A NOVEL DESIGN TOOL FOR HEAT PUMP SYSTEMS

Martin Forsén, M. Sc.,Per Lundqvist, Docent

Department of Energy Technology, Division of Applied Thermodynamics and Refrigeration,
Royal Institute of Technology, KTH, Stockholm, Sweden.

ABSTRACT

Designing a competitive heat pump system for heating purposes involves thorough economic and
technical considerations. In contrast to other conventional heating systems the performance of a heat
pump system is strongly influenced by the temperature levels in the heat distribution system and the
temperature and the characteristics of the heat source. The use of deep boreholes or horizontal ground
coils as heat source that offer a relatively high and stable temperature level increases the performance in
cold climates but also the investment cost. The high investment cost associated with a heat pump can only
be justified if the running cost and emissions of green house gases are much less than that of a
conventional heating system. This raises the need to evaluate the seasonal performance factor as well as
the design of the heat source. There are a number of different tools and software available on the market
today. Most of which are either looking into the design of the heat pump unit or the heat source. This
paper describes a novel design tool developed in co-operation with the actors on the Swedish heat pump
market. The tool has been developed in order to perform system performance analysis as well as the
design of vertical boreholes.

Key Words: simulation, borehole design, domestic heat pump

1 INTRODUCTION

Energy utilization in the built environment is one of the most important aspects that have to be
addressed in the near future. In order to reach the targets of the Kyoto-protocol, the energy utilization in
the built environment has to go through a transition. Up to now most of our space conditioning systems
contribute to the global warming. Environmentally benign heating systems have to be introduced on a
large scale in order to reduce the emissions of green house gases. The use of small-scale bio-fuel
furnaces, heat pumps and an extensive use of district heating based on bio fuel have to be implemented.
None of these techniques will on its own be able to fulfil the transition towards sustainable energy supply.
All techniques will complement each other and find their markets where the conditions are most
beneficial.

Domestic heat pumps are one of the most efficient ways to provide space heating and preparation of
sanitary hot water. Even though technical know-how on the heat pumping technology is well proven, it
has not yet reached public acceptance worldwide. In Europe, a sustainable market has only been
established in small countries like Sweden, Switzerland and parts of Austria. The market for domestic
heat pumps in Sweden has, during the last decade, gone through an enormous development (see Fig. 1).
The total sales of domestic heat pumps reached over 66 thousand units 2004 (SVEP 2005). On top of that
somewhere in between 40 000-50 000 reversible air/air heat pumps, of which only a minor part is
included in the statistics compiled by the SVEP, were sold 2004. All together more than 100 thousand
heat pumps were thus sold in Sweden 2004, a country consisting of approximately 1.6 million single-
family houses. Due to the escalating price of oil and electricity in conjunction with the increase on energy
related taxes the market for heat pumps continuous to grow at a high pace.

1
 70 000 66 154

60 000

48 806
50 000

39 602
40 000 36 568

30 000
24 253
20 296
17 023 18 699
20 000
12 431
10 202 9 028
10 000

0
1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004

Fig. 1 Heat pump market development in Sweden 1994-2004 (Swedish Heat Pump Association
2005)

The high investment cost affiliated with ground source heat pumps (GSHP) can only be justified if
the operating costs prove to be considerable lower than traditional heating systems. This raises the
demand for an accurate tool for design and evaluation of different heat pump systems.

2 SIMULATION MODELS FOR HEAT PUMP SYSTEMS

There are a number of different tools available for energy performance calculation on the market.
The fact that each tool is designed for its own objective and target group, lead to significant differences in
terms of the structure for calculation, input, output, user interface and accuracy. A close examination of
all available products three characteristic groups of tools can be distinguished. The main divergence
between the groups are how the calculations are structured. Energy performance calculations are
structured in static calculations, quasi-static calculations or dynamic calculations.

