A Thermo-Economic Analysis of Working Fluid and Cycle Combinations For Industrial Heat Pump Applications

Download as pdf or txt
Download as pdf or txt
You are on page 1of 14

DOI: 10.18462/iir.gl2022.

0027

A Thermo-economic Analysis of Working fluid and Cycle combinations


for Industrial Heat Pump Applications
Gustavo J. OTERO RODRIGUEZ*, Andrew J. MARINA, Wouter DE VRIES,
Michel VAN DER PAL, Simon SPOELSTRA
TNO
1755 ZG, Petten, the Netherlands, e-mail
*Corresponding author: Gustavo.OteroRodriguez@tno.nl

ABSTRACT
Successful implementation of heat pumps in the industry requires an adequate balance between two
opposing goals: minimizing the system's capital expenditure (CAPEX) and maximizing the thermodynamic
cycle's coefficient of performance (COP). In this study, we identify for several external conditions the working
fluid and cycle layout which combined can provide end-users with the most economically attractive solution
for the installation of a heat pump in their specific process. Consequently, we coupled numerical models of
heat pump cycles and thermo-physical properties of working fluids with specifically developed cost models
for heat pump components. The simulation results indicate that there is a trade-off between CAPEX and COP.
Other economic metrics highlight the importance of a higher COP, resulting in a low operational expenditure
(OPEX). Natural refrigerants result in higher COP (lower OPEX), but the added costs due to explosive
environment safety for flammable refrigerants hamper the CAPEX.
Keywords: Industrial heat pump, thermo-economic optimization, CAPEX and OPEX tradeoff

1. INTRODUCTION
Industrial heat pumps are a viable alternative to fossil fired boilers to reduce the greenhouse gas (GHG)
emission of industrial processes with a low temperature (<200°C) heat supply. Industrial processes are
responsible for a quarter of the total energy consumption (Eurostat 2019) and approximately 20% of the total
GHG emissions in Europa (EEA 2019); this does not include the indirect emissions attributed to external
energy supply. One of the European Council's measures to reduce the GHG emissions is to improve the
energy efficiency (European Council 2014). An industrial heat pump, utilizing waste heat as a source stream,
works as a multiplier; by spending 1 unit of (renewable) electricity, an industrial heat pump can generate 3
or 4 units of heat (depending on the heat pump performance) at the required temperature.
. Industrial heat pumps are considered a promising option and are gaining market interest. It is
expected that industrial heat pumps will be more integrated into industrial processes coming years through
demonstration projects (i.e., European call: Horizon-CL5-2021-D4-01). The technology that is applied is
usually a Reverse Rankine or closed-cycle compression heat pump. This technology originates from
refrigeration technology and has been adapted to fulfil the requirements for heating applications. Despite
the promising outlook, there are still several research and development areas that can improve the technical
and financial characteristics of these systems.
During the design of a heat pump, the manufacturer is subjected to a conflicting trade-off between
minimizing the Capital expenditure (CAPEX) and the operational expenditure (OPEX); in other words there is
a balance between the total cost of the installation and the energy efficiency of the system. Adequately
balancing this trade-off is vital to improving the business case for the end-user.
This study aims to identify the working fluid and cycle layout, which combined, can provide end-users
with the most attractive economic conditions for the installation of a heat pump in their process. More
specifically, we derive a selection map for vapor compression heat pumps, in terms of the working fluid and
cycle layout, as a function of the external process conditions, for the production of steam at temperatures
up to 200°C. A selection map can provide end-users with the most attractive solution in terms of short and/or
long-term goals. To achieve this task , we coupled numerical models of heat pump cycles and thermophysical
properties of working fluids with cost models developed for heat pump components. We modelled over

15th IIR-Gustav Lorentzen conference on Natural Refrigerants | June 13-15 | Trondheim, Norway
twenty working fluids with six different heat pump layouts (e.g., standard, economizer cycle, with a flash
tank) for four heat capacities (0.5 - 5 MW), four sink temperatures (120-200°C), and four external
temperature lifts (40 - 100 K).
The structure of the present article is as follows: first, the thermo-economic model is described, including
the optimization procedure. Later, we introduce the boundary conditions in terms of external (sink/source
temperature and heat duty) and internal (working media and heat pump layout). Next, we discuss the results
of the thermo-economic optimization in section 4, which presents the industrial heat pump selection map.
The final section of this article presents conclusions and future works.

2. THERMO-ECONOMIC MODEL OF A VAPOR COMPRESSION HEAT PUMP


. We depict a schematic of a vapor compression heat pump in Figure 1. As the aim is a selection
guideline, we developed a techno-economic model of a vapor compression heat pump that accounts for
changes in the working media and thermodynamic cycle layout.
This section describes a general method for evaluating the thermodynamic performance, overall
capital cost, and economic metrics for a vapor compression heat pump. We apply this methodology to
compare different working media and heat pump layouts in terms of thermodynamic performance, the total
cost of the system, and financial viability. We implemented the thermo-economic model in Engineering
Equation Solver (EES) while the post-processing and general optimization were done in Python.

