Automation, or Russian Roulette? - IDEALS @ Illinois: IDEALS Home
Minimizing TEWI by Charge Reduction in a Compact Chiller - Ideals
Transcript of Minimizing TEWI by Charge Reduction in a Compact Chiller - Ideals
Minimizing TEWI by Charge Reduction in a Compact Chiller
ACRC TR-176
For additional information:
Air Conditioning and Refrigeration Center University of Illinois Mechanical & Industrial Engineering Dept. 1206 West Green Street Urbana,IL 61801
(217) 333-3115
P. R. Barnes and C. W. Bullard
August 2000
Prepared as part of ACRC Project 69 Stationary Air Conditioning System Analysis
C. W. Bullard, Principal Investigator
The Air Conditioning and Refrigeration Center was founded in 1988 with a grant from the estate of Richard W. Kritzer, the founder of Peerless of America Inc. A State of Illinois Technology Challenge Grant helped build the laboratory facilities. The ACRC receives continuing support from the Richard W. Kritzer Endowment and the National Science Foundation. Thefollowing organizations have also become sponsors of the Center.
Amana Refrigeration, Inc. Ar~elik A. S. Brazeway, Inc. Carrier Corporation Copeland Corporation DaimlerChrysler Corporation Delphi Harrison Thermal Systems Frigidaire Company General Electric Company General Motors Corporation Hill PHOENIX Honeywell, Inc. Hussmann Corporation Hydro Aluminum Adrian, Inc. Indiana Tube Corporation Invensys Climate Controls Lennox International, Inc. Modine Manufacturing Co. Parker Hannifin Corporation Peerless of America, Inc. The Trane Company Thermo King Corporation Valeo, Inc. Visteon Automotive Systems Whirlpool Corporation Wolverine Tube, Inc. York International, Inc.
For additional information:
Air Conditioning & Refrigeration Center Mechanical & Industrial Engineering Dept. University of Illinois 1206 West Green Street Urbana,IL 61801
217 3333115
Abstract
MINIMIZING TEWI BY CHARGE REDUCTION IN A COMPACT CHILLER
A simulation model was developed to investigate strategies for reducing total equivalent
warming impact (TEWI) in compact water chillers. The focus was on minimizing R-410A
refrigerant charge while increasing efficiency. Compact flat plate heat exchangers with
refrigerant channels similar in scale to microchannels (Dh= 0.7 mm and 0.8 mm for the
condenser and evaporator, respectively) appear capable of reducing total system charge about
80% compared to conventional air-air split systems. Results are also compared to those obtained
for highly efficient air-to-air unitary systems, in which minimum-TEWI design strategies require
larger heat exchangers having greater charge. Overall the two approaches achieve comparable
reductions in global warming impacts; the chiller depends more on reducing direct emissions,
compared to unitary systems' dependence on reducing indirect emissions through use of flat
mUlti-port tubes with folded fins. These results are tentative, because the simulations did not
include detailed analysis of possible opportunities for improving the chiller technology by
optimizing the air and water pumping requirements in the secondary loop. The primary benefit
of the chiller technology, relative to air-air unitary, appears to lie in its compatibility with the use
of toxic or flammable refrigerants.
