Air-Water Heat pump study

João Belo
1

1. Instituto Superior Técnico; Technical University of Lisbon; Avenida Rovisco Pais, Nº1,
1049-001 Lisboa; Email Address: joao.belo@ist.utl.pt.
Abstract: The present work main objective was to create a mathematical model, using the
VBA™ language, for a Bosch Air-Water Heat Pump. The model analyses its behaviour on different
ambient conditions. The set of equations is solved with the Broyden method improved with the
Sherman-Morrison formula.
The model includes energy and heat transfer balances for the Evaporator and Condenser, to
simulate the heat exchanges and the work of the compressor and Evaporator Fan. The Compressor
performance is predicted through the system’s evaporating and condensing temperatures.
Different changes on the vapour compression cycle are analysed, phase separation and the
introduction of an internal heat exchanger to the system. The influence of the Fan’s power is
evaluated as well as modifications of the Evaporator’s geometry with the purpose of maximizing the
coefficient of performance (COP).
The conceived model is able to predict the real installation operation, producing identical
results from the case study presented. For equal input conditions the internal heat exchanger cycle
produces more heat and the higher COP.
The study revealed that the evaporator Fan is over dimensioned for obtaining the highest
COP. Geometry changes indicate a limited grow of COP with the heat transfer area and air flow rate.
Keywords: Heat pump, Performance coefficient, efficiency, phase separation, internal heat
exchanger

NOMENCLATURE
Au
ì
Global heat transfer coefficient [W/K] 
CC Cylinder displacement [m
3

c
p
Specific Heat 
C0P Coefficient of performance 
b
ì
Enthalpy at point i [J/Kg.K] 
m
uì¡
Air mass flux [Kg/s] 
m
c¡up,1.1

Refrigerant mass flux at the 
evaporator on case 1.1 [Kg/s] 
m
c¡up,1.2

Refrigerant mass flux at the 
evaporator on case 1.2 [Kg/s] 
m
¡c]
Refrigerant mass flux [Kg/s] 
m
wt¡
Water mass flux [Kg/s] 
N
]ìns
Number of fins 
N
L
Number  of longitudinal row of tubes 
N
1
Number of transversal row of tubes 
P
L
Longitudinal Pitch [m] 

GREEK SYMBOLS
[ Plate corrugation angle [º] 
AP
ìnst
Installation pressure loss 
e Heat Exchanger efficiency 
p
comp
Compressor isentropic efficiency 













P
1
Transversal Pitch [m] 
N
¡ot
Rotational speed [rpm] 
µ
cond
Condenser heat [W] 
µ
c¡up
Evaporator heat [W] 
I
cond
Condensation temperature [°] 
I
c¡up
Evaporation temperature [°] 
I
ì
Temperature at point I [°] 
t
]ìn
Fin thickness [m] 
I

uì¡
Volumetric flux of air [m
3
/s] 
w
comp
Compressor power [W] 
w
comp,ìdcuI
Compressor ideal power [W] 
w
comp,¡cuI
Compressor real power [W] 
w
]un
Fan power [W] 
w
pump
Water Pump power [W] 


p
]un
Fan efficiency 
p
¡oI
Compressor volumetric efficiency 
0
Fin corrugation angle/Medium 
temperature difference between two 
fluids 
u Specific volume 



1. Introduction
In a more industrialized world, where the energetic needs grow everyday, it becomes more
important to give an answer to the increasing energy needs, in an efficient and quicker way. For
human commodity, water heating represents the fourth biggest energy consumption in the
commercial sector (Hepbasli, et al., 2009). Many of the existing systems produce heat through electric
resistances or fossil fuel energy conversion with limited efficiency. To compensate this problem, the
use of a vapour compression cycle grew in its importance, because it is able to produce a greater
amount of heat, using the same electric power, with a lower cost and more efficiently.
The main objective of the work is the development of a computational model of an existing
heat pump, using Excel™. The model uses Broyden’s method improved with the Sherman-Morrison
formula (Kelly, 2003) to resolve the non-linear system of equations. Fluid properties were evaluated
using REFPROP, developed by NIST (National Institute of Standards and Technology, USA).
The models allow also the evaluation of different cycle configurations, phase separation, in
which two different options were studied, the introduction of a flash tank (feeding the saturated vapour
to the compressor after mixing with the vapour exiting the evaporator); and feeding the compressor
with saturated vapour and recirculating the saturated liquid on the Evaporator. The last configuration
used was the introduction of an internal heat exchanger (assuming an efficiency e = u.8), to increase
the vapour temperature, utilising the heat available on the saturated liquid of the condenser at a
higher temperature. The behaviour of the model was tested varying the compressor inlet temperature,
air flux and water flux. An evaluation of the cycle’s performance coefficient (COP) was carried out
considering a self-regulated fan, modelled by varying the fan’s power consumption, and making
changes on the evaporator geometry.
C0P =
µ
cond
w
comp
+w
]un
+w
pump