Figure 1: Schematic of a vapor compression heat pump

2.1. Thermodynamic model and equipment sizing


We made a thermodynamic model for three types of heat pump layouts: simple Reverse Rankine
cycle, Economizer cycle, and Flash tank cycle. In addition, an internal heat exchanger (or suction gas
superheater) was included in all layouts.
The numerical calculation of the thermodynamic cycle is based on energy and mass balances. The
compressor is modeled using an isentropic (𝜂𝑖𝑠𝑒𝑛 = 70%), mechanical (𝜂𝑚𝑒𝑐 = 90%), and volumetric
efficiency (linear function of the volume ratio derived from data from compressor manufacturer). We have
assumed no pressure drops within the piping or heat exchangers of the system. Moreover, we model the
expansion valves as isenthalpic. Furthermore, the evaporator and condenser are modelled using the pinch
point temperature difference with the source and sink media. Note that the pinch point temperature
difference is an input to the thermodynamic model, which can be optimized; however, this will impact the
heat transfer area and, therefore, the capital cost of the heat exchanger.
Other thermodynamic inputs used were:
• The outlet of the condenser is saturated liquid.
• the superheat at the compressor outlet (before the condenser)

15th IIR-Gustav Lorentzen conference on Natural Refrigerants | June 13-15 | Trondheim, Norway
• the outlet superheat of the evaporator for the heat pump layouts with internal heat exchanger (IHX),
and
• the mass flow fraction that goes through the economizer (for economizer heat pump layouts).
These parameters can also be optimized for better thermodynamic performance (or lower investment). cycle
layout.

2.1.1. Performance indicators


We calculated the coefficient of performance (COP) and the thermodynamic efficiency (𝜂1𝑠𝑡𝐿𝑎𝑤 ) of the heat
pump. The COP is the ratio between the heat output and the power used by the heat pump:
𝑄̇𝑐𝑜𝑛𝑑
𝐶𝑂𝑃 = 𝑊̇𝑐𝑜𝑚𝑝
Eq. (1)

The thermodynamic efficiency is the ratio of the coefficient of performance of the system with respect to the
ideal heat pump cycle:
𝐶𝑂𝑃
𝜂1𝑠𝑡𝐿𝑎𝑤 = 𝐶𝑂𝑃𝑖𝑑𝑒𝑎𝑙
Eq. (2)

The ideal COP is defined as:


𝑇̅𝑠𝑖𝑛𝑘
𝐶𝑂𝑃𝑖𝑑𝑒𝑎𝑙 = 𝑇̅𝑠𝑖𝑛𝑘 −𝑇̅𝑠𝑜𝑢𝑟𝑐𝑒
Eq. (3)

External temperatures were used for the definition of the ideal performance. Moreover, 𝑇̅𝑠𝑖𝑛𝑘 and 𝑇̅𝑠𝑜𝑢𝑟𝑐𝑒
are logarithmic mean temperatures that consider temperature glides.

2.1.2. Equipment sizing


To estimate purchased equipment cost, we need to estimate the capacity or size of each piece of
equipment within the heat pump system. This section discusses sizing of heat exchangers, compressor,
electric motor, piping, storage vessel, and valves.
The area of heat exchangers is the crucial parameter for the cost of this component and we therefore
correlated it to the equipment cost. The heat transfer analysis is based on the logarithmic mean temperature
difference (LMTD) method. As mentioned, the pinch temperature difference (closest approach between the
hot and cold stream) is imposed. To calculate the heat transfer area 𝐴𝐻𝑇 , we use:

𝑄̇ = 𝐻𝑇𝐶 ∗ 𝐴𝐻𝑇 ∗ Δ𝑇𝐿𝑀 Eq. (4)

Where Δ𝑇𝐿𝑀 is the logarithmic mean temperature difference:


(𝑇 −𝑇𝐶𝑆,𝑜𝑢𝑡 )−(𝑇𝐻𝑆,𝑜𝑢𝑡 −𝑇𝐶𝑆,𝑖𝑛 )
Δ𝑇𝐿𝑀 = 𝑙𝑛[(𝑇𝐻𝑆,𝑖𝑛 Eq. (5)
𝐻𝑆,𝑖𝑛 −𝑇𝐶𝑆,𝑜𝑢𝑡 )/(𝑇𝐻𝑆,𝑜𝑢𝑡 −𝑇𝐶𝑆,𝑖𝑛 )]

With the subscripts HS and CS as the hot side and cold side, respectively. The overall heat transfer coefficient
is the summation of the heat resistance across the heat exchanger (neglecting the resistance due to
conduction through the metal and fouling):
1 1 1
𝐻𝑇𝐶
= 𝐻𝑇𝐶 + 𝐻𝑇𝐶 Eq. (6)
𝐶𝑆 𝐻𝑆

𝑊
Constant heat transfer coefficient as given by Zuhlsdorf et al. (2018) were used: Evaporation 3000 ,
𝑚2 𝐾
𝑊 𝑊 𝑊
Condensation 2400𝑚2 𝐾, Liquid 1500 𝑚2 𝐾, and Gaseous 1200 𝑚2 𝐾.
The compressor has a large impact on the heat pump overall performance and the total cost of the
installation. We parametrized the compressor with the power input and the swept volume:
𝑚̇ 1
𝑣̇ = 𝜌 𝜂
Eq. (7)
𝑖𝑛 𝑣𝑜𝑙

The motor driving the compressor is sized with the mechanical power provided:

