Applied Thermal Engineering 114 (2017) 1073–1081

Contents lists available at ScienceDirect

Applied Thermal Engineering

journal homepage: www.elsevier.com/locate/apthermeng

Research Paper

Sizing effects on the energy performance of reversible air-source heat

pumps for office buildings
Matteo Dongellini ⇑, Claudia Naldi, Gian Luca Morini
Department of Industrial Engineering (DIN), Alma Mater Studiorum – University of Bologna, Viale Risorgimento 2, 40136 Bologna, Italy

h i g h l i g h t s

A study on the annual performance of air-to-water heat pumps is presented.

The influence of HP modulating capability, climate and system sizing is shown.

Mono-compressor, multi-compressor and inverter-driven heat pumps are considered.

SCOP is maximum when HP capacity is tuned on building peak load.

Best SEER is obtained for HPs slightly oversized with respect to design peak load.

a r t i c l e i n f o a b s t r a c t

Article history: This paper deals with the numerical analysis of the energy performance of HVAC systems for heating and
Received 17 August 2016 cooling, based on a reversible electrical air-source heat pump. The aim of this study is to highlight in
Revised 24 October 2016 which way the energy consumption of these systems is influenced by the heat pump modulating capa-
Accepted 4 December 2016
bility, the heat pump sizing and the climate. A numerical model based on the bin method and the energy
Available online 5 December 2016
signature procedure, able to take into account the performance of the heat pump at partial load, has been
followed in order to calculate the seasonal and annual performance of the HVAC system coupled to a typ-
Keywords:
ical office building, located in three different European locations (Frankfurt, Istanbul, Lisbon). Different
Air-source heat pump
Reversible heat pump
kinds of heat pumps (mono-compressor on-off, multi-compressor and inverter-driven heat pumps) have
Seasonal performance factor been considered. The numerical results presented in this paper point out that the heat pump sizing
Annual performance factor strongly affects seasonal and annual HVAC energy consumption: the ratio between the full thermal
Heat pump oversizing capacity of the heat pump and the building peak load strongly influences the energy performance of
Energy efficiency the system. It has been demonstrated that, depending on the season, the seasonal performance of the
above-mentioned kinds of heat pumps are influenced in a different way by their sizing.
Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction
During the last decade, the European Union (EU) has made a
strong effort in order to achieve a significant reduction of both
new and old building energy consumptions. A series of specific
Directives [1–3], aimed to promote the use of renewable energies
and to improve the energy efficiency in buildings and HVAC sys-
tems, have been published and their application is deeply changing
the national regulations. The Directive 2009/28/EC (RES Directive)
[1], in particular, highlights the potential of technologies based on
renewable energy sources and promotes this new generation of
high-efficiency HVAC systems to provide heating, cooling and air
ventilation in buildings. Heat pump systems are one of the most
suitable solutions for energy efficiency improvement in buildings.
⇑ Corresponding author.
E-mail address: matteo.dongellini@unibo.it (M. Dongellini).
In 2013 about 56% of the total heat pumps sold in Europe were
reversible units, and about 91% of them were air-source systems
[4]. It has been demonstrated that air-source heat pumps can con-
tribute to reduce the primary energy consumption of a building
with respect to traditional systems [5]; however, the potential
energy saving linked to the use of air-source heat pumps depends
on the correct selection of the heat pump size, on the basis of the
building seasonal energy needs and on outdoor conditions [6]. In
the last years, many researchers have investigated which parame-
ters mainly influence the seasonal energy performance of air-
source heat pumps [7–10] and specific numerical models have
been proposed in order to estimate the Seasonal Energy Efficiency
Ratio (SEER) and the Seasonal Coefficient Of Performance (SCOP) of
these devices coupled to a building. As an example, Kinab et al. [11]
presented a numerical model of a modulating reversible air-source
heat pump, by means of which the optimal design of the system
can be obtained. The interaction between the heat pump character-