Table of Contents
List of Tables ............................................................................................................. iii List of Figures ........................................................................................................... iv Nomenclature ............................................................................................................ v 1. Introduction ........................................................................................................... 1 2. Baseli ne systems ................................................................................................ 2
2. 1 Air-to-air split systems .................................................................................... 2
2.2 Compact hermetic chiller ............................................................................... 4 3. Minimum-TEWI chiller ........................................................................................ 8 4. Tradeoffs near the optimum .......................................................................... 15
4. 1 Standard operating conditions .................................................................... 15 4.2 Increased Tevap ............................................................................ .................... 16 4.3 Effect of water loop pressure drop ............................................................. 17
5. Conclusions ........................................................................................................ 20 6. References ............................................................................................................ 22 Appendix A. Baseline systems .......................................................................... 25
A. 1 Baseline residential split system ................................................................ 25 A.2. Baseline chillers ............................................................................................ 26
A.2.1 Geometry of commercially-available plate heat exchangers .................. 26 A.2.2 Correlation selection for plate heat exchangers ..................................... 27
A. 2. 2. 1 Single-phase correlations ....................................................................... 29 A.2.2.2 Evaporation ............................................................................................ 32 A.2.2.3 Condensation ......................................................................................... 35
A.2.3 Typical CBE simulation results ................................................................. 37 Appendix B. Details of ideal chiller optimization ........................................ 41
B. 1 Model development. ...................................................................................... 41 B.1.1 Design operating conditions ..................................................................... 41 B.1.2 Assumed model inputs .............................................................................. 41 B.1.3 Search variables ......................................................................................... 44 B.1.4 Model correlations ...................................................................................... 44
B.2 Chiller optimization .................. ...................................................................... 45
11
List of Tables
Table 2.1 Split system power comparison ...................................................................................... 3 Table 2.2 Hermetic chiller model inputs ......................................................................................... 5 Table 2.3 CBE results ...................................................................................................................... 6 Table 3.1 Correlations ..................................................................................................................... 8 Table 3.2 System comparison with standard inputs ...................................................................... 12 Table 4.1 Optimal chiller with higher Tevap ................................................................................ 18 Table 4.2 Tradeoffs when ~Pwloop=O .......................................................................................... 19 Table A.l Alc split system operating conditions at design point.. ................................................ 26 Table A.2 SWEP plate geometry .................................................................................................. 28 Table A.3 Correlation range .......................................................................................................... 31 Table A.4 Split systemlCBE model comparison ........................................................................... 39 Table B.l Design operating conditions ......................................................................................... 41 Table B.2 Model inputs ................................................................................................................. 42 Table B.3 Search variable constraints ........................................................................................... 44 Table B.4 Model comparison ........................................................................................................ 49
iii
List of Figures
Figure 2.1 Split system TEWI comparison ..................................................................................... 3 Figure 2.2 Split system/CBE TEWI comparison ............................................................................ 6 Figure 3.1 Condenser temperature profile of optimized chiller .................................................... 10 Figure 3.2 Evaporator temperature profile of optimized chiller ................................................... 11 Figure 3.3 Power requirements ..................................................................................................... 11 Figure 3.4 Charge comparison ...................................................................................................... 13 Figure 3.5 TEWI comparison ........................................................................................................ 13 Figure 3.6 Minimum-TEWI condenser ......................................................................................... 14 Figure 4.1 Minimum-charge at different COPs ............................................................................ 15 Figure 4.2 TEWI tradeoffs at standard inputs ............................................................................... 16 Figure 5.1 TEWI comparison for all options ................................................................................ 21 Figure A.l Alc split system charge distribution [glkW cooling capacity] .................................... 25 Figure A2 Typical brazed plate heat exchanger. .......................................................................... 