2.1 Static calculations

All available sales support software, provided by the Swedish manufacturers of heat pumps, stems
from this category. The static calculation models are based on “bin-data” i.e. climate data is described in a
table of the occurrence of outdoor temperatures. All relevant energy transfer processes are then calculated
for each outdoor temperature. The result is then multiplied by the annual duration of the specific outdoor
temperature. Adding all these results together forms the annual performance. Steady-state conditions are
assumed for all conditions. The obvious advantage offered by this technique is the simplicity, which leads
to fast computations and relatively low demands on climate data. “Bin” wheater data is available at a
relatively low cost and there are a number of reasonable accurate synthetic climate generators available.
The inherent disadvantages with static calculations are that no considerations for seasonal variations can
be taken in to account in the calculation, thus the characteristics of a ground heat source is not adequately
treated. Furthermore variations in electricity rates will not be reflected in the calculations.

2
2.2 Quasi-static calculations

In quasi-static simulations all calculations are performed in chronological order. The year is divided
in an appropriate number of time steps. The choice of time step is based on the application, available
climate data and calculation capacity. Typical time steps are 1 hour, 1 day, 1 week, 1 month. The shorter
time step chosen the higher accuracy may be acquired. For all calculations steady-state conditions are
assumed for each time step. The greatest benefits associated with a quasi-static calculation model are that
seasonal effects will be reflected in the calculations as well as the opportunity to make comparisons
between the calculations and monitored results in existing installations. The most commonly used time
step is one hour. The advantages with a quasi-static calculation model appear at the price of more detailed
climate data. The minimum time step is dependent on the time constant for the process being examined.
There is a limit where the condition can no longer be considered to be at a steady state. In those cases a
dynamic calculation method have to be applied.

2.3 Dynamic calculations

If the time constant for any of the described components is longer than the desired time step in a
quasi-static calculation a dynamic method has to be applied. A simulation of the dynamic interaction
between a set of components raises the demand for a meticulous description of all components. The
amount of required input increase the burden on the user and limits the availability to a smaller group.
The obvious advantage is that dynamic processes may be revealed in the simulation. The choice of
calculation method will be decided on the overall aim of the simulation and competence level of the
anticipated users. The general accuracy and required calculation capacity is depicted in Fig. 2.

Dynamic
Computation Time

Quasi-Static
Short Time Step

Quasi-Static
Long Time Step

Static

Accuracy
Fig. 2 Characteristics of different calculation methods

2.4 Choice of calculation method

In the initial phase of all software development the overall aim has to be defined. The aim and the
identified users of the final product sets the boundaries for the user interface and serve as a guide in
choosing the most appropriate calculation structure. Our objective, in this work, is to develop a design
tool to be used in the contact between the contractor and the consumer. The results of the calculations
shall, in an adequate way reflect realistic operating conditions and functionality of the simulated
installation. The result of the simulation shall provide the consumer with information on the reduction of
operating costs as well as required design of a borehole, in the case of a GSHP application. Based on the
requirements for the simulation and taking into account for the identified group of users a quasi-static
calculation structure has been chosen. The time step of 1 hour has been found to be most appropriate for
this application.

3
3 THE SIMULATION MODELS

As the tool is designed to evaluate the potential for cost reduction by the use of heat pump
technology in an existing building, the calculations are based on knowledge of present energy utilisation.
A further limitation is that ducted air systems for heat distribution are not included in the models due to
negligible existence on the Swedish market. The general simulation flow is depicted in Fig. 3 and in short
described as follows. Based on the annual heat demand and climate data for the chosen location, an
energy profile is established. The energy profile reveals the heating demand for each occurring outdoor
temperature. The information in the climate database and energy profile is then used, in order to establish
the heat demand, on an hourly basis. Given the design temperatures for the hydronic heat distribution
system, the temperature levels (supply- and return temperature) may be acquired for all prevailing
outdoor temperatures. This step is omitted in the case of an air-air heat pump. In the case of a GSHP a
desired mean temperature for the incoming brine is set. Together with the information of the hourly heat
demand, all governing operating conditions for the heat pump are set and all energy transfers may be
acquired.