15th IIR-Gustav Lorentzen conference on Natural Refrigerants | June 13-15 | Trondheim, Norway
𝑊 ̇
𝑊̇𝑚𝑜𝑡𝑜𝑟 = 𝑐𝑜𝑚𝑝 Eq. (8)
𝜂𝑚𝑒𝑐

The diameter of the pipe between each cycle component (𝐷𝑝𝑖𝑝𝑒,𝑖 ) is sized according to typical fluid velocities
applied in refrigeration units, selecting a fluid velocity of 1 m/ s and 12.5 m/s for the liquid and vapor sections,
respectively. Thereafter, the length of the piping for each section (𝐿𝑝𝑖𝑝𝑒,𝑖 ) is defined with the following
expression:

𝐿𝑝𝑖𝑝𝑒,𝑖 = 𝐿𝑚𝑖𝑛 + 𝑓𝑝𝑖𝑝𝑒 𝐷𝑝𝑖𝑝𝑒,𝑖 Eq. (9)

Where the minimal length of the pipe is set to 𝐿𝑚𝑖𝑛 =2 m and the piping distance factor of 𝑓𝑝𝑖𝑝𝑒 =20.
The storage vessel is sized by calculating the amount of liquid in the system (within the liquid lines) and
multiplying it by a factor (1.25):

𝑉𝑜𝑙𝑠𝑡𝑜𝑟𝑎𝑔𝑒𝑉𝑒𝑠𝑠𝑒𝑙 = 1.25 ∗ 𝜋 ∗ ∑ 𝐿𝑝𝑖𝑝𝑒,𝑙𝑖𝑞 ∗ 𝐷𝑝𝑖𝑝𝑒,𝑙𝑖𝑞 Eq. (10)

Finally the valves within the system are characterized using the pipe diameter at the inlet of the valve.

2.2. Cost models


To calculate the total cost of the installation (TCI), we sum the cost of each component (heat
exchangers, compressors, valves, piping, storage vessel, etc.) and add the indirect cost of the system
(engineering cost, assembly cost, among others). The main advantage of this approach is that all cost
correlations are expressed as a function of the component size/capacity. This enables the analysis of the
economic benefits as a function of the system scale. Most studies focus on the cost of the most expensive
components (i.e., compressor and heat exchangers) and multiplying their purchased cost by a factor (of
around 4, see Zuhlsdorf 2018 and Kosmadakis 2020 for example) to estimate the TCI.
Alternatively, we gathered cost data of standard equipment within a heat pump to derive cost
models. We dimension each component and correlate a geometrical characteristic to the cost of the
equipment (Bejan et al. 1996). Moreover, we searched in open literature for cost models — especially for the
compressor and heat exchangers— used in previous techno-economic models for heat pumps (Van de Bor
and Infante Ferreira 2013, Ommen et al 2015 and Zühlsdorf et al 2018).
We developed cost correlations from quotation data available internally at TNO, from partners (e.g.,
Bronswerk Heat transfer) and/or from information available in an independent cost estimate data for the
process industry (Dutch Association of Cost Engineers (DACE) price booklet). From the cost data, a Purchased
equipment cost (PEC) function is constructed as proposed by Bejen et al. 1996 where the purchase cost of an
equipment is a function of the size or capacity. For most equipment, we use the following exponential
expression:
𝑋 𝛼
𝑃𝐸𝐶𝑌 = 𝑃𝐸𝐶𝑊 (𝑋 𝑌 ) Eq. (11)
𝑊

When more accurate, a linear function of the capacity is used:

𝑃𝐸𝐶𝑌 = 𝐶𝐴 𝑋𝑌 + 𝐶𝐵 Eq. (12)

We checked the validity of the cost models using quotations available within TNO. The references used came
from previous and ongoing projects. We chose the cost models — for the techno-economic model of the
heat pump— that had the lowest relative difference with the quotations.

2.2.1. Other direct costs


The remaining direct cost component consists of the purchased equipment installation (or assembly
costs), the piping, and the electrical and instrumentation equipment:
• Assembly cost: dependent on the hourly rate and approximate how long the assembly may take (in weeks
or days)

𝐴𝑠𝑠𝑒𝑚𝑏𝑙𝑦𝑐𝑜𝑠𝑡 = 𝐻𝑜𝑢𝑟𝑙𝑦𝑟𝑎𝑡𝑒 ∗ (#𝑤𝑒𝑒𝑘𝑠 ∗ 40) Eq. (13)

15th IIR-Gustav Lorentzen conference on Natural Refrigerants | June 13-15 | Trondheim, Norway
• Piping costs: we have derived the piping cost model using data from the DACE price booklet.
• Instrumentation: The cost regarding the instrumentation and control of the system are related to the
number of sensors (mainly temperature and pressure) in the heat pump (one per piping connection
between components) and the control panel/electronics:

PE𝐶𝑖𝑛𝑠𝑡𝑟𝑢𝑚𝑒𝑛𝑡𝑎𝑡𝑖𝑜𝑛 = 𝑃𝐸𝐶𝑝𝑟𝑒𝑠,𝑠𝑒𝑛𝑠𝑜𝑟 𝑁𝑝𝑟𝑒𝑠,𝑠𝑒𝑛𝑠𝑜𝑟 + 𝑃𝐸𝐶𝑡𝑒𝑚𝑝,𝑠𝑒𝑛𝑠𝑜𝑟 𝑁𝑡𝑒𝑚𝑝,𝑠𝑒𝑛𝑠𝑜𝑟 + 𝑃𝐸𝐶𝑐𝑜𝑛𝑡𝑟𝑜𝑙,𝑝𝑎𝑛𝑒𝑙 Eq. (14)