26 Figure A.3 Plate spacing ............................................................................................................... 28 Figure A.4 Refrigerant-side velocity comparison ......................................................................... 28 Figure A5 Water-side velocity comparison ................................................................................. 29 Figure A.6 Single phase heat transfer correlations ....................................................................... 30 Figure A7 Single phase pressure drop correlations ...................................................................... 32 Figure A8 Evaporation heat transfer correlation .......................................................................... 33 Figure A9 Evaporation pressure drop correlation ........................................................................ 34 Figure A10 Condensation heat transfer correlation ..................................................................... 35 Figure A.ll Condensation pressure drop correlation .................................................................... 36 Figure A12 Split system and CBE power .................................................................................... 38 Figure A13 Split system and CBE charge distribution ................................................................ 38 Figure A14 Split system and TEWI comparison ........................................................................ .40 Figure B.l Chiller loop layout (10.6 kW unit) ............................................................................. .43 Figure B.2 Heat exchanger layout for current investigation ......................................................... 45 Figure B.3 Split system and optimum chiller power. .................................................................... 46 Figure B.4 Energy-charge comparison ........................................................................................ 47 Figure B.5 Split sytem and optimized chiller charge distribution ................................................. 48 Figure B.6 Split system and optimized chiller TEWI comparison .............................................. .48
iv
Nomenclature
A area m2
b plate spacing mm Cp specific heat kJ/kg-K D diameter mm Dh hydraulic diameter mm f Darcy friction factor fQ heat flux assumption for charge inventory g gravitational constant mls2
G mass flux per channel kglm2-s h heat transfer coefficient W/m2-K
lfg enthalpy of vaporization kJ/kg k thermal conductivity W/m-K KH Hughmark flow parameter L length m LMTD log-mean temperature difference °C m mass g ill mass flow rate kgls Np number of plates Perit critical pressure kPa q heat flow rate kW " heat flux kW/m2 q
T temperature °C V overall heat transfer coefficient W/m2-K
VA conductance W/K V ve1cocity mls
V volumetric flow rate m 3/s w plate width mm
VI power W Wg refrigerant gas density weighting factor x vapor quality
Dimensionless Groups "
Bo boiling number Bo=-q-G· i fg
Fr Froude number 2· !1P/pV~x
Nu Nusselt number hD/k
Pr Prandtl number IlCp/k
v
Re
Greek Symbols a ~ ~p
~T
11 pump
11s
~
P
Subscripts 2ph amb avg blower bulk calc cond c, cold comp dis eq evap fan h, hot In
I liq 10 loop out r sub suct sup tot w x
Reynolds number
Lockhart Martinelli parameter
void fraction corrugation angle, from vertical pressure drop temperature change water pump efficiency isentropic compressor efficiency viscosity density
two-phase ambien average evaporator blower, or sum of all fans in cold water loop pertaining to the entire bulk flow calculated value condenser cold water loop compressor discharge line equivalent evaporator condenser fan hot water loop inlet liquid property liquid line liquid only property water loops outlet refrigerant subcooled suction line superheated total water, water-side as a function of quality
vi
pVD/Jl
° kPa °C
kglm-s kglm3
1. Introduction
The Montreal Protocol mandated the phase-out of hydrochlorofluorocarbons (HCFCs),
requiring selection of new refrigerants and new technologies (Sands et al. 1997). To characterize
the global warming effects of new systems, the total equivalent warming impact (TEWI)
accounts for the release of refrigerant into the atmosphere (direct effects) and the release of
carbon dioxide from electricity generation (indirect effects).
The TEWI for a 10.6 kW residential air conditioner using R410A is about 10% direct and
90% indirect (Kirkwood and Bullard, 1999). Since the direct portion is a function of the amount
of charge and the loss rate, a 45% reduction in charge could improve TEWI as much as a 5%
increase in COP. Many alternatives to HCFCs are not greenhouse gases, but may be either toxic
or hazardous (e.g. butane, propane, ammonia) and therefore it is necessary to minimize charge
when those refrigerants are used. This study uses a simulation model to investigate various
means of reducing TEWI, with the primary focus on strategies for minimizing charge without
decreasing COP.
Section 2 describes two residential-scale "baseline" systems from which improvements
can be measured. One is a conventional U.S.-style split system; the other is a hermetic chiller
utilizing compact brazed plate heat exchangers. The residential scale was chosen only to provide
a familiar starting point for the analyses which are normalized in per unit cooling capacity, in the
interest of generalizing the resulting insights across a broader range of unitary alc system and
chiller sizes.
Section 3 briefly describes the optimization process and explains the results of the TEWI
minimization. Section 4 describes the sensitivity analysis performed near the optimum with
respect to several model assumptions. Conclusions are summarized in Section 5.
Appendix A describes the baseline systems: conventional split system and compact
brazed plate system. Appendix B provides a detailed discussion of the flat plate model
development.
1
.. ~-
2. Baseli ne systems
2.1 Air-to-air split systems A conventional air-to-air residential split system uses copper tubes and aluminum fins to
transfer heat between the refrigerant and air. Typically the condenser and compressor located
outside the house require long liquid and suction lines (>7.5 m each) to the evaporator inside the
house. The heat exchangers and long liquid line account for nearly 90% of total charge
(Andrade and Bullard, 1999).
Kirkwood and Bullard (1999) explored the extent to which TEWI could be reduced in
systems with refrigerant-air heat exchangers by using microchannel heat exchangers. They
examined microchannels because of their compactness for a given heat transfer capacity and
pressure drop, compared to traditional round-tube/plate-fin heat exchangers. Their simulations
suggested that TEWI could be reduced by approximately 13% compared to a conventional
R410A system, at the ARI 210/240-B standard rating condition (26.7 °C DB indoor, 19.4 °C WB
indoor, 27.8 °C DB outdoor). This improvement was achieved by increasing COP (4.5 vs. 3.8)
and reducing charge (235 glkW vs. 258 glkW), thus decreasing both the indirect and direct
components ofTEWI (Figure 2.1).