http://dx.doi.org/10.1016/j.applthermaleng.2016.12.010
1359-4311/Ó 2016 Elsevier Ltd. All rights reserved.
1074 M. Dongellini et al. / Applied Thermal Engineering 114 (2017) 1073–1081
Nomenclature
APF Annual Performance Factor [–]
BES Building Energy Signature [–]
Cc degradation coefficient for on-off cycles [–]
COP Coefficient Of Performance [–]
DD Degree Days [–]
EER Energy Efficiency Ratio [–]
E energy [kW h]
I total solar irradiation [W/m2]
h heat transfer coefficient [W/m2 K]
N HVAC system operating hours [h]
P heat pump thermal capacity [kW]
PLF Part Load Factor [–]
PLR Part Load Ratio [–]
SCOP Seasonal Coefficient Of Performance [–]
SEER Seasonal Energy Efficiency Ratio [–]
T temperature [°C]
Greek symbols
a downsizing/oversizing parameter [–]
b unbalanced loads indicator [–]
istics and the building heating/cooling loads has been investigated
numerically by Madonna and Bazzocchi [12]. The authors pointed
out that the ratio between heating and cooling peak loads strongly
affects the annual performance of air-source heat pumps; in build-
ings with unbalanced loads the heat pump seasonal efficiency
could experience a marked decrease, down to 25%, due to the
excess of on-off cycles.
Busato et al. [13] reported a SCOP comparison between single-
stage electrical and gas-driven air-to-water heat pumps, installed
in residential buildings, by considering three different Italian sites.
The study shows that heat pumps have a huge potential to reduce
primary energy consumption and CO2 emissions, but they are char-
acterized by a greater investment cost with respect to traditional
systems.
The analysis of the literature in this field puts in evidence that
the evaluation of the energy saving potentiality of HVAC systems
based on different kinds of electric-driven heat pumps (i.e.
mono-compressor, multi-compressor, inverter-driven), with
respect to traditional heating and cooling systems, is still incom-
plete. Some authors focused their research on the optimization of
mono-compressor [8–9,13] or inverter-driven [7,10,12] heat
pumps, but a complete comparison between the seasonal perfor-
mance obtainable by adopting these heat pumps as a function of
the design thermal loads of the building is still missing. In this
paper, the estimation of the seasonal performance, both in heating
and cooling mode, of different models of electric-driven air-to-
water heat pumps, located in different European climates, is
shown. The aim of the work is the assessment of the optimal sizing
rules for this kind of heat pumps, in order to obtain the best Annual
Performance Factor (APF) as a function of climate, building thermal
loads and heat pump characteristics. The relationship between
outdoor conditions, heating and cooling loads, modulating capabil-
ity and the system efficiency is presented in order to provide a
complete set of information to HVAC designers.
2. The case study
2.1. Building characteristics
A typical three-story office building having the characteristics
summarized in Table 1 has been considered. The working profile
D extra infrared radiation (external surface-sky) [W/m2]
e absorptance coefficient of external surface [–]
Subscripts
avg average value
base reference value
BU delivered by back-up
c cooling value
des design value
fc full capacity value
el electric value
ext outdoor value
h heating value
HP delivered by heat pump
month monthly value
o at outer surface
sa sol-air
supply delivered value
of the HVAC system coupled to the office building reflects its occu-
pancy profile: during workdays (from Monday to Friday) the heat-
ing/cooling plant is switched on from 6.00 to 20.00, on Saturday
the HVAC system is activated from 6.00 to 14.00 while on Sunday
the system is turned off.
2.2. Climatic data and choice of the sites
In order to take into account the influence of different climates
on the annual performance of the heating/cooling system, the
office building has been placed in three different locations: Frank-
furt (50.6°N, 8.4°E), Istanbul (41.0°N, 41.6°E) and Lisbon (38.4°N,
9.7°E), which can be considered representative, according to the
goal of this work, of Colder, Average and Warmer heating reference
climates indicated by the standard EN 14825 [14] (even if different
from Helsinki, Strasburg and Athens, cited by the standard).
The local outdoor temperature bin trend can be calculated by
means of the hourly outside temperature distribution of the Typi-
cal Meteorological Year (TMY) defined for the specific location, as
pointed out by Naldi et al. [15]. In this paper, bins of both heating
and cooling season have been calculated by using the TMYs of the
Meteonorm database (version 5.0.13) included in TRNSYS 17 [16].
In Table 2 the duration of the heating and cooling season is
reported together with the typical value of the Heating Degree
Days (DDh) and Cooling Degree Days (DDc) of each location. The
calculation of DDc has been performed by using the sol-air temper-
ature Tsa [17], by taking into account the effect of incident solar
radiation on building envelope. Sol-air temperature is an indicator
more useful than external air temperature to consider the severity
of the cooling season and it is defined as:
eI  D
T sa ¼ T ext þ ð1Þ
ho
where e is the surface solar absorptance of the building envelope, I
is the global solar irradiance on the envelope, ho is the heat transfer
coefficient for long-wave radiation and convection at the outer sur-
face and D is the extra infrared radiation due to difference between
the external air temperature and the apparent sky temperature (cal-
culated by means of UNI TS 11300-1:2014 [18]).
To evaluate the effect of defrosting cycles during the heating
season, the percentage of the heating season in which the outdoor
conditions fall within the frosting region [19] (i.e. air temperature
 M. Dongellini et al. / Applied Thermal Engineering 114 (2017) 1073–1081 1075

Table 1
Geometrical and thermal characteristics of the selected office building.

Net floor area Total surface area Total windows area Net volume S/V ratio External wall U-value Floor U-value Windows U-value
(m2) (m2) (m2) (m3) (m1) (W/m2 K) (W/m2 K) (W/m2 K)
787 1416 139 2125 0.52 0.37 0.46 1.70

Table 2
Main climatic data of the selected sites.

Frankfurt Istanbul Lisbon

Heating season
Length of the season From 1st October to 30th April From 1st November to 15th April From 1st December to 31st March
DDh (Tbase,h = 20 °C) 3233 1973 863
Frosting period 60% 36% 3%
Cooling season
Length of the season From 1st June to 31st August From 1st June to 30th September From 1st May to 30th September
DDc (Tbase,c = 26 °C) 405 1079 1251