2.2.2. Indirect costs


Opposed to the direct costs, indirect costs do not become a permanent part of the facilities. Still they
are required for the completion of the project (Bejan et al. 1996). In the current analysis, we lumped indirect
costs (e.g., engineering, added safety costs due to explosive environments and other contingencies) into a
factor. We define the engineering and contingency costs as an added factor to the sum of all the purchased
equipment costs. For contingencies and/or a safety factor, we can increase the percentage of engineering
cost. Moreover, for flammable working media where higher safety measures (ATEX environment1) are
necessary, we have added an extra percentage of 20-30%.

𝑇𝐶𝐼 = (1 + 𝑃𝑒𝑟𝑐𝑒𝑛𝑡𝑎𝑔𝑒𝐸𝑛𝑔/𝐴𝑇𝐸𝑋,𝑐𝑜𝑠𝑡 ) ∗ (∑ 𝑃𝐸𝐶 + 𝐴𝑠𝑠𝑒𝑚𝑏𝑙𝑦𝑐𝑜𝑠𝑡 ) Eq. (15)

2.3. Economic metrics


The economics of a heat pump is determined by compared costs versus income. The operational cost
can be characterized by annual cash flows for incomes and expenses. In the case of a heat pump, the
expenses consist of the cost for the consumed electricity (𝐶𝐹𝑒𝑙 = 0.053 €/𝑘𝑊ℎ). The income cash flow
supply is achieved by the supply of heat that is transferred to the heat sink. If the heat pump replaces a
natural gas boiler, the income is considered the savings obtained by not operating the boiler with natural gas
(𝐶𝐹𝑠𝑢𝑝𝑝𝑙𝑦 = 0.027 €/𝑘𝑊ℎ). We consider the cash flow associated with the waste heat source as free
(𝐶𝐹𝑠𝑜𝑢𝑟𝑐𝑒 = 0.0€/𝑘𝑊ℎ𝑟). Finally, there is an added cash flow regarding the cost of CO2 (𝐶𝐹𝐶𝑂2 =
0.0177820 €/𝑘𝑊ℎ); for the heat pump, it is a saving with respect to the conventional natural gas boiler. The
values of these cash flows were taken from a 2020 PBL analysis on heat pumps subsidies (Marsidi and Lensink
2020) and therefore do not consider the rising energy prices of 2021.
The comparison of annual cash flows and on-time expenses requires the calculation of present time
values of future costs, which according to Bejan et al. 1996 can be done using the capital recovery factor
(CRF):
𝑛
𝑖𝑒𝑓𝑓 (1+𝑖𝑒𝑓𝑓 )
𝐶𝑅𝐹 = 𝑛 Eq. (16)
(1+𝑖𝑒𝑓𝑓 ) −1

Where n is the lifetime of the heat pump (20 years) and 𝑖𝑒𝑓𝑓 is the effective interest rate defined as
1+𝑖
𝑖𝑒𝑓𝑓 = 1+𝑖 𝑖𝑛𝑡𝑒𝑟𝑒𝑠𝑡 − 1 Eq. (17)
𝑖𝑛𝑓𝑙𝑎𝑡𝑖𝑜𝑛

where we use 2.5% and 1.5% as the interest rate and inflation rate, respectively, taken from PBL.
We evaluated several economic indicators used in investment decisions .
Net present value (NPV): is the difference between the present value of cash inflows and the present value
of cash outflows over a period of time. The projects with the highest present values are given the highest
preference among various alternatives (Bejan et al. 1996).
𝐶𝐹𝑠𝑢𝑝𝑝𝑙𝑦 +𝐶𝐹𝐶𝑂2 −𝐶𝐹𝑒𝑙 ±𝐶𝐹𝑠𝑜𝑢𝑟𝑐𝑒
𝑁𝑃𝑉 = −𝑇𝐶𝐼 + Eq. (18)
𝐶𝑅𝐹

Pay-back time (PBT): is defined as the length of time required for the cash inflows received from a project to
recover the original cash outlays required by the initial investment (Bejan et al.1996).
𝑇𝐶𝐼
𝑃𝐵𝑇 = 𝐶𝐹 +𝐶𝐹
Eq. (19)
𝑠𝑢𝑝𝑝𝑙𝑦 𝐶𝑂2 −𝐶𝐹𝑒𝑙 ±𝐶𝐹𝑠𝑜𝑢𝑟𝑐𝑒

1
According to the European directive, an ATEX environment considers the safety of a workplace concerning
explosion hazardous areas and equipment. Extra measures need to be taken to create a safe environment
concerning explosion hazards, making this an additional installation cost.

