. It is worth noting that DP11# uses a GHE which is three times shorter than the one used by case DP00. The ASHP performance is the lowest (EER value of 3.61 value) and a more interesting gap is evident, especially related to a larger exploitation of the GHE. In summer season, the overall needs of air conditioning cover a period of 1304 h; when the DSHP is used, the ground exploitation goes from a minimum of 488 h (37.4%, DP11#) to a maximum of 635 h (48.7%, DP11+). The average heat flux occurring at the flat-panel GHE (qFP_ave) shows a minimum of 50.4 W per length metre of trench in GCHP cases (DP00, DP01) and a maximum of 142.8 W/m in DSHP case (DP11#), which demonstrates the high performance of the flat-panel. qFP_ave values are quite comparable in winter and summer for DSHP cases, while winter values halved those in summer for GCHP cases. This difference is always attributable to the exploitation time of the ground, unchanged for DSHP cases (500–700 h) unlike GCHP cases (with more than 5000 h in wintertime against 1300 h in summertime).
As expected from the annual trend of flat-panel temperature, all DSHP cases show higher efficiencies than GCHP cases. Yet, they still perform better than the ASHP. Regarding the heat flux per metre of trench, DSHP cases provide values which are several times those given by GCHP cases. During the whole year, DP00 and DP01 show COP/EER values of 3.42, the ASHP of 3.37 and the DSHP from 3.48 to 3.56. Overall, the average heat flux occurring at the flat-panel, coupled with a DSHP system, is three to five times higher than that of a GCHP.
Finally, an analysis about how the PCM behaves in the different cases was carried out. Thus,
Figure 12 shows the solid phase time series in terms of mass fraction for the trench. In winter, a more efficient use of PCM1 is made by DP01 (
Figure 12a) than by DP11 (
Figure 12b). Hence, PCM1 solidifies reaching a peak of 100% (DP01), while it is of 80% in DP11. A slight improvement is achieved by DP11*, while DP11+ gets slightly lower results (
Figure 12c,d). In summertime, the fraction of PCM2 that becomes liquid is very similar in all cases. This result could also suggest the advisability of using backfill materials in the trench with higher thermal conductivity.
To complete the results depiction, a sequence of the thermal field at 41.333 days of simulation time is presented in
Figure 13 for different cases. The first two images (DP00, DP01) show the large exploitation carried out by the GCHP in comparison with the DSHP; no relevant differences are shown when PCMs are used in coupling with a GCHP. More interesting differences are detectable among DSHP cases. In the sequence DP10, DP11 and DP11*, the ground temperature rises in the domain from the bottom to the top, whilst the flat-panel temperature decreases. Temperature distribution for DP11 and DP11# cases are quite similar, while DP11+ shows the highest temperature values among all.