Climate Design parameters Laboratory

database hydronic system measurements

Annual heat Heat demand Operating conditions

demand Energy profile hourly basis hourly basis
Temp levels heat Energy
generator distribution system calculator

Design parameters
ground loop

Fig. 3 Simulation flow

The following sections describe the procedures to obtain the energy profile, temperature levels for
the simulated heat distribution system, design of vertical ground loop and modelling of the heat pump
performance.

3.1 Establishing the energy profile

The procedure to establish the energy profile is based on a simplification that all the heat losses from
a building can be lumped together in one general heat loss factor (HLF). The HLF is determined from the
annual heat demand and the number of degree hours at the location.

& = HLF (t &

Q indoor − t outdoor ) − Q intgain

Q A = HLF ⋅ degh
(1a,b)

The concept of degree hours is widely used in many countries. The definition of how to calculate the
number of degree hours may vary (Forsén 2000, SMHI), but the general idea remains the same. In this

4
context the degree hours are calculated for each day when the mean temperature is below 11 °C. The
internal heat gain (tintgain) is set to 3 K.

8760
degh = ∑ (t
i =1
indoor − t intgain − t outdoor (i )) (1c)

3.2 Heat distribution system

The general methodology to determine the supply- and return water- temperature for a radiator system has
been described in the literature (Peterson, Nilsson 1988). The temperature levels of the hydronic heat
distribution system are derived from an overall heat balance. At steady state condition the heat supplied to
the radiator must equal the heat supplied to the room and the heat loss through the climate shield of the
house. The heat supplied to the hydronic distribution system may be decided by equation 2a.

Q& = m& cp ( t supply − t return )

(2a)
Q& = c ( t
1 supply−t )
return

Heat loss from the climate shield may be determined by the following expression.

Q& = c 2 ( t indoor − t outdoor ) (2b)

The overall heat transfer coefficient for the radiator can not be treated as a constant since it depends
on the logaritmic mean temperature difference of the radiator. The relation for the overall heat transfer
coefficient can be written as follows.

n
⎛ ∆t ⎞
U = U DOT ⎜⎜ ⎟⎟
⎝ ∆t DOT ⎠
t supply − t return (2b,c)
∆t =
⎛ t supply − t indoor ⎞
ln⎜⎜ ⎟⎟
⎝ t return − t indoor ⎠

Index DOT denotes values at design outdoor temperatur and index n denotes the radiator exponent.
The overall heat transfer coefficient at DOT ( U DOT ) and the exponent n are given in catalogs from
radiator manufacturers. The following relation can be written for the heat transfer from the radiator.

n
⎛ 1 ⎞
Q& = U DOT A⎜⎜ ⎟ ⋅ (∆t )1 + n (3a)
⎟
⎝ ∆t DOT ⎠

Which may be rewritten in

Q& = c3 (∆t )1+ n (3b)

5
 By knowledge of the heat power demand ( Q& ), tsupply, treturn, toutdoor, tindoor, and the radiator exponent all
constants c1, c2, and c3 may be determined, thus the supply and return temperatures may be found for all
occurring outdoor temperatures.