15th IIR-Gustav Lorentzen conference on Natural Refrigerants | June 13-15 | Trondheim, Norway
Specific cost of heat (SCH): This cost relates the annual fuel cost, the annual operating or income cost, as
well as the annual equivalent of the investment cost, to the amount of supplied heat per year (for a certain
number of operating hours).
𝐶𝐹𝑒𝑙 −±𝐶𝐹𝑠𝑜𝑢𝑟𝑐𝑒 +𝑇𝐶𝐼∗𝐶𝑅𝐹
𝑆CH = 𝑄̇𝑠𝑖𝑛𝑘 ∗𝑂𝐻
Eq. (20)

Note that we did not consider operation and maintenance costs; maintenance costs are often taken as
2-4% of the original investment cost per year (Bejan et al.1996). Moreover, we have taken the operating hours
(OH) as 8000 hours per year.

2.4. Optimization
We depict a schematic of the overall thermo-economic model of a heat pump in Figure 2. The coloured
squares in this figure highlight the different inputs, outputs, and parameters to optimize. Hereunder we
explain/describe these variables ending with the optimization objective and procedure.

Figure 2: Schematic of the techno-economic model of a heat pump

The blue squares are the inputs to the models:


• the external boundary conditions,
• sizing parameters used for the equipment sizing,
• purchased equipment data and cost indicators used for the derivation of the cost models, and
• economic inputs used to calculate the economic metrics.
The outputs, in the red squares, from the model are threefold:
• thermodynamic performance of the heat pump (e.g., COP) which is related to the operational cost of the
installation (OPEX),
• the total cost of the installation or capital expenditure (CAPEX), and
• the economic metrics that consider both the CAPEX and OPEX of the installation.
Finally, the yellow squares represent the parameters that we can optimize. As shown in Figure 3, we perform
two optimizations: (1) an individual optimization of a given system as a function of the thermodynamic inputs
and (2) a general optimization were we compare the different heat pump layout and working media
combinations and define a selection map as a function of the external boundary conditions. Both
optimizations aim towards the following uncoupled goals:
a) Maximize the COP of the heat pump. As a consequence, the OPEX will be minimized.
b) Minimize the TCI of the heat pump; this is a short-term economic goal resulting in a lower CAPEX.
c) Maximize the financial viability of the project by optimizing the economic metrics. For instance,
maximizing the net present value or minimizing the pay-back time and/or specific cost of heat.

15th IIR-Gustav Lorentzen conference on Natural Refrigerants | June 13-15 | Trondheim, Norway
Figure 3: Optimization procedure

3. BOUNDARY CONDITIONS

3.1. External boundary conditions


We have focused our analysis on heat pumps that recover heat from liquid water streams to
generate steam at the sink for heating capacities below 10 MW. In many cases, steam fulfils the heat
requirement in industry (Arpagaus2016) as low-pressure steam is used as energy carrier in many industries
such as paper, food, and chemical sector (Arpagaus 2018). Therefore, we analyze the production of steam at
several temperature levels (𝑇𝑠𝑖𝑛𝑘 ∈ 120 − 200℃). Most of the potential of heat pumps installation in Europe
are for heating capacities of less than 10 MW (Marina 2021). Consequently, we focused our analysis within a
range of 0.5 - 10 MW heat output (Marina 2021). The last external boundary condition that we consider is
the external temperature lift, which defines the temperature difference between the sink outlet and the
source inlet . We impose temperature lifts from 20 – 100 K.

3.2. Internal boundary conditions: working fluids and heat pump layout
The analysis primary focus is to develop a selection map for heat pump layout and working fluid
combinations that provides the optimum depending on the goal (i.e., thermodynamic performance, lowest
capital cost, or most financially interesting project). We chose twenty-five (25) working fluids and three
different heat pump layouts (simple, economizer, and flash tank with or without IHX depending on fluid type).
Table 1 lists all the investigated working fluids. Most of the refrigerants in the list have been
mentioned/studied in literature as possible working fluids for high-temperature heat pumps. We have added
other working media due to their high-performance potential at high temperature (e.g., Dimethyl Carbonate,
cyclohexane, cyclopentane). We differentiate between the working fluids in terms of their type of
compression (wet or dry compression) and their flammability. The former depends on each working fluid's
thermodynamic characteristics: the liquid-vapor saturation distribution in the temperature-entropy diagram.
In terms of the latter, if a working media is flammable extra costs (25% added to the total direct costs) are
included in the installation due to the required ATEX environment. We disregarded particular combinations
of working fluids and external boundary conditions because the condensing temperature of the cycle is
higher than the working media critical temperature or if the evaporating pressure of the heat pump is below
sub-atmospheric2.

2
It is not recommended to have part of the heat pump system under vacuum due to safety reasons and air ingress.

15th IIR-Gustav Lorentzen conference on Natural Refrigerants | June 13-15 | Trondheim, Norway
Table 1. List of working media
Working fluid Tcrit (°C) Compression Flammability
Water 373.95 Dry -
Dimethyl Carbonate (DMC) 283.85 Wet 3
Cyclohexane 280.45 Wet 3
Methanol 240.23 Dry 3
Cyclopentane 238.57 Wet 3
Acetone 234.95 Dry 3
Hexane 234.67 Wet 3
Isohexane 224.55 Wet 3
Pentane 196.55 Wet 4
Diethyl Ether 193.55 Wet 4
Isopentane 187.20 Wet 4
R1336mzz(Z) 171.35 Wet -
Novec 649, 1230 168.66 Wet -
R1233zd(E) 166.45 Wet -
Neopentane 160.59 Wet 4
R1224yd(Z) 155.54 Wet -
Trans-Butene 155.46 Wet 4
Butane 151.98 Wet 4
R1234ze(Z) 150.12 Wet -
Butene 146.14 Wet 4
Isobutene 144.94 Wet 4
Isobutane 134.66 Wet 4
Ammonia 132.41 Dry 1
Dimethyl ether 127.23 Dry 4
R1234ze(E) 109.36 Wet -