( 1 )
There has been some research on heat pumps. Guo et al. (2011), studied the design
optimization of an air-source heat pump water heater, working according to the Shanghai climate. The
study revealed that during winter the heating capacity decreases gradually with the increase of the
inlet water temperature, the opposite occurs on summer. This happens until the enthalpy difference
between the inlet and outlet of the condenser dominates the heat transfer.
Figure 1 - Heating capacity vs T
euap

(Guo-Yuan et al. 2008)
Figure 2 - Power vs. T
euap

(Guo-Yuan et al. 2008)
Guo-Yuan et al. (2008) studied a heat pump with a flash tank to compare its performance with
a “sub-cooled” system. The study revealed that the flash tank increases more the heating capacity
when compared with the internal heat exchanger (Figure 1). The compressor power changes slightly
due to the change of the pressure ratio (Figure 2).
2. Model
2.1 Compressor
The compressor is characterised by the isentropic and volumetric efficiencies defined as:
p
comp
=
w

comp,ìdcuI
w

comp,¡cuI
( 2 )
p
¡oI
=
m
¡c]
× u
CC × (N
¡ot
¡6u)
( 3 )
The values were computed from the catalogue data provided by Bosch and correlated as
polynomials as a function of the evaporating and condensing temperatures.
2.2 Condenser
The condenser is a double-wall plate heat exchanger with 14 chevron angled plates of AISI
316L. The calculation of the convection coefficient of the refrigerant fluid is made by the correlation
provided by Longo (2010), which uses a Nusselt correlation for film condensation for vertical surfaces.
Based on the mass flux it is observed that heat transfer is controlled by film condensation and the
effects of forced convection are small. For water, the correlation proposed by Han et al. (2003) is
used, which is a function of the Reynolds and Prandtl numbers and the plate corrugation angle [.
The water pump has 3 operating positions, changed by a switch.
2.3 Evaporator
The evaporator is a finned-tube heat exchanger, with herringbone fins, to transfer heat from
air to the refrigerant. The vapour leaves the evaporator with an over heat of 5°. The correlations
selected for the convection coefficients are, Wang et al. (1999), for air and Hewitt et al. (1994) for the
refrigerant. The average properties of the refrigerant were evaluated by considering a medium quality
inside the evaporator.
The fan works at a nominal power of 100W, but in order to change the rotational speed, it is
used a capacitive condenser on the electrical power source, working presently at powers ~50W, using
a frequency controller. For both situations the global efficiency was characterised and considered in
the simulations.
3. Equation system
After modelling each component of the cycle, the model was formulated with the following set
of equations to be solved by Broyden method.
Fon
F

a|r
× _
ΔP
ìnst
p
]un
]
w

]un
-1 = u
( 4 )
E:oporotor
`
1
1
1
1

m
¡c]
× (b
1
-b
4
)
m
uì¡
× c
p
cir
× (I
uì¡,ìn
-T
a|r,uut
)
-1 = u
( 5 )
m
¡c]
× (b
1
-b
4
)
(Au)
c¡up
× 0
c¡up
× |I
uì¡,ìn
-T
euap
|
-1 = u ( 6 )
ConJcnscr
`
1
1
1
1