Kirkwood's design decreased TEWI considerably, but was limited by several factors.
First, the search was limited to "off the shelf' microchannel tubes and other components.
Secondly, air has a high thermal resistance which requires large area, which in tum limits heat
exchanger charge-reduction strategies. Larger heat exchangers increase COP, but also require
additional charge, even with microchannel tubes. This explains the minimal charge reduction
shown in Figure 2.l. There is still some potential improvement in the microchannel design by
either developing new microchannel technologies, or decreasing liquid line length (10.8 m in
Kirkwood's simulations) to reduce total charge. Table 2.1 compares the systems, with all figures
normalized per kW of cooling capacity.
The details of the TEWI calculations are shown below:
TEWI = Indirect Effect + Direct Effect Indirect Effect = Power. Run Time. Mass CO 2
Where: Power = total electric power consumed by the unit [kWeJ
2
(2.1) (2.2)
Run Time = hours the unit runs per year [900 hr/yr, Illinois] Mass CO2 = mass of CO2 produced in electric generation [0.65 kgC02lkWhe]
Direct Effect = Charge. Loss Rate. GWP (2.3) Where:
Charge = total refrigerant charge of the system [kg] Loss Rate = rate of refrigerant leakage per year [assumed 4%/yr] GWP = global warming potential of refrigerant [1730 kgC02IkgR41OA]
The C02 emission rate for electricity generation and global warming potential was
obtained from Sand et ai. (1997). The refrigerant leakage rate and run time were obtained from
Kirkwood and Bullard (1999).
180~-----------------.
II Indi rect [J Oi rect 160 +------{!
~140+---
m ~ 120 +---
8. ~ 100
~ o 80+---
~ ...... 60+---
~ .- 40 +---
20+---
0+---
Conventional split system Microchannel split system
Figure 2.1 Split system TEWI comparison
T bl 21 S r a e iPllt s stem power com~anson Conventional Microchannel split system split system
W tot [WIkW cooling capacity] 256 221
W comp [WIkW] 198 187
WpUmPing [WIkW] 58 34
Wblower [WIkW] 40 17
Wcondfan [WIkW] 18 17
3
....
2.2 Compact hermetic chiller An alternative way to minimize the direct TEWI effect would be to minimize charge by
building a small chiller to take advantage of the compactness obtainable with refrigerant-to-water
heat exchange instead of refrigerant-to-air. Commercially available compact brazed plate heat
exchangers (CBEs) are used in a wide variety of applications (food processing, chemical reaction
processes and pharmaceutical industries). Due to their very compact nature, high surface
volume ratios, relatively low pressure drops, and their ability to utilize chevrons and bumps
imprinted on the plates, they rely more on heat transfer coefficient and less on area to transfer
heat.
Many studies have examined liquid-liquid heat transfer and pressure drop in CBEs:
Buonopane and Troupe (1969), Okada et al. (1972), Cooper (1974), Focke et al. (1985), Rortgen
(1988), Shah and Focke (1988), Roetzel et al. (1994), Yang and Rundle (1994), Bogaert and
Boles (1995), Talik et al. 1995, and Muley and Manglik (1999). However, only a few sources
have examined evaporation and condensation in CBEs: Panchal et al. (1983), Marvillet (1991),
Haseler and Butterworth (1995), Yan and Lin (1999), and Yan et al. (1999), with only the two
Yan studies providing correlations. The Yan correlations for two-phase heat transfer and
pressure drop were used in a simulation model with correlations from Shah and Focke for single
phase heat transfer and Focke et al. for single-phase pressure drop. These were chosen by
applying criteria described in Appendix A.
The compact chiller using CBEs was modeled according to the ARI 550/590 standard
rating condition for chillers (0.054 Us per kW at 29.4 DC inlet condenser water, 0.043 Us per
kW at 6.7 DC outlet evaporator water). It was assumed that the connecting line lengths could be
quite short, as shown in Table 2.2. To provide chilled water to every room, a total length of
150 m was assumed for the cold water pipes, while the hot water loop was assumed to be 50 m.