below 6 °C and relative humidity above 50%), is shown in Table 2
for each location. A penalty coefficient has been applied on the heat
pump performance during the hours falling in the frosting region,
by following the simplified model by Vocale et al. [20]. With the
method suggested in [20], the number of defrost cycles can be esti-
mated by considering both the external conditions and the heat
pump characteristics; in this way it becomes possible to link the
penalty coefficient to the heat pump modulating capability.
The bin distributions of the selected locations present different
trends, as evidenced in Fig. 1. Fig. 1a highlights that the heating
season in Frankfurt is the coldest among the selected climates: it
presents a design temperature Tdes,h of 10 °C with an average
value Tavg,h of 4.9 °C. Istanbul and Lisbon are characterized by a
milder heating season, with Tdes,h equal to 3 °C and 4 °C,
respectively, and Tavg,h equal to 8.3 °C and 11.6 °C, respectively.
The hot seasons of the selected locations are more similar to one
another, as evidenced by Fig. 1b. The most severe condition for the
heat pump is set in Lisbon, with a value of the design temperature
Tdes,c of 32 °C and with an average outdoor temperature Tavg,c equal
to 20.7 °C. Istanbul cooling season is characterized by similar
values (Tdes,c equal to 30 °C and Tavg,c equal to 22.4 °C) but the
length of the season is one month shorter than in Lisbon. Finally,
Frankfurt presents a cooling design temperature of 29 °C and an
average air temperature equal to 18.1 °C; in this location bins are
shifted towards lower outdoor temperature values.
2.3. Heating and cooling loads calculation

250 The calculation of building net thermal energy needs for space
Bin (hours)

Frankfurt heating and cooling has been performed on a monthly basis

200 Istanbul according to the standard EN ISO 13790:2008 [21]. As a result,
the monthly net thermal energy required by the building for
Lisbon
heating and cooling (Emonth,h and Emonth,c, respectively) has been
150
obtained.
Considering an operating time for the HVAC system that follows
100 the activation profiles described previously, both during the heat-
ing and cooling season, it is possible to calculate the monthly mean
50 thermal power that the heat pump (Pmonth,h/c) has to supply to the
building by dividing the thermal energy needs for the number of
0 operating hours (Nmonth).
-10 -8 -6 -4 -2 0 2 4 6 8 10 12 14 16 18 20 The monthly values of Pmonth,h/c can be correlated to the monthly
Outdoor air temperature (°C) average values of the outdoor air temperature (or sol-air tempera-
(a) ture in summer) via the Building Energy Signature (BES) method
(Fels [22], UNI EN 15603 [23]), which considers the outdoor tem-
200
Bin (hours)

perature (or the sol-air temperature during the hot season) as

Frankfurt the only forcing term of building thermal needs.
Istanbul In Fig. 2 Pmonth,h/c values are represented as functions of the out-
150
Lisbon door monthly average temperature in winter and sol-air tempera-
ture in summer; the linear interpolation of these data represents
100 the BES, both for heating and cooling season, associated to the
office building in the three selected sites. BESs obtained for the
same building located in Frankfurt, Istanbul and Lisbon are repre-
50 sented in Fig. 2a–c, respectively.
By observing Fig. 2, it is evident that in Frankfurt the heating
peak load (Pdes,h = 31.3 kW) is larger than the cooling peak load
0
(Pdes,c = 13.6 kW); on the other hand, in Istanbul seasonal peak
5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35
Outdoor air temperature (°C)
loads are more balanced (Pdes,h = 20.1 kW, Pdes,c = 15.8 kW) like in
(b) Lisbon, where the maximum peak load is associated to the cooling
season (Pdes,h = 14.1 kW, Pdes,c = 16.3 kW). Frankfurt is representa-
Fig. 1. Bin trend for the selected locations during heating (a) and cooling season (b). tive of typical situations in which the maximum thermal load is
1076 M. Dongellini et al. / Applied Thermal Engineering 114 (2017) 1073–1081

40
Thermal power (kW) Building heating load

Building cooling load

30
Pdes,h

20

Pdes,c
10

Tdes,h Tdes,c
0
-10 -5 0 5 10 15 20 25 30 35 40
Outdoor air temperature (°C) (a)

40
Thermal power (kW) Building heating load

Building cooling load

30

20 Pdes,h
Pdes,c

10

Tdes,h Tdes,c
0
-10 -5 0 5 10 15 20 25 30 35 40
Outdoor air temperature (°C) (b)

40
Thermal power (kW) Building heating load

Building cooling load

30

20
Pdes,c
Pdes,h
10

Tdes,h Tdes,c
0
-10 -5 0 5 10 15 20 25 30 35 40
Outdoor air temperature (°C) (c)

Fig. 2. Building energy signature for heating and cooling season: Frankfurt (a), Istanbul (b) and Lisbon (c).

reached during winter and strongly unbalanced loads are present
moving from the heating to the cooling season. Istanbul is repre-
sentative of sites in which maximum thermal loads are always
reached during winter, but loads moving from the cold to the hot
season are only slightly unbalanced. Lisbon is representative of
locations in which the cooling season is more severe than the heat-
ing one, but loads are balanced from one season to the other.
An indicator of the existence of seasonal unbalanced loads can
be introduced by the ratio b:
Pdes;c
b¼ ð2Þ
Pdes;h
when b = 1 the absolute value of the design thermal power is the
same both for heating and cooling; when b > 1 the building thermal
loads during summer are larger than the winter loads and the oppo-
site occurs for b < 1. The three examples in this paper are character-
ized by b = 0.43 (Frankfurt), b = 0.79 (Istanbul) and b = 1.16 (Lisbon).
 M. Dongellini et al. / Applied Thermal Engineering 114 (2017) 1073–1081 1077