m
¡c]
× (b
2
-b
3
)
m
wt¡
× (h
wtr,uut
-b
wt¡,ìn
)
-1 = u ( 7 )
m
¡c]
× (b
2
-b
3
)
(Au)
cond
× 0
cond
× |T
cund
-I
wt¡,ìn
|
-1 = u ( 8 )
The first equation is the energy balance to the fan, where p
]un
is the fan efficiency allowing for
the calculation of the air flow based on the power supply to the fan. Equations (4) and (5) are the
energy balances to the Evaporator. From these equations the outlet air temperature and the
evaporation temperatures are calculated. Equations (6) and (7) are the energy balances of the
condenser, dependent on the water outlet and condensing temperatures. The equations are non-
dimensional, to allow the Broyden method to work with non-dimensional and normalized residues in
order to reduce any errors from the minimization of the function residues.
The study of the different configurations implies certain alterations to the equation system. For
the phase separation, the main changes are the different refrigerant mass flux that enters the
evaporator. With an energy balance to the flash tank, it is possible to account the vapour fraction
extracted. Using the mass flux crossing the compressor (m
¡c]
) as reference, the mass flux crossing
the evaporator in each case is,
Case 1.1 m
c¡up,1.1
= _
b
4
-b
Iìq,sut
b
]g
_m
¡c]
( 9 )
Case 1.2 m
c¡up,1.2
= _
b
4
-b
¡up,sut
b
Iìq,sut
-b
6
_m
¡c]
( 10 )
On equation (10) the quality at point 6 (exit of the evaporator) was considered as x = 1,
saturated vapour.
Introducing a new heat exchanger in the cycle will introduce two new equations to the
previous set, to compute the new variables, the vapour temperature at the compressor inlet and the
sub-cooled liquid temperature at the exit of the new heat exchanger. The new equations are the
energy and thermal balance to the internal heat exchanger.
Eq. A
m
¡c]
× (b
1
-b
A
)
m
¡c]
× (b
3
-h
B
)
-1 = u ( 11 )
Eq. B
m
¡c]
× (h
1
-b
A
)
(Au)
ìnt
× 0
ìnt
× |I
cond
-I
A
|
-1 = u ( 12 )
The evaluation of 0
ìnt
is made by a Newton-Raphson method for the e -NIu method for non-
mixed cross flow heat exchangers with e = u.8
In order to cope with the changing conditions in the environment, the fan must be able to
change the rotational velocity to optimize the COP. This test was conducted by changing the power of
the fan in the simple cycle from 5W to 100W to find where the optimal operational power is. It should
be mentioned that the fan considered was always the same and therefore the best efficiency was
achieved for the full capacity.
The last study made, was the optimization of the evaporator geometry. Changes to the
geometric parameters were made to define what are the possible changes to optimize the COP. The
changed parameters were analysed: Number of parallel streams; number of transversal and
longitudinal lines (N
L
, N
C
); transversal and longitudinal pitches (P
1
, P
L
); number of fins (N
]ìns
); and fin
thickness (t
]ìn
).
4. Results
The validation of the computational model is performed by comparing the results of the case
study, presented by Bosch, with the model results at catalogue conditions. The data from the
compressor is available for the reference conditions considering inlet temperature of 20°C, while test
data was obtained with inlet air temperature of 7°C and inlet water temperature of 40°C. The inlet
compressor temperature in this later case is a result from the calculation.
Table 1 - Comparison of real model with theoretical model
 
Nominal 
data 
T
cump,|n
  2û°  AT
SH
5° 
Results  Difference Results  Difference
m
re¡
kg¡s 6.39E‐03  6.81E‐03  6.54%  -  - 
Q
euap
kw 0.832  0.883  6.14%  -  - 
Q
cund
kw 1.350  1.397  3.46%  -  - 
W
cump
w 410  418  1.96%  -  - 
COP AJ. 3.29  3.34  1.47%  -  - 
T
a|r
- T
euap

⁰C
9.0  -  -  8.3  ‐7.98% 
T
cund
-T
wtr

⁰C
10.0  -  -  10.4  4.10% 
The computational model can reproduce the compressor catalogue with deviations lower than
~7%. Regarding the heat transfer, the resulting temperature difference after applying the model under
the real operating conditions (5ºC over heat), produced a result with a difference of ~8% in the
evaporator and ~4% in the condenser, compared to the results obtained using the provided data.
Figure 3 and 4 present the heating capacity and COP according to the type of alteration to the
base cycle. The result was a decrease of both values that were not anticipated. This was due to the
calculated reduction in convection heat transfer coefficient for the refrigerant inside the evaporator.
The refrigerant evaporation is possibly dominated by nucleate boiling instead of forced convection, as
modelled and the use of lower vapour fractions may increase the coefficient while with the present
approach it was reduced. Hence this result requires further attention. The model responds well to the
heat transfer characteristics in the evaporator, that is a decrease of the coefficient leads to a larger
temperature difference, lowering the evaporation temperature and therefore decreases the
performance of the compressor.

4.1 Parametric tests
To evaluate the effect of operating conditions three tests were conducted: the increase of the
inlet compressor temperature and the variation of the volumetric flux of water. The first test revealed
an increase of µ
cond
with I
comp,ìn
while µ
c¡up
decreases and the compressor power slightly increases.
These effects cause an increase of the COP in every type of cycle. Therefore it is not interesting to
isolate the line from the evaporator to the compressor and if some heat can be recovered from the
compressor cooling it will have a positive effect.
The change on the water flux shows a first moment where there is a decrease of w
comp
and
an increase of µ
cond
and µ
c¡up
. At approximately ~3 L/min the curves resume almost constant. The
COP increases, but while for water the increase levels of for values above 3 l/min that is the value
used, for the air flow the increase does not level off. It should be mentioned that the value of COP in
this test is only from the compressor so the effect of energy consumption in the fan is not considered.
Figure 4 – Change of COP with type of cycle
Figure 3 – Change of Q

cund
with type of cycle

Figure 5 - Power vs. F

wtr

4.2 Fan optimization
Figure 6 represents the variation of COP with the power of the fan. In the blue line the fan
efficiency is variable according to the model created for this component (green line). In red, the
efficiency is kept constant and equal to 0.3. It is seen an inflection on both curves, revealing an
operation point where the COP is maximum, which is (w
]un
; C0P) = (9 ; 2.89) for variable efficiency
and (w
]un
; C0P) = (1S ; S.1) for constant efficiency. In the other line the power consumption of the
pump was also included.