We evaluate three different heat pump layouts: the standard Reverse Rankine, the economizer cycle,
and the flash tank cycle. We depict all the heat pump layouts assessed in this project in Figures 4 and 5 for a
dry and wet expansion fluid, respectively. The wet compression fluids need a suction gas superheater at the
compressor; therefore, for such fluids, we have added an internal heat exchanger to provide superheat at
the inlet of the compressor. Unlike the standard Reverse Rankine layout, the economizer and flash tank cycle
consists of a two-stage system where not the entire refrigerant is compressed from evaporating to
condensing pressure. In the flash tank cycle, the entire refrigerant is expanded after the condenser to a mid-
pressure where the liquid and vapor of the refrigerant are separated. The remaining liquid is further
expanded to the evaporating pressure whereas the vapor is fed to the high-pressure compressor. In the
economizer cycle, the refrigerant is separated into two streams after the condenser: one of them is expanded
to mid pressure3 and is then evaporated utilizing the sub-cooled heat from the other refrigerant stream. The
latter goes through an expansion to the evaporating (low) pressure.

3
The mid pressure is defined such that the pressure ratio of both compressor is equivalent; therefore the pressure ratio of each
compressor is the square root of the overall pressure ratio.

15th IIR-Gustav Lorentzen conference on Natural Refrigerants | June 13-15 | Trondheim, Norway
Figure 4: Schematic of a heat pump for dry compression fluids

Figure 5: Schematic of a heat pump for wet compression fluids (including an internal heat exchanger IHX )

4. HEAT PUMP SELECTION MAPS


In this section of the report, we present and discuss the results from the thermo-economic optimization
of the heat pump. We have derived heat pump selection maps for the different optimization goals: higher
thermodynamic performance, cheaper installation, and financially attractive project. We represent the
selection map as a function of the sink temperature and the temperature lift, a template is depicted in Figure
6. We do not include the dependency concerning the heating capacity (third axis of the selection map)
because the 𝑄̇𝑠𝑖𝑛𝑘 has a small influence on the selection maps (less than 5% difference in the economic
metrics) for working media and heat pump layout. However, the heating capacity has a large effect on the
absolute economic metrics of the selected map; a larger heating capacity (i.e., 5 MW) is economically more
profitable.

15th IIR-Gustav Lorentzen conference on Natural Refrigerants | June 13-15 | Trondheim, Norway
Figure 6: Heat pump selection map as a function of the external boundary conditions

4.1.1. Heat pump selection map for maximum COP


Figure 7 depicts the selection map for heat pump layout and working media for maximum COP. The
flash tank with an internal heat exchanger (FlashTankIHX) is the heat pump layout that gives the best
thermodynamic performance from the three heat pump layouts analyzed. In terms of the working media, a
fluid with a wet compression performs better in terms of COP according to our thermodynamic model. The
hydrocarbons work best at a lower sink temperature (for 𝑇𝑠𝑖𝑛𝑘 ∈ 120 − 160 ℃) and DMC (at least for source
temperature larger than 120 ℃) for higher sink temperature (for 𝑇𝑠𝑖𝑛𝑘 ∈ 180 − 200 ℃); this result can be
related to the critical temperature of the working media. From the selection map (Figure 7) it is clear that
there is a correlation between heat pump layout + working media with respect to the source (evaporating,
see figure 1) temperature; the diagonal distribution is for a constant source temperature (𝑇𝑠𝑜𝑢𝑟𝑐𝑒 = 𝑇𝑠𝑖𝑛𝑘 −
Δ𝑇𝑙𝑖𝑓𝑡 ). For example, the flash tank + cyclo-pentane combination (in red in Figure 7) is for a constant source
temperature of 80℃.

Figure 7: Heat pump layout and working media selection map for a maximum COP

4.1.2. Heat pump selection map for lowest TCI


Figure 8 depicts the selection map for heat pump layout and working media for the cheapest heat
pump installation. The selection map for TCI does not match with the previous discussion for the COP. This
confirms our hypothesis that there is a tradeoff between the CAPEX and the OPEX.
The optimization routine usually selects the simplest heat pump layout: the standard heat pump with
an internal heat exchanger. This result is a consequence of fewer components. There are exceptions for
extreme cases like high or low temperature lift or sink temperature, where there are fewer fluids available
or a larger chance of temperature crossing within in the heat exchangers.