2.4. Description of heat pumps characteristics
The HVAC system coupled to the office building is based on a
single reversible electric heat pump, which satisfies both heating
and cooling loads; electric heaters have been considered as back-
up system during the cold season, whereas no back-up systems
are considered for the hot season. The building terminal units con-
sist of two-pipe fan-coils, sized with a supply water temperature of
45 °C during winter and 7 °C in summer.
Three different heat pump typologies are considered: mono-
compressor ON-OFF Heat Pumps (ON-OFF HPs), Multi-
Compressor Heat Pumps (MCHPs) and Inverter-Driven Heat Pumps
(IDHPs).
For each heat pump typology, different units, which differ for
their nominal thermal capacity, have been selected: the most pow-
erful heat pump model for each typology has been chosen in order
to match the peak load during the most severe season. Then, other
smaller units have been considered, in order to evaluate the influ-
ence of heat pump sizing on the seasonal and annual performance.
For Frankfurt and Istanbul climates (b < 1), the choice of the small-
est nominal thermal power is based on the cooling peak load. For
Lisbon (b > 1), the heating and cooling loads are almost balanced
but the cooling season is slightly more severe; therefore, the heat
pump sizing is based on the cooling peak load. In fact, even if smal-
ler heat pumps would be more suitable during the cold season,
then the cooling demand would not be satisfied during the whole
hot season, since no back-up system for cooling is present in the
HVAC system.
Eleven different commercial reversible air-to-water heat pumps
have been considered; technical data at full capacity given by the
manufacturer are reported in Tables 3–5 for ON-OFF HPs, MCHPs
and IDHPs, respectively. Each heat pump is identified by a progres-
sive number (# HP unit).
The calculation of the seasonal energy performance of the
above-mentioned heat pumps, coupled to the office building
described before, has been performed by following the numerical
method proposed by Dongellini et al. [24], which takes into
account the different modulating capability of the heat pump
typologies and calculates the heat pump efficiency at partial loads.
This aspect is crucial for a correct analysis of the seasonal and
annual energy performance of the system, because heat pumps
work at full load only for a limited period of the whole heating/-
cooling season. More in detail, the evaluation of SCOP and SEER
for the proposed case study has been carried out according to the
‘‘fixed outlet” approach reported by the standard EN 14825 [14];
in the present analysis, the supply water temperature has been
fixed at 45 °C during the heating season and at 7 °C during the
cooling season.
At partial loads, the simplest solution to fit the heat pump
capacity to the building load is to perform a series of on-off cycles,
which, however, have a strong impact on heat pump efficiency and
consequently on the overall seasonal energy performance of the
unit [25]. In fact, for each restart, the compressor restores the pres-
sure difference between condenser and evaporator and, during off
periods, energy consumption due to auxiliary appliances is still
active. The capability of multi-compressor and inverter-driven
heat pumps to reduce the use of on-off cycles is one of the reasons
of their better performance compared with mono-compressor on-
off heat pumps. In addition, it is important to take into account that
at partial loads the heat exchanger surfaces become oversized and
the pressure ratio between condensation and evaporation is
reduced; all these effects improve the performance of these heat
pump typologies at partial loads.
To take into account the above-mentioned degradation of COP
or EER at partial loads, several researchers [6,8,10–12,15] and stan-
dards [14] have proposed to correct the full load COP/EER by means
of a part load factor (PLF), defined as follows:
PLR
PLF ¼ ð3Þ
ð1  C C Þ þ C c PLR
where Cc is a parameter that takes into account on-off losses and
the part load ratio (PLR) is defined as the percentage of the full load
capacity delivered by the heat pump at partial load. As highlighted
by Madonna and Bazzocchi [12], the value of Cc strongly influences
on-off losses: for Cc = 1 no degradation effects are considered, while
cycling losses increase if Cc decreases. The value of Cc suggested by
UNI EN 14285 [14] is 0.9, but experimental tests have highlighted
that values from 0.7 to 0.9 can be in some cases more realistic: as
an example, Madonna and Bazzocchi [12] have used for their
variable-speed air-to-water heat pump values of the degradation
coefficient equal to 0.73 and 0.83 for heating and cooling modes
respectively.
For MCHPs and IDHPs, the model used in this paper [24] allows
to take into account the modulation range of these devices: within
this range, the model calculates heat pump performance at partial
load by evaluating the characteristic curves of the unit on the basis
of the number of compressors switched on (MCHPs) or the effec-
tive inverter frequency (IDHPs). Outside the modulation range,
heat pump COP/EER is corrected by means of PLF from Eq. (3), by
assuming Cc = 0.9.
3. Discussion of the results
The seasonal energy performance of a HVAC system character-
ized by a single reversible unit, is heavily affected by the device siz-
ing. In the case of strongly unbalanced seasonal peak loads, as

Table 3
Thermal/cooling capacity and COP/EER at full load of the selected on-off heat pumps.

# HP unit 1 2 3 4
HP typology ON-OFF HP ON-OFF HP ON-OFF HP ON-OFF HP
Heating mode (Tsupply,h = 45°C)
Text (°C) Ph (kW) COP Ph (kW) COP Ph (kW) COP Ph (kW) COP
7 17.8 2.55 20.0 2.56 21.9 2.58 26.4 2.59
2 20.6 2.94 23.6 2.96 25.9 2.97 31.6 2.95
7 22.9 3.31 27.0 3.33 29.8 3.36 36.3 3.37
12 25.7 3.74 31.0 3.76 34.3 3.81 41.1 3.78
Cooling mode (Tsupply,c = 7°C)
Text (°C) Pc (kW) EER Pc (kW) EER Pc (kW) EER Pc (kW) EER
20 22.3 4.28 27.2 4.30 31.0 4.31 35.9 4.33
25 21.2 3.73 26.1 3.80 29.5 3.76 34.3 3.82
30 20.2 3.30 24.7 3.32 27.8 3.26 32.7 3.35
35 19.2 2.82 23.3 2.88 26.1 2.79 30.9 2.86
1078 M. Dongellini et al. / Applied Thermal Engineering 114 (2017) 1073–1081

Table 4
Thermal/cooling capacity and COP/EER at full load of the selected multi-compressor heat pumps.