4.3 Evaporator’s Geometry
The tests revealed a limited growth of COP with the increase of the heat transfer area. The
COP increases until a maximum point, and after it the transferred heat remains constant or starts do
decrease (Figure 8). Compared with the existing system some improvement can be obtained by
increasing one row of tubes in the longitudinal direction.
250
450
650
850
1050
1250
1450
1 3 5
W
c
o
m
p
Q
e
v
a
p
Q
c
o
n
d
[
W
]
V
wtr
[L/min]
Qcond,base
Qcond,11
Qcond,12
Qcond,hx
Qevap,base
Qevap,11
Qevap,12
Qevap,int
Wcomp,bas
e
Wcomp,11
Wcomp,12
0%
5%
10%
15%
20%
25%
30%
35%
2,7
2,8
2,9
3
3,1
3,2
5 15 25 35 45 55 65 75 85 95
η
f
a
n
C
O
P
W
fan
[W]
Variable eff. Constant eff. fan_eff
Figure 6 - COP vs. Fan power

Another point to take into account is the variation of the convection coefficients. A decrease of
the flow area leads to an increase of the Reynolds number. With the increase of Re, the convection
coefficients computed from the heat transfer correlations will increase and so the heat transferred
between the air and refrigerant. In general an increase in the heat transfer area leads to some
improvement, it should be considered the costs to build/change the heat exchanger.
5. Conclusions
The implementation of a computational model to calculate the performance of a heat pump
was achieved successfully. The results are well fitted with the results from the real installation.
The introduction of an internal heat exchanger led to the larger heat output and COP
compared to the other configurations studied. The phase separation configurations produced results
similar to the base cycle.
There is a benefit of increasing the compressor inlet temperature. Due to the increase of the
COP with this temperature, it is suggested that further studies are taken to implement extra heating in
the compressor admission line, e.g. using the heat loss from the compressor.
Considering the work in the fan the COP presents a peak. The selected fan is oversized and
the simulations considered included a large decrease of its efficiency when reducing the capacity. An
appropriate fan should be selected for the reduced air flow to improve the global COP.
The result of tests changing the water flow rate and the geometry of the evaporator showed
that the potential to increase the performance of the heat pump is limited, although the best option
would be to add a row of tubes in the longitudinal direction. Further evaluation of this modification
should be done considering correlations that account for the number of tubes in the flow direction.

2,78
2,79
2,8
2,81
2,82
2,83
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
C
O
P
N
L
N
T
Figure 8 - COP vs. Number of lines and columns
References
1. Guo, J. J., Wu, J. Y., Wang, R. Z., Li, S. 2011. Esperimental Research and Operation
Optimization of an Air-source Heat Pump Water Heater. Applied Energy. 2011, Vol. 88, pp.
4128-4138.
2. Guo-Yuan, Ma e Hui-Xia, Zhao. 2008. Experimental Study of a Heat Pump System with
Flash-Tank Coupled with Scroll Compressor. Energy and Buildings. 40, 2008, pp. 697-701.
3. Han, Dong-Hyouck, Lee, Kyu-Jung and Kim, Yoon-Ho. 2003. The Characteristics of
Condensation in Brazed Plate Heat Exchangers with Different Chevron Angles., Journal of
the Korean Physical Society, Vol. 43, No. 1, July 2003, pp. 66-73.
4. Hepbasli, Arif and Kalinci, Yildiz. 2009. A review of heat pump water heating systems.
Renewable ans Sustainable Energy Reviews. 2009, Vol. 13, pp. 1211-1229.
5. Hewitt, G. F., Shires, G. L., Bott, T. R. 1994. Process Heat Transfer. s.l. : Begel House,
1994.
6. Huber, David and Walter, Heimo. 2010. Forced convection heat transfer in the transition
region between laminar and turbulent flow for a vertical circular tube. Lastest Trends on
Theoretical and Applied Mechanics - Fluid Mechanics and Heat & Mass Transfer. 2010.
7. Kelly, C. T. 2003. Solving Nonlinear Equations with Newton’s Method, Fundamentals of
Algorithms. 2003.
8. Longo, Giovanni A. 2010. Heat transfer and pressure drop during HFC refrigerant saturated
vapour condensation inside a brazed plate heat exchanger. International Journal of Heat and
Mass Transfer 2010, Vol. 53, pp. 1079-1087.
9. Wang, C. C., Du, Y. J., Chang, Y. J., Tao,W. H. 1999. Airside Performance of Herringbone
Fin-and-Tube Heat Exchangers in Wet Conditions. The Canadian Journal of Chemical
Engineering. 1999, Vol. 77(6), pp. 1225-1230.