3.3 Design of vertical ground loop

There are numerous models and for the design of vertical boreholes described in the literature. An
overview of existing software was performed by Hellstöm and Sanner (2001). The models are generally
based on the line source theory or the cylindrical heat source method derived by Carslaw and Jaeger
(1947). The cylindrical heat source method was later on expressed in a more convenient manner for this
application by Ingerzoll et al (1954). In common for most of the models described in the literature as well
as commercially available design tools are that they are giving a detailed description of all heat transfer
processes that take place in the ground loop but only a rudimentary description of the heat pump. This is a
drawback since the performance of the heat pump is interlaced to the temperature levels of the
heating/cooling distribution system. The general simulation tool TRNSYS however, offer the opportunity
to incorporate detailed ground source simulation into a complete building system. The user interface
within TRNSYS is developed for use of a vast range of applications and consequently not all that easy to
deal with. Bernier and Randriamiarinjatovo (2004) have implemented a complete set of models for
system simulation in the more user-friendly environment offered by EES (2000). The system simulation
presented by Bernier and Randriamiarinjatovo incorporates simulation of the building heat/cooling load
as well as the heat pump and the ground loop. In this work, much inspired by the work of Bernier (2000),
the cylindrical heat source method is applied. Variations in the thermal load of the borehole are treated by
superposition.

3.3.1 General assumptions

In order to calculate the temperature at the wall of the borehole (TV), the undisturbed ground
temperature (TB) for the location has to be established. The undisturbed ground temperature is the natural
temperature of the ground unaffected by any ground heat exchanger, at a depth where seasonal variations
are not observed. The estimation of the undisturbed ground temperature is based on the annual mean
temperature of the location. This temperature is however an underestimation for locations that for longer
periods are covered by snow. The snow cover serves as insulation for the ground, which results in that the
undisturbed ground temperature will be somewhat higher than the annual mean temperature for the
location. A correction of 1.5 K/100 days of coherent snow cover is recommended in Swedish national
guidelines (VVS-Handboken 1963). Furthermore a correction of 1,6 K/100 m for the natural vertical
temperature gradient is assumed as an average of Swedish bedrock.

TB
Tv

R
Q

Fig. 4 Geometry of vertical borehole

6
3.3.2 Heat transfer model

The borehole is considered as an infinitively long cylinder embedded in infinite bedrock with known
heat transfer characteristics. All heat transfer is assumed to be perpendicular to the borehole. An
analytical solution for the temperature decrement around the cylinder has been derived by Ingerzoll et al
(1954).

∞
Q& e−β Fo(t ) − 1 ⎡ r r ⎤ ∂β
2

TB − TV(t) =
Lλπ 2 ∫
0
⎢ j0 ( β )Y1 (β ) − j1 (β )Y0 ( β )⎥ 2
j1 (β ) + Y1 (β ) ⎣ R
2 2
R ⎦β
(4)

For simplicity reasons part of the expression above is expressed in terms of a function of the Fourier
number

∞
e − β Fo ( t ) − 1 ⎡ r ⎤ ∂β
2

j0 ( β )Y1 ( β ) − j1 ( β )Y0 ( β )⎥ 2 = f (Fo )

1 r
π2 ∫
0
2 2 ⎢
j1 ( β ) + Y1 ( β ) ⎣ R R ⎦β
(5)

The expression, which is valid for constant heat transfer rate, may be used for variable heat transfer
rate by the use of superposition. The method of superposition or load aggregation is in detail described in
other publications (Kavanaugh and Rafferty 1997) and (Bernier 2001). Using the simplified expression
above and applying the superposition method of three different heat transfer rates result in the following
expression.

[ f (Fo ) − f (Fo )]+ Lλ [ f (Fo ) − f (Fo )]+ Lλ f (Fo )

⋅ ⋅ ⋅

TB − TV ( t3 ) =
Q1 Q2 Q3
t3 − 0 t3 −t1 t3 −t1 t3 − t 2 t 3 −t 2 (6)
Lλ

As the ground loop is designed for many years of continuous operation the simulation is performed
for 11 years of operation. The mean cooling load over 1 year is used for the first 10 years of operation and
thereafter the mean cooling load for each day of the year is used. In this way the number of terms in
equation 6 is reduced to 366.