# HP unit 5 6 7 8
HP typology MCHP MCHP MCHP MCHP
Compressors 2 2 2 2
Heating mode (Tsupply,h = 45°C)
Text (°C) Ph (kW) COP Ph (kW) COP Ph (kW) COP Ph (kW) COP
7 14.6 2.47 16.7 2.42 20.0 2.48 24.6 2.44
2 19.1 3.04 21.8 2.97 26.0 3.00 32.1 2.99
7 22.2 3.36 25.1 3.31 30.1 3.35 37.3 3.33
12 25.6 3.67 29.1 3.71 34.8 3.75 43.0 3.72
Cooling mode (Tsupply,c = 7°C)
Text (°C) Pc (kW) EER Pc (kW) EER Pc (kW) EER Pc (kW) EER
20 21.8 4.62 25.3 4.68 30.7 4.64 37.1 4.67
25 20.8 4.13 24.2 4.18 29.3 4.08 35.6 4.15
30 19.9 3.59 22.8 3.62 27.9 3.57 34.0 3.66
35 18.8 3.15 21.6 3.18 26.4 3.11 32.4 3.20

Table 5
Thermal/cooling capacity and COP/EER at full load of the selected inverter-driven heat pumps (data evaluated for inverter frequency equal to 120 Hz).

# HP unit 9 10 11
HP typology IDHP IDHP IDHP
Frequency range (Hz) 20–120 20–120 20–120
Heating mode (Tsupply,h = 45°C)
Text (°C) Ph (kW) COP Ph (kW) COP Ph (kW) COP
7 15.1 2.11 21.3 2.10 24.4 2.14
2 18.6 2.52 26.1 2.54 30.1 2.55
7 21.2 2.82 29.6 2.76 33.2 2.78
12 24.6 3.22 33.8 3.18 38.6 3.19
Cooling mode (Tsupply,c = 7°C)
Text (°C) Pc (kW) EER Pc (kW) EER Pc (kW) EER
20 21.9 3.89 30.7 3.85 34.4 3.81
25 20.9 3.37 29.4 3.33 32.9 3.31
30 19.7 2.89 28.0 2.85 31.3 2.86
35 18.5 2.46 26.6 2.42 29.6 2.45

obtained for Frankfurt, the heat pump can fit exactly only one of
these two peak loads.
In Table 6 the heating and cooling full capacity (Pfc,h, Pfc,c) of the
selected heat pumps, evaluated at the design outdoor temperature,
are shown and compared with the heating and cooling peak loads
of the office building situated in the three considered locations. It is
possible to check the heat pump oversizing/downsizing with
respect to the design loads by introducing the heating and cooling
oversizing parameters (ah, ac) defined as:
Pfc;h  Pdes;h
ah ¼
Pdes;h
P fc;c  Pdes;c
ac ¼ ð5Þ
Pdes;c
A value of a equal to 0 means that the selected heat pump is
able to deliver exactly the building design thermal load; for a < 0
the heat pump is downsized with respect to the building loads,
while a > 0 indicates an oversized heat pump.
By fixing a heat pump, its downsizing/oversizing depends on
the location of the building. For example, the MCHP #6 is strongly
oversized during the cooling season, with respect to the cooling
peak loads of the office building situated in Frankfurt and Istanbul
ð4Þ
(ac = 70% and ac = 45%, respectively). On the other hand, the heat
pump #6 is downsized with respect to the heating loads in Frank-

Table 6
Full heating and cooling capacity (Pfc,h Pfc,c) at the design temperature of the selected heat pumps and percentage of their oversizing (or downsizing) in heating and cooling mode
for the typical office building as a function of its location.

Location Frankfurt Istanbul Lisbon

Unit full Down/Oversizing Unit full Down/Oversizing Unit full Down/Oversizing
capacity at Tdes parameter capacity at Tdes parameter capacity at Tdes parameter
(kW) (kW) (kW)
# HP unit Pfc,h Pfc,c ah ac Pfc,h Pfc,c ah ac Pfc,h Pfc,c ah ac
1 (ON-OFF) 16.3 20.4 48% +50% 18.5 20.2 8% +28% 21.5 19.8 +53% +22%
2 (ON-OFF) 18.2 25.0 42% +84% 20.9 24.7 +4% +57% – – – –
3 (ON-OFF) 19.9 28.2 36% +107% 22.9 27.8 +14% +76% – – – –
4 (ON-OFF) 23.2 33.0 26% +143% – – – – – – – –
5 (MCHP) 13.4 20.1 57% +48% 16.5 19.9 18% +26% 20.3 19.4 +44% +19%
6 (MCHP) 15.4 23.1 51% +70% 18.8 22.9 6% +45% – – – –
7 (MCHP) 18.4 28.2 41% +107% 22.4 27.9 +12% +77% – – – –
8 (MCHP) 22.6 34.3 28% +153% – – – – – – – –
9 (IDHP) 14.4 20.0 54% +47% 16.4 19.7 18% +25% 19.6 19.2 +39% +18%
10 (IDHP) 20.2 28.3 35% +108% 23.2 28.0 +16% +77% – – – –
11 (IDHP) 23.3 31.6 26% +133% – – – – – – – –
 M. Dongellini et al. / Applied Thermal Engineering 114 (2017) 1073–1081 1079