15th IIR-Gustav Lorentzen conference on Natural Refrigerants | June 13-15 | Trondheim, Norway
The Hydrofluoroolefins (HFOs) —R1233zd(E) and R1234ze(Z) — have taken the place over from the
hydrocarbons for low sink temperatures (for 𝑇𝑠𝑖𝑛𝑘 ∈ 120 − 140℃) because the ATEX environment is not
needed for such fluids. For higher sink temperatures, hydrocarbons are the cheapest option because the
model discriminates the selected HFOs; their critical temperature is relatively low compared to the
hydrocarbons. There is a need for HFOs (or non-flammable fluid) with a higher critical temperature to develop
cost-effective high-temperature heat pumps.
From the selection map optimizing for TCI, it is clear that there is a correlation between heat pump
layout + working media concerning the sink (condensing) temperature. This is again opposite from the results
using the COP as the optimized variable.

Figure 8: Heat pump layout and working media selection map for a minimum total investment cost including the
ATEX environment costs

4.1.3. Heat pump selection map for lowest SCH


Figure 9 depicts the selection map for heat pump layout and working media for the most financially
viable project. We depict the resulting selection map only for the SCH, but identical results were found with
the NPV. Similar results were also found for the PBT with some distinct exceptions; it must be noted that we
use the simple PBT where the value of money does not change in time; this is the reason why we choose the
SCH and NPV over the PBT.
The selection map for the SCH matches quite well with the one from the COP; the distribution is
concerning the source temperature. The OPEX has a more significant influence than the CAPEX over the
economic metrics for the economic inputs used in this study. In terms of the working media, wet compression
fluids are mostly selected, except methanol at high sink temperatures (>160°C) and low-temperature lifts.
The HFOs are optimum for a sink temperature of 120°C, similar to the selection map using TCI as the
optimized parameter. This is a consequence of the extra safety measures for explosive hazardous
environment cost.
In terms of the heat pump layouts, the flash tank layout performs best in general. At higher sink
temperatures, the selected heat pump layouts are similar to those from the optimization using TCI. From
these results, at high sink temperature and low-temperature lift, the CAPEX of the system has a larger
influence over the economic metrics to the selected heat pump layout and working media. Further analysis
must be performed to validate this hypothesis

15th IIR-Gustav Lorentzen conference on Natural Refrigerants | June 13-15 | Trondheim, Norway
Figure 9: Heat pump layout and working media selection map for the specific cost of heat including the ATEX
environment costs

5. CONCLUSIONS
Heat pump technologies can significantly reduce the CO2 emissions from industrial processes. Their
successful implementation requires an adequate balance between two opposing goals: minimizing the
system's capital expenditure (CAPEX) and maximizing the thermodynamic cycle's coefficient of performance
(COP). This conference paper presents a detailed thermo-economic study of high-temperature heat pumps.
Various parameters were optimized to develop a selection guideline for working fluid and layout as a function
of external conditions for short and long-term financial attractiveness.
The selection maps for working fluid and heat pump layout correspond to previous TNO studies. In
general, there is a correlation between the working media and heat pump layout with respect to one of the
external temperatures: the source temperature for COP and SCH, and the sink temperature for TCI. The
selection map derived from the optimization of the COP correlates with the optimum economic metrics for
most of the external temperature conditions. The OPEX has a significant influence —and even more so with
the rising energy prices in 2021—on the SCH, NPV, among others. Moreover, the heat pump layout with a
flash tank and internal heat exchanger results in most cases in better thermodynamic performance and lower
operational costs. On the other hand, in terms of minimizing the CAPEX, the simplest heat pump layout
(standard heat pump cycle) is preferred. The ATEX costs influence the selection maps for TCI and other
economic metrics at lower sink temperatures. There are no HFOs available — at least not considered in this
study— for high-temperature (>140°C) heat pumps. Our recommendation is to optimize the heat pump for
one of the economic metrics (specific cost of heat or net present value) as it considers both the CAPEX and
OPEX.
Future studies will include reporting these results to end-users and making them aware of the trade-
off between OPEX and CAPEX. Moreover, we plan to use the results from this study —the identified working
media and heat pump layouts— to generate an experimental plan for high-temperature heat pumps.

15th IIR-Gustav Lorentzen conference on Natural Refrigerants | June 13-15 | Trondheim, Norway
ACKNOWLEDGEMENTS
The project is partially funded by subsidies of the Topsector Energy from the Ministry of Economic Affairs
and Climate Policy (EZK) of the Netherlands — the project number is TEEI118001.