3.3.3 Determination of borehole resistance and brine temperature

The method previously described will establish the temperature at the borehole wall. The following
text will describe the method used to obtain the thermal resistance of the borehole and the temperature of
the brine entering the evaporator. Configuration of the borehole heat exchanger (BHE) is in this study
limited to single U-pipe surrounded by ground water. This is the prevalent configuration on the Swedish
market. Back-filling of boreholes is restricted to a very small number of installations in Sweden. Figure 4
depicts a cross-section of the idealised borehole.

7
 TV

TBrine

Fig. 5 Cross-section of borehole

The temperature difference in between the borehole wall (TV) and the temperature of the brine
(TBrine) is obtained by the use of electric resistance equivalence.

RS

RH2O

TV TBrine
RS

Fig. 6 Borehole resistance

q&
TV − TBrine =
Rb
Rs
Rb = + R H 2O (7 a , b )
2

The thermal resistance of the pipe (Rs) is determined by:

ln( D y / Di ) 1
Rs = + (8)
2πλ pipe πDiα i

Heat transfer coefficient on the inside of the pipe (αi) is determined by:

8
 Re ≥ 2300
Nu = 0.023 Re 0.8 Pr 0.4
Re < 2300
Nu = 1.86(Re Pr d i / L)1 / 3 (9a, b)

The thermal resistance induced by the material surrounding the pipes (RH2O) in the borehole is
calculated by use of a shape factor (Sb). The shape factor suggested by Remund (1999) is used in this
project.

1
R H 2O = (10)
S b λ H 2O

An energy balance of the borehole loop then gives the incoming brine temperature (Tin brine) to the
evaporator.

q&
Tin brine = TB + (11)
2m& C p

3.4 Modelling of the heat pump

As the overall aim of the project is to develop a tool for evaluation of existing heat pumps, the
modelling of the heat pump is done by the use of performance data files. The performance data files are
produced by use of monitored results from performance tests run according to the European norm
EN 14511. In order to reach good accuracy the number of test points has been increased. The
performance data files have to go through an evaluation by a third party before it is being approved for
use in the software. Heat output as well as electric input has to be monitored at the following test points.

Brine-water heat pump: In addition to the test points in the table below, the maximum temperature of
the water leaving the condenser has to be specified. The EN 14511 gives the flow rate for the brine- and
water-circuit.

Table 1 Test points for brine-water heat pumps

Water temp leaving condenser 35°C 45°C 55°C

Brine temp entering evaporator -5°C x x x
0°C x x x
+5°C x x x

Air-water heat pump: In addition to the test points in the table below, the lowest operation
temperature has to be specified. The outdoor temperatures are given for dry bulb temperature and wet
bulb temperature (with in the brackets).

9
 Table 2 Test points for air-water heat pumps

Water temp leaving condenser 35°C 45°C 55°C

Outdoor temperature -7°C(-8°C) x x x
+2°C(+1°C) x x x
+7°C(+6°C) x x x

Capacity controlled air-air heat pump: Heat output and electric input for the given test points in the
table below are to be taken as the average value over a whole test cycle, including defrosting period, if
required. If the heat pump is supplied with an auxiliary heating cable the electric input for that has to be
included. In addition the lowest operation temperature has to be specified.

Table 3 Test points for capacity controlled air-air heat pumps

Relative capacity 100% 75% 50%

Outdoor temperature -15°C x
-7°C(-8°C) x
+2°C(+1.0°C) x x
+7°C(+6°C) x x x

Linear interpolation is used to obtain the performance in between test points.

4 IMPLEMENTATION OF PROPOSED MODELS IN COMPUTER SOFTWARE

All the models that are described in the previous sections have been implemented in the
commercially available software, Prestige, distributed by The Swedish Heat Pump Association. Much
attention has been paid into the development of a user-friendly interface; enabling sound evaluations of
heat pump systems for users with unpretentious experience in system simulation and advanced simulation
tools. The software is distributed with climate data files for more than 100 Swedish locations and a
number of standardised heat pump performance files. There are at present time more than 600 licensed
users of the software.