furt (ah = 51%) but in Istanbul it is only slightly downsized
(ah = 6%). The value of ah and ac is reported in Table 6 for each
heat pump and climate.
For each case, the thermal energy delivered by the heat pump
and by the electric back-up (if needed), as well as the electrical
energy overall consumptions of the system have been calculated,
both for heating and cooling modes. The seasonal performance
indexes linked to the heating season (SCOP) and to the cooling sea-
son (SEER) have been evaluated:
EHP;h þ EBU;h
SCOP ¼
Eel;HP;h þ Eel;BU;h
EHP;c
SEER ¼
Eel;HP;c
where EHP and EBU are the thermal energy delivered by the heat
pump and by the back-up system, respectively, and Eel,HP and Eel,BU
are the electric energy consumptions of the heat pump and of the
back-up system, respectively.
The Annual Performance Factor (APF), able to describe the
annual energy performance of the system (by taking into account
both the heating and cooling behavior in a while) and defined as:
EHP;h þ EBU;h þ EHP;c
APF ¼
Eel;HP;h þ Eel;BU;h þ Eel;HP;c
is also calculated for all the cases investigated in this work.
In Table 7 the values assumed by the indexes of Eqs. (6)–(8) are
shown for each case. These performance values can be useful in
order to evidence the influence of the heat pump typology, climate
and unit sizing on the seasonal and annual energy performance of
the system.
In a site like Frankfurt characterized by a heating season more
severe than the cooling season (b = 0.43), the heat pump size is
generally chosen accordingly to the heating peak load. In this case,
units #4, #8 and #11 are the closest to the heating peak load, with
downsized full capacity at design outdoor temperature (#4:
ah = 26%, #8: ah = 28%, #11: ah = 26%). Since in Frankfurt the
the smallest units, i.e. models #1 (ah = 8%), #5 (ah = 18%) and
#9 (ah = 18%) for ON-OFF HPs, MCHPs and IDHPs, respectively.
In Lisbon the cooling season is more severe than the hot season
and the heat pump selection is based on the cooling peak load.
Hence, only the smallest units (#1: ac = 22%, #5: ac = 19%, #9:
ac = 18%) have been considered in simulations. The HVAC system
in Lisbon gives the best annual performance among the considered
sites.
In order to outline general rules for the optimal sizing of HVAC
systems based on reversible heat pumps, in Fig. 3 the SCOP trend
ð6Þ calculated with different heat pumps is represented as a function
of ah.
For downsized heat pumps (negative ah), the SCOP decreases
rapidly because it is negatively influenced by the electric heaters
ð7Þ
used as back-up system. When the heat pump capacity is tuned
on the building thermal needs (ah between 20% and 20%), the
SCOP reaches its maximum value for ON-OFF HPs; this value is gen-
erally lower than those obtained with MCHPs and IDHPs, due to
their enhanced capability to reduce the number of on-off cycles
by modulating the thermal power delivered. By adopting MCHPs
and IDHPs it is possible to obtain higher SCOP values even for
ah > 20% with respect to ON-OFF HPs.
In Fig. 4 the SEER trend is represented as a function of ac; results
point out that, by scaling from larger to smaller units, the SEER
ð8Þ increases because the chillers reduce their need of on-off cycles.
The maximum SEER is obtained for values of ac close to zero; the
values of SEER for ON-OFF HPs are lower than those obtained for
IDHPs and MCHPs; in addition, Fig. 4 highlights that the influence
of ac on the value of SEER is larger for ON-OFF HPs with respect to
IDHPs and MCHPs due to the enhanced modulating capability of
the latter two heat pump typologies.
3,4
3,2
3,0
2,8
SCOP

cooling peak load is lower than the heating one, units #4, #8 and
#11 result largely oversized with respect to the cooling peak load, 2,6
up to 153% for unit #8 (see Table 6). Therefore, during the cooling
season the chillers have to deliver only a small fraction of their full ON-OFF HP
2,4
cooling capacity by using their modulation capability. MCHP
In Istanbul heating and cooling peak loads are similar (b = 0.73), 2,2 IDHP
but the cold season is still more severe than the hot season; conse-
quently, heat pump sizing is usually based on the heating design 2,0
-80% -60% -40% -20% 0% 20% 40% 60%
load. Owing to a lower value of heating peak load, units #4, #8
h
and #11 have not been considered for Istanbul simulations due
to their large cooling and heating capacity. The results point out Fig. 3. SCOP as a function of the heating heat pump downsizing/oversizing
that the best annual performance is achievable by considering parameter (ah).

Table 7
Seasonal and annual performance indexes as functions of the considered heat pump and of the location.