It was assumed that the fan power requirements would be identical to the split system outdoors
(19 W/kW) and halved indoors due to the absence of ductwork (20 W/kW). Superheat was set to
7 DC as recommended by SWEP (1992) for that company's compact brazed plate heat
exchangers. Subcooling was set equal to 2 DC in order to minimize the amount of liquid in the
condenser, while still ensuring full condensation. The compressor was assumed to have an
4
isentropic efficiency of 0.7 and UA of 15.8 W/K, identical to the scroll compressor simulated by
Andrade and Bullard (1999). Water pumping power was calculated by Equation 2.4, and pump
efficiency was assumed to be 0.6 (Hall, 2000).
. V(M> HX + ilP1oop ) W = water pump (2.4)
11 pump
The compressor power for the chiller is greater due mainly to the difference in standard
rating conditions for the split system versus the chiller. As shown in Table 2.3, the split system
condenses at 39 °e and evaporates at 9 °e. The eBE system with 60 plates in both the
condenser and evaporator (denoted 60x60) has average condensing and evaporating temperatures
of 37 °e and 5 °e, respectively. This increases compressor power to 208 W/kW compared to
198 W/kW for the conventional split system. The two water loops introduce 9 W/kW that did
not exist in the conventional split system. However, because indoor blower power was reduced
by half, the total power consumed by the chiller (255 W/kW) about equal to the conventional
split system (256 W/kW).
T bl 22 H h'll d I ' a e ermetIc c 1 er mo e mputs
variable value variable value
qevap 10.6 kW l1s,comp 0.7
Wblower 20W/kW UAcomp 15.8 W/K
Weondfan 18 W/kW Lsuction 0.5m
TWe,out 6.7°e ~ischarge 0.5m
TWc,in 29.4 °e Lliquid 0.2m
Vw•e 1.73 m3/hr Dsuction 10.7 mm
"w,e 2.17 m3/hr Ddischarge 12.7 mm
ilTsup 7°e D1iquid 3.0mm
ilTsub 2°e L1oop,cold 150m
Tamb 27.8°e L1oop,hot 50m
11 pump 0.6 D1oop,h&c 25.4 mm
5
Table 2.3 CBE results Conventional CBE split system (60 plates in both HX's)
W IOI [WIkW] 256 255
W comp [WIkW] 198 208
WpUmPing [W/kW] 58 47
Wblower [W/kW] 40 20
WCOrul fan [WIkW] 18 18
WCOldlOOP [WIkW] -- 5
WhOllOOP [WIkW] -- 4
m [glkW] 258 100
Tcond,avg [0C] 39 37
Tevap,avg [0C] 9 5
The compact design of the CBEs, and the shorter liquid line, assumed for the remotely
located hermetic packaged chiller, provide substantial charge reduction (100 glkW vs. 258
glkW). As a result, TEWI is 7% less for the CBE system than the conventional split system, but
still 7% higher than the microchannel split system (Figure 2.2).
180,---------------------------------.
160
~ 140
l 8. 120
~ 100
~ 80 (,)
~ ..... 60
~ 40
20
o Conventional split system
Microchannel split system
60XSO Plate CBE
Figure 2.2 Split systemlCBE TEWI comparison
6
....
As mentioned earlier, one of the advantages of eBEs is that they can be used in a wide
variety of applications, but they are used primarily in liquid-liquid heat exchange applications.
As such, the design is not necessarily optimized for refrigerant evaporation and condensation
heat transfer. Initial analyses indicated that the best way to minimize charge was by decreasing
plate spacing, increasing heat transfer at the same time, paying a slightly higher price in pressure
drop. The resulting ideal plate heat exchanger had plate geometries well outside those used to
develop the correlations (Appendix A). Therefore, the model was altered by using different
correlations before conducting the optimization for minimum TEWI (briefly described in
Section 3, described in detail in Appendix B).
7
" ,,"
3. Minimum-TEWI chiller
Initial analyses suggested that the optimal heat exchanger geometries were well outside
the range of channel aspect ratios used to develop the correlations for the chevron plate CBBs.
The CBB model was then modified by replacing the heat transfer and pressure drop correlations
as described in Table 3.1. These correlations were developed for flow in smooth tubes and
applied to rectangular channels, using the hydraulic diameter calculated according to Equation
2.5. The optimizations had heat flux and mass flux values within the range of the correlations
and are extrapolated only on diameter.