NOMENCLATURE
Acronyms 𝑄̇ : Heat duty [𝑘𝑊]
ATEX: Atmospheres Explosibles 𝑠: Entropy [𝑘𝐽/(𝐾 𝑘𝑔)]
CAPEX: Capital expenditure 𝑇: Temperature [℃] or [K]
CF: Cash flow [€/kWh] 𝑣̇ : Volumetric flow rate [𝑚3 /𝑠]
CRF: Capital recovery factor [-] 𝑉𝑜𝑙: Volume [𝑚3 ]
COP: Coefficient of performance 𝑊̇ : Work [𝑘𝑊]
DACE: Dutch Association of Cost Engineers
DMC: Dimethyl carbonate Greek letters
GHG: Greenhouse gases Δ: Difference
HFO: Hydrofluoroolefins ΔT𝐿𝑀 : Log. mean temperature difference [𝐾]
HTC: Heat transfer coefficient [𝑊/(𝑚2 𝐾)] η: Efficiency
IHX: Internal heat exchanger Φ: Exergy [𝑘𝐽/𝑘𝑔]
LMTD: Logarithmic mean temperature difference 𝜌: Density [𝑘𝑔/𝑚3 ]
NPV: Net present value 𝜇: Viscosity [𝑁 𝑠/𝑚2 ]
OH: Operating hours [ℎ𝑟/𝑦𝑒𝑎𝑟]
OPEX: Operational Expenditure Subscripts
PBL: Planbureau voor de Leefomgeving Comp: Compressor
(Netherlands Environmental Assessment Agency) Cond: Condenser / condensing stream
PBT: Pay-back time [𝑦𝑒𝑎𝑟] El/elect: Electricity
PEC: Purchased equipment cost [€] Evap: Evaporator / evaporating stream
SCH: Specific cost of heat Gas: Gaseous stream
TCI: Total cost of investment Liq: Liquid stream
TNO: Toegepast Natuurwetenschappelijk Isen: Isentropic
Onderzoek (Applied Scientific Research) Mec: Mechanical
NG: Natural gas
Vol: Volumetric
Symbols CS: Cold side stream of a heat exchanger
𝐷: Diameter [𝑚] HS: Hot side stream of a heat exchanger
𝑓: Factor in: inlet stream
𝑖: Interest rate [−] out: outlet stream
ℎ: Enthalpy [𝑘𝐽/𝑘𝑔]
𝐿: Length [𝑚]
𝑚̇: Mass flow [𝑘𝑔/𝑠]
𝑛𝐿𝑖𝑓𝑒 : Life time [𝑦𝑒𝑎𝑟]

REFERENCES
Eurostat, 2019. Energy Balances.

European Environment Agency, 2019. Greenhouse gas emissions by aggregated sector,


https://www.eea.europa.eu/data-and-maps/daviz/ghg-emissions-by-aggregated-sector-5 (accessed July 3,
2020).

European Council, 2014. Climate and energy policy framework.

Klein, S., 2006. Engineering Equation Solver (EES) V10.2. Fchart software, Madison, USA www.fchart.com

DACE price booklet, 2020. Independent cost estimate data for the process industry, edition 35,
https://www.dacepricebooklet.com/.

15th IIR-Gustav Lorentzen conference on Natural Refrigerants | June 13-15 | Trondheim, Norway
C. Arpagaus, F. Bless, J. Schiffmann, S.S. Bertsch, 2016. Pom-pes à chaleur à multiples températures: Une
synthèse de la littérature, International Journal of Refrigeration 69 437–465.
doi:10.1016/j.ijrefrig.2016.05.014.URL http://dx.doi.org/10.1016/j.ijrefrig.2016.05.014

C. Arpagaus, F. Bless, M. Uhlmann, J. Schiffmann, S.S. Bertsch, 2018. High temperature heat pumps: Market
overview, state of the art, research status, refrigerants, and application potentials, Energy 152
985–1010.doi:10.1016/j.energy.2018.03.166.

J. Hamilton, J. Miller, 1990, A simulation program for modeling an air-conditioning system, ASHRAE
transactions 96 (1), 213–221.

G. Angelino, C. Invernizzi, 1988. General method for the thermodynamic evaluation of heat pump working
fluids, International journal of refrigeration 11 (1) 16–25.

F-Chart, 1992. Engineeringequationsolver(ees),http://fchartsoftware.com/ees/.

T. Ommen, J.K. Jensen, W.B. Markussen, L. Reinholdt, B. Elmegaard, 2015. Technical and economic working
domains of industrial heat pumps: Part 1 - Single stage vapour compression heat pumps, International Journal
of Refrigeration 55, 168–182. doi:10.1016/j.ijrefrig.2015.02.012.

A. Bejan, G. Tsatsaronis, M.J. Moran, 1995. Thermal design and optimization, John Wiley & Sons.

B. Zühlsdorf, W. Meesenburg, T.S. Ommen, J. E. Thorsen, W.B. Markussen, B. Elmegaard, 2018. Improving
the performance of booster heat pumps using zeotropic mixtures, Energy 154 390–402.
doi:10.1016/j.energy.2018.04.137.URLhttps://doi.org/10.1016/j.energy.2018.04.137

D.M. Van de Bor, C.A. Infante Ferreira, 2013. Quick se-lection of industrial heat pump types including the
impact of thermodynamic losses, Energy 53 312–
322.doi:10.1016/j.energy.2013.02.065.URLhttp://dx.doi.org/10.1016/j.energy.2013.02.065

Marina, A.J., Spoelstra, S., Zondag, H.A., & Wemmers, A.K. 2021. An estimation of the European industrial
heat pump market potential, Renewable and Sustainable Energy Reviews 139: 110545.

Marsidi, M. and Lensink, S., 2020. Conceptadvies SDE++ 2021 Grootschalige Warmepompen.

Kosmadakis, G., Arpagaus, C., Neofytou, P., & Bertsch, S., 2020. Techno-economic analysis of high-
temperature heat pumps with low-global warming potential refrigerants for upgrading waste heat up to
150°C. Energy Conversion and Management, 226, 113488.

Tsotsas, Evangelos, and Arun S. Mujumdar, 2011. Modern drying technology, volume 4: Energy savings. John
Wiley & Sons.

15th IIR-Gustav Lorentzen conference on Natural Refrigerants | June 13-15 | Trondheim, Norway

You might also like