5 FURTHER WORK

In order to accomplish the user-friendly interface a lot of the parameters used in the underlying
models are hidden for the user and thus hindering more experienced users for advanced parameter studies.
This automatically raises the demand for a more advanced version of the software, which is currently
under development. An English user interface and climate files for locations outside Sweden will be
implemented to facilitate international use. In order to supply the user with information of thermal
conductivity of the ground rock collaboration with Geological Survey of Sweden has been initiated.

10
ACKNOWLEDGEMENTS

This project would not have been possible without the funding and support from The Swedish
Energy Agency, The Swedish National Testing and Research Institute, NIBE AB, IVT Industrier AB,
Thermia Värme AB and ELFORSK.

NOMENCLATURE

A Area m²
Cp Specifik heat capacity J/(kg·K)
degh Degree hour Kh
Dy,Di Dy, Diameter external, Di, Diameter internal m
Fo Fourier number -
HLF Heat loss factor W/K
L Borehole length M
• Mass flow rate kg/s
m
N Radiator exponent -
Nu Nusselt number -
Pr Prandtl number -
• Heat load W
Q
R, r Radius, Thermal resistance M, (m K)/W
Re Reynolds number -
Sb Shape factor -
T, t, Temperature, time in equation 6 K, s eqn 6
α Heat transfer coefficient W/(m²·K)
λ Thermal conductivity W/(m·K)

REFERENCES

Bernier, B., Randriamiarinjatovo, D. 2004. Annual Simulations of Heat Pump Systems With Vertical
Ground Heat Exchangers. Proceedings from The bi-annual conference of IBPA-Canada, Se-Sim
2004, June 9-11, 2004, Vancouver, BC, Canada.

Bernier, B. 2000. A Review of the Cylindrical Heat Source Method for the Design and Analysis of
Vertical Ground-Coupled Heat Pump Systems, 4th International Conference, Heat Pumps In Cold
Climates, Caneta Research, August 17-18, Ottawa, Canada.

Bernier, B. 2001. Ground-Coupled Heat Pump System Simulation. ASHRAE Transactions 107 part
1.

Carslaw, H. S., Jaeger, J. C. 1947. Conduction of Heat In Solids. Oxford, U.K.

Forsén M. 2000. Degree Days At 13 Nordic Locations, Dept. of Energy Technology, Royal Institute
of Technology. Stockholm.

Hellström G., Sanner B. 2001. PC-programs and modelling for borehole heat exchanger design,
International Geothermal Days Germany, Bad Urach, ISS, Skopje, Macedonia.

11
Ingersoll L. R., Zobel O. J., Ingersoll A. C.1954. Heat Conduction With Engineering, Geological and
Other Applications, Revised Edition, McGraw-Hill.

Kavanaugh S.P., Rafferty K. 1997. Ground-Source Heat Pumps: Design of Geothermal Systems for
Commercial and Institutional Building. ASHRAE, Atlanta, GA, USA.

Klein S. A. 2000. EES- Engineering Equation Solver, F-chart software, Madison, Wisconsin, USA.

Nilsson P-E. 1988. Anslutning av värmepumpar till befintliga värmesystem. Värmetekniska och
ekonomiska begränsningar, Document D6, Chalmers University of Technology.

Peterson F. Reglerkurvan, Debatt #1, A4-Serien #100, Dept. of Energy Technology, Royal Institute of
Technology. Stockholm.

Remund C. P. 1999. Borehole Thermal Resistance: Laboratory and Field Studies, ASHRAE
Transactions 105 part 1.

Swedish Metrological and Hydrological Institute, SMHI. Produktblad Graddagar, SMHI, Norrköping.

Swedish Heat Pump Association, SVEP 2005. Unofficial sales statistics. SVEP Information &
Service AB, Stockholm.

VVS-Handboken 1963. Förlags AB VVS, Stockholm, Sweden.

12