2·w·b Dh =--
w+b
Table 3.1 Correlations Conventional
Flat plate chiller CBE chiller
Heat transfer
Condensation Yan, Lio & Lin Dobson-Chato
Evaporation Yan&Lin Wattelet-Chato
Single-phase Shah & Focke Dittus-Boelter
Pressure drop
Condensation Yan, Lio & Lin de Souza-Pimenta
Evaporation Yan & Lin de Souza-Pimenta
Single-phase Focke ASHRAE
Charge Rice, with Hughmark void fraction
(2.5)
The optimization analysis favored smaller diameters, but previous experiments with
R410A in microchannels suggest that extrapolation errors are small (Stott et ai., 1999). The first
step was to maximize COP, within roundoff to two decimal places, by decreasing refrigerant
plate spacing and increasing the number of plates to decrease pressure drop. Then heat
exchanger geometry was adjusted to minimize charge for that specified value of COP, using
direct search and variable metric optimization algorithms. By maximizing COP first, the
compressor discharge pressure decreased and the suction pressure increased. This forced the
refrigerant outlet temperatures within 3.6 DC and 0.6 DC of the water inlet temperature for the
counterflow condenser and evaporator, respectively. To decrease charge at the high COP,
geometry was changed in a way that increased condenser pressure drop so that the refrigerant
8
....
outlet temperature was within 0.1 °C of the water inlet. As seen with the CBE chiller, the
biggest difference in compressor power is due to the different operating conditions, which were
specified by the standards.
The minimum-charge condenser has many long narrow plates (Np=200, L=3.8 m, w=1.9
mm) spaced closely together (br=O.4 mm) as shown in Table 3.2. The small refrigerant ports
create a mass flux in each channel 25% higher than the conventional split system. Due to the
narrow plate width, water-side plate spacing increases to 4.6 mm (vs. 1.6 mm for the CBE) in
order to decrease pressure drop. However, the result is a water mass flux per channel 3.9 times
higher than the conventional CBE. Since both water and refrigerant have high mass flux, overall
heat transfer coefficient is 138% higher than the conventional split system (3480 W/m2-K vs.
1600 W/m2-K). The high heat transfer coefficient allows for lower LMTD, 2.2°C (versus 5.6 °C
for the conventional split system) to decrease condensing temperature, as shown in Table 3.2.
The lower LMTD is obtained by decreasing the average condensing temperature to 34°C (at the
cost of increasing pressure drop) to have an approach temperature within 0.1 °C (Figure 3.1)
CBEs have many chevron bumps that restart the laminar boundary layer to create high
water heat transfer coefficients (-9500 W/m2-K) at low Reynolds numbers (-660). The
optimization analysis pointed towards geometries that lay outside the range of published
correlations for unsteady developing flow over chevron plates. Therefore, the current
investigation used correlations for smooth flat plates very closely spaced, relying on turbulent
flow (Re:::::2100) to achieve higher heat transfer coefficients (-4750 W/m2-K) than could be
obtained with lamilar flow between smooth plates. Even though the minimum-TEWI chiller has
a condenser 3.8 m long (compared to 0.4 m for the CBE) and higher Reynolds number, water
side pressure drop is 8.6 kPa, while the CBE water pressure drop is 12 kPa due to the chevron
bumps.
The minimum-charge evaporator is much more sensitive to refrigerant pressure drop than
the condenser. While the evaporator has many plates, slightly longer than standard CBEs
(Np=200, L=0.7 m), they are much shorter than the condenser plates. The evaporator also tended
toward the minimum plate spacing on the refrigerant side, 0.4 mm, but the plates are much wider
than the condenser, 19 mm, (narrower than standard CBE width of 71 mm). The small
9
"~'
refrigerant channels have lower mass flux per channel than conventional split systems (76
kglm2-s vs. 157 kglm2-s), but have a much greater surface area and smaller hydraulic diameter
(0.8 mm versus 9.2 mm) providing higher refrigerant heat transfer coefficients. By using water
instead of air and the higher refrigerant heat transfer coefficient, the optimal chiller evaporator
has a higher overall heat transfer coefficient than conventional split systems (1872 W/m2-K vs.
700 W/m2-K). The refrigerant-side heat transfer area is much higher than the conventional split
system (2.6 m2 vs. 1.1 m2) and therefore, due to high U, has a much lower LMTD (3.6 °C versus
11°C). The evaporating temperature could not increase above 5 °C due to the low water
temperature required for the secondary loop, and the specified 7 °C superheat. The evaporator
temperature profile of the minimum-TEWI chiller is shown in Figure 3.2.
Despite the lower evaporating temperature (5°C vs. 9°C), compressor power was only
1 WIkW higher than the conventional split system, due to its lower average condensing
temperature (34°C vs. 39°C). As shown in Figure 3.3, total pumping power for the optimal
chiller was 21 % lower than the conventional split system. Due to the massive reduction in
charge (82%, Figure 3.4) the minimum-TEWI chiller reduces total TEWI 13% compared to the
conventional split system (Figure 3.5).
60
55
50 0' o
;:' 45
40
35 k-refrigeram
~----30 water~
Length [m]
Figure 3.1 Condenser temperature profile of optimized chiller
10
"~'
.. ~'
10
~ 8
6 refrigerant
4~~~~~~~~~~~~~~~
o 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Length [m]
Figure 3.2 Evaporator temperature profile of optimized chiller
300~--------------------------------~
• Compressor [J Blower • Cond. Fan [J Cold Loop • Hot Loop
250 +----
'>: ... '1 200 +---
B ~ 150 +---
~ ......
1100 +--
a.
50+---
0+---
Conventional split system Mn-TEW I chiller
Figure 3.3 Power requirements
11
a e iystem companson WIt T bI 32 S
TEWI [kgC02lkW-year] Indirect TEWI Direct TEWI COP mtot [glkW]
Wcomp [W/kW]
WpUmPing [WlkW]
Wblower [WlkW]
Wcondfan [WlkW]
WCOldlOOP [WlkW]
WhotlOOP [WlkW]
Tavlt [0C]
AHT [m2]
LMTD [0C]
mass [glkW] D [mm]
~ w [mm] <Il
br [mm] s::: .g bw [mm] 0 u L [m]
No --# ref circuits APref [kPa]
APwcond [kPa] APw100D [kPa]
Tavg [0C]
AHT [m2] LMTD [0C]
mass [glkW] D [mm]
.... w [mm]
~ br [mm]
~ bw [mm] > ~ L [m]
N~ --# ref circuits
APref [kPa]
APwevao [kPa] APw100D [kPa]
- value tS at mmtmum search bound + value is at maximum search bound
Conventional split system
168 150 18
3.77 258
198
58
40
18
--
--39 1.6 5.6 144 9.1 ------
37.1 --
1.5 lO4 ----9
1.1 11.0 54 9.2 ------
6.3 --6
5.5 ----
12
....
stan ar mputs d d'
Min-TEWI chiller
147 144 3
3.96 47
199
46
20
18
5
3
34 1.7 2.2 9
0.7 1.9
0.4-4.8 3.8
200+ 99 355 8.6 29
5 2.6 3.6 8
0.8 19.0 0.4-0.6
0.69 200+
99 2.9 9.2 64
300.-----------------------------------, I!iI Condenser CI EloElporator • Uquid Une CI Suction Une • Discharge Une II Other
250+---
200 +---
I 150 +----1 ...... E
100 +---
50+---
0-/---
Conventional split system Win-TEWI chiller
Figure 3.4 Charge comparison
180.-----------------------------------, II Indirect CI Direct
160 +----------l
'i:' 140 +----m ~ 120 +----8-~ 100 +---
8 80+----
~ ...... 60+---i ~ 40+---
20+---
0+---
Conventional split system Min-TEWI chiller
Figure 3.5 TEWI comparison
The optimally sized condenser is actually very compact. Figure 3.6 shows a plate 3.8 m
long, but only about 2 mm thick and around 120 mm wide (allowing 0.2 mm channel dividers).
13
· .' "~'
Such a plate might be made of copper, steel or aluminum, and bent to fit into a more compact
package. Ends could be cut to accommodate brazed or welded fittings that would separate the
refrigerant and water channels at the inlets and outlets.
Figure 3.6 Minimum-TEWI condenser
14
,.'
4. Tradeoffs near the optimum
4. 1 Standard operating conditions
The results presented in Section 3 led to a chiller geometry in which TEWI was
minimized at the maximum COP achievable at the given operating conditions. The simulations
indicated that another 28% reduction in charge could be obtained, at a cost of 7% reduction in
COP, as shown in Figure 4.1a. Figure 4.1a shows minimum charge tradeoffs at various COP
values, while Figure 4.1 b shows the associated plate width and length tradeoffs for the
evaporator and condenser.
As COP decreased from the maximum of 3.96, charge was minimized by decreasing
condenser length, with little change to condenser plate width, and by simultaneously decreasing
evaporator width, with little change to evaporator plate length, as shown in Figure 4.1b. For both
heat exchangers, the number of plates remained at the arbitrarily selected upper bound (200)
while the refrigerant-side plate spacing remained 0.4 mm. Water-side plate spacing decreased
from 4.8 mm to 2.1 mm for the condenser and from 0.6 mm to 0.4 mm for the evaporator. These
effects increase all components of power (compressor and both water pumps) by increasing
condensing temperature, decreasing evaporating temperature and increasing water-side pressure
drop for both heat exchangers.
a. 0 (.)
4.0,----------------_
...
3.9 r max COP
3.8
3.7
3.6
3.5 30 35 40 45 50
charge [glkW capacity]
a) COP-charge tradeoffs
4
3.5
3
2.5
:g:2 ....I
1.5
0.5
~ ¢ maxCOP , I
~ ! ~ I
~ ~ ~
8
t. Evaporator _. ~- . Condenser
max COP
~ • o+---~---~---,____--~
o 5 10 w[mm]
15
b) HX geometry tradeoffs
20
Figure 4.1 Minimum-charge at different COPs
15
· ...
Manufacturing and cost considerations may also limit realization of the COP-maximizing
design. For example, the 19 mm wide by 0.4 mm tall rectangular evaporator channel would
require thick wall to withstand the pressures of R41OA. As seen in Figure 4.1 b, the charge
minimizing design strategy allows for substantially narrower channels, but entails a COP
penalty. Similarly, the 3.7 m length of the optimal condenser could be a problem, but Figure
4.1 b shows how the length can be reduced as the maximum-COP constraint is relaxed.
Figure 4.2 shows TEWI for each of points in Figure 4.1 compared to the 60x60 plate
CBE and the conventional split system. Despite the lower COP, nearly every point has a lower
total TEWI than the CBE due to the drastic reduction in charge.
Of course, if the objective was not to minimize charge, plate lengths and widths could
probably be reduced with very little sacrifice of COP.
170~------------------------------------------~
'i:' co
160
~ 150 ... CI) Q.
3: 140
~130 ~ ...... 3: 120 W I-
110
100
II Indirect 0 Direct
- w
" 'a co en () Cij E Q) c 1ii 02
----_../ Y
;:; tn a:: c >. ~ en min-charge tradeoffs 0
co c x 8 0
co
Figure 4.2 TEWI tradeoffs at standard inputs
4.2 Increased Tevap
Due to the low water temperature in the evaporator (Twe.Dut= 6.7 DC) evaporating
temperature was 5.1 °C in the first stage of optimizations. In the previous optimizations, chilled
16
water returned at 12.2 °C as dictated by the ARI standard. The return temperature is well below
15.7 °C, the dewpoint of the bulk air specified in the ARI 21O/240-B air-conditioning rating
condition. In order to continue to allow for dehumidification in all the rooms of the house, the
return temperature was held constant at the specified 12.2 0c. The water outlet temperature was
then increased to 9 °C and water flow rate increased to still achieve the same amount of heat
transfer. In the earlier optimizations, the temperature of the refrigerant exiting the evaporator
could be increased no higher than 12.1 °C due to the specified 7°C superheat (SWEP, 1992). To
allow evaporating temperature to increase, superheat was arbitrarily decreased to 1°C,
simulating the impact of using some kind of liquid overfeed system or a low-side receiver.
The maximum COP with the higher water temperature was 4.38, achieved by evaporating
at 8.7 °C and condensing at 33.2 °C, as shown in Table 4.1. Compressor power decreased 20%
from the original optimization. However, the higher water flow rate increased water pumping
power from 5 to 21 W/kW cooling capacity. Water pressure drop through the evaporator was
decreased slightly by increasing plate width (to 22.2 mm) and water-side plate spacing (to 0.8
mm). The increased plate width required higher refrigerant