Location Frankfurt Istanbul Lisbon

# HP unit SCOP SEER APF SCOP SEER APF SCOP SEER APF
1 (ON-OFF) 2.49 3.09 2.54 2.77 3.27 2.95 2.72 3.13 2.94
2 (ON-OFF) 2.57 2.91 2.60 2.69 3.14 2.85 – – –
3 (ON-OFF) 2.62 2.76 2.63 2.65 3.01 2.78 – – –
4 (ON-OFF) 2.59 2.63 2.60 – – – – – –
5 (MCHP) 2.53 4.29 2.64 3.15 4.34 3.52 3.31 4.22 3.79
6 (MCHP) 2.71 4.08 2.80 3.19 4.17 3.51 – – –
7 (MCHP) 2.89 3.85 2.96 3.16 4.05 3.45 – – –
8 (MCHP) 2.93 3.61 2.99 – – – – – –
9 (IDHP) 2.29 4.35 2.40 2.93 4.33 3.35 3.23 4.25 3.75
10 (IDHP) 2.59 4.07 2.69 2.89 4.13 3.27 – – –
11 (IDHP) 2.66 4.02 2.75 – – – – – –
1080 M. Dongellini et al. / Applied Thermal Engineering 114 (2017) 1073–1081

4,5 fact, the enhanced modulating capability of these devices strongly

cuts down the difference between heat pump capacity and build-
ing thermal loads, avoiding on-off cycles penalization.
4,0
Furthermore, Fig. 5 highlights the influence of unbalanced
building loads on the annual system performance: APF is lower
3,5 in correspondence of lower values of b. In fact, a thermal load
SEER

strongly unbalanced during winter (i.e. a low b value) leads to a

higher use of electric back-up heaters.
3,0
These results suggest, for an optimal exploitation of the energy
ON-OFF HP
saving potential of heat pumps, a building design able to obtain
2,5 MCHP balanced heating and cooling loads (b ffi 1), so the heat pump can
IDHP cover both heating and cooling loads by reducing the use of
2,0 back-up systems. MCHPs and IDHPs must be preferred to ON-OFF
0% 20% 40% 60% 80% 100% 120% 140% 160% 180% HPs due to their enhanced modulating capability, which reduces
c the use of on-off cycles during the year; nevertheless, their use
in buildings with low b (b < 0.5) markedly reduces the energy
Fig. 4. SEER as a function of the cooling heat pump downsizing/oversizing
savings.
parameter (ac).
As a rule for HVAC designers, the best energy performance of
HVAC systems based on a single reversible electric heat pump is
reached if a multi-compressor or an inverter-driven heat pump is
4,0 selected with a slight oversizing with respect to the annual build-
ON-OFF HP
ing thermal loads, according to the overall parameter atot intro-
duced in this paper.
3,6 MCHP =1.16

IDHP

3,2
4. Conclusions
APF

2,8 In this work the behavior of a HVAC system based on a reversi-

ble air-to-water heat pump coupled to a typical office building dur-
2,4 ing both heating and cooling season has been analyzed. Electric
=0.79 resistances have been considered as back-up system during winter.
=0.43 The main seasonal and annual energy performance indexes have
2,0 been calculated by means of a model based on the bin-method
-60% -40% -20% 0% 20% 40% 60%
and the building energy signature. Mono-compressor on-off,
tot
multi-compressor and inverter driven heat pumps have been con-
Fig. 5. APF as a function of the overall heat pump downsizing/oversizing parameter sidered to take into account the different capability of thermal
(atot). power modulation of these devices.
The obtained results evidence that downsizing/oversizing
degree of the considered heat pump with respect to the thermal
loads of the building is a crucial parameter that influences the
In order to analyze the annual energy performance of the HVAC energy performance of the HVAC system. It has been demonstrated
system, an overall indicator of the downsizing/oversizing degree of that during winter, when the full thermal capacity delivered by the
the system (atot) has been defined: selected heat pump at design conditions ranges from 20% to +20%
of the building design thermal load, the SCOP value becomes max-
DDh ah þ DDc ac
atot ¼ ð9Þ imum, especially for buildings characterized by balanced loads
DDh þ DDc
between winter and summer. Inverter driven and multi-
compressor heat pumps are able to guarantee larger SCOP than
where ah and ac are weighted on the degree days linked to the heat-
those obtainable by adopting on-off heat pumps (+20%). The lar-
ing (DDh) and cooling (DDc) seasons.
gest values of SEER are reached if the heat pump oversizing is lower
In Fig. 5 APF is shown as a function of atot. Values of the unbal-
than 40% of the design building cooling load.
ance indicator b (see Eq. (2)) are also reported for each case. It is
In addition, it has been demonstrated that the Annual Perfor-
evident that the best annual performance is obtained by adopting
mance Factor (APF) of the system is a function of the unbalance
MCHPs or IDHPs slightly oversized (0% < atot < 40%) with respect to
indicator b and of the overall index of the downsizing/oversizing
building design energy needs.
degree of the system atot. The variation of APF with b and atot is
ON-OFF HPs are less influenced by the overall oversizing per-
strongly dependent on the kind of heat pump considered (on-off,
centage than the other heat pump typologies: from atot = 30% to
multi-compressor, inverter driven), which means that the rules
atot = 20–40%, APF increases of about 10% for on-off units and of
for the optimal sizing of a heat pump change with the typology
about 20% for other heat pump typologies. This result highlights
of the selected heat pump.
the influence of heat pump sizing on the annual energy perfor-
mance of the HVAC system.
Fig. 5 confirms that ON-OFF HPs are able to reach lower APF val-
ues (16%) with respect to MCHPs and IDHPs. Reported data high- Acknowledgements
lights also that the more accurate control of the thermal power
delivered by multi-compressor and inverter-driven heat pumps The authors acknowledge the support provided by the manu-
can guarantee larger enhancements on APF, especially with unbal- facturer Galletti Group Spa (Italy) which made the technical data
anced loads (b < 1 or b > 1) and a strong heat pump oversizing. In sheets of their electric heat pumps available for this analysis.
 M. Dongellini et al. / Applied Thermal Engineering 114 (2017) 1073–1081 1081

References [13] F. Busato, R.M. Lazzarin, M. Noro, Energy and economic analysis of different
heat pump systems for space heating, Int. J. Low-Carbon Technol. 7 (2012)
104–112.
[1] European Commission, Directive 2009/28/EC of the European Parliament and
[14] European standard EN 14825:2013, Air conditioners, liquid chilling packages
of the Council of 23 April 2009 on the Promotion of the use of energy from
and heat pumps, with electrically driven compressors, for space heating and
renewable sources, 2009.
cooling – Testing and rating at part load conditions and calculation of seasonal
[2] European Commission, Directive 2010/31/EU of the European Parliament and
performance, 2013.
of the Council of 19 May 2010 on the Energy Performance of Buildings, 2010.
[15] C. Naldi, M. Dongellini, G.L. Morini, E. Zanchini, Comparison between hourly
[3] European Commission, Directive 2012/27/EU of the European Parliament and
simulation and bin-method for the seasonal performance evaluation of electric
of the Council of 25 October 2012 on Energy Efficiency, amending Directives
air-source heat pumps for heating, in: Proceedings of Building Simulation
2009/125/EC and 2010/30/EU and repealing Directives 2004/8/EC and 2006/
Applications 2015–2nd IBPSA Italy, Bolzano, Italy, 2015.
32/EC, 2012.
[16] S.A. Klein et al., TRNSYS 17: A Transient System Simulation Program, Solar
[4] T. Nowak, P. Westring, Heat pump market development in Europe, in:
Energy Laboratory, University of Winsconsin, Madison, USA, 2010.
Proceedings of 7th EHPA European Heat Pump Forum, Berlin, Germany, 2014.
[17] Handbook, ASHRAE fundamentals, Non-residential cooling and heating load
[5] R. Chargui, H. Sammouda, Modeling of a residential house coupled with a dual
calculation procedures, American Society of Heating, Refrigerating and
source heat pump using TRNSYS software, Energy Convers. Manage. 81 (2014)
AirConditioning Engineers, Atlanta, USA, 2001.
384–399.
[18] Italian standard UNI TS 11300-1:2014, Energy performance of buildings – Part
[6] W. Wu, W. Shi, B. Wang, X. Li, Annual performance investigation and economic
1: Evaluation of energy need for space heating and cooling, 2014.
analysis of heating systems with a compression-assisted air source absorption
[19] J.H. Zhu, Y.Y. Sun, W. Wang, S.M. Deng, Y.J. Ge, L.T. Li, Developing a new
heat pump, Energy Convers. Manage. 98 (2015) 290–302.
frosting map to guide defrosting control for air-source heat pump units, Appl.
[7] S. In, K. Cho, B. Lim, C. Lee, Partial load performance test of residential heat
Therm. Eng. 90 (2015) 782–791.
pump system with low-GWP refrigerants, Appl. Therm. Eng. 85 (2015) 179–
[20] P. Vocale, G.L. Morini, M. Spiga, Influence of outdoor air conditions on the air
187.
source heat pumps performance, Energy Proc. 45 (2014) 653–662.
[8] W. Chen, A comparative study on the performance and environmental
[21] European standard ISO 13790:2008, Energy performance of buildings –
characteristics of R410A and R22 residential air conditioners, Appl. Therm.
Calculation of energy use for space heating and cooling, 2008.
Eng. 28 (2008) 1–7.
[22] M.F. Fels, PRISM: an introduction, Energy Build. 9 (1986) 5–18.
[9] L. Palmiter, J.-H. Kim, B. Larson, P.W. Francisco, E.A. Groll, J.E. Braun, Measured
[23] European standard UNI EN 15603, Energy performance of buildings — Overall
effect of airflow and refrigerant charge on the seasonal performance of an air-
energy use and definition of energy ratings. 2008.
source heat pump using R-410A, Energy Build. 43 (2011) 1802–1810.
[24] M. Dongellini, C. Naldi, G.L. Morini, Seasonal performance evaluation of
[10] L. Junghans, Evaluation of the economic and environmental feasibility of heat
electric air-to-water heat pump systems, Appl. Therm. Eng. 90 (2015) 1072–
pump systems in residential buildings, with varying qualities of the building
1081.
envelope, Renew. Energy 76 (2015) 699–705.
[25] H.I. Henderson, D. Parker, Y.J. Huang, Improving DOE-2’s RESYS routine: user
[11] E. Kinab, D. Marchio, P. Riviere, A. Zoughaib, Reversible heat pump model for
defined functions to provide more accurate part load energy use humidity
seasonal performance optimization, Energy Build. 42 (2010) 2269–2280.
predictions, in: 2000 ACEEE Summer Study Proceedings, 2000.
[12] F. Madonna, F. Bazzocchi, Annual performances of reversible air-to-water heat
pumps in small residential buildings, Energy Build. 65 (2013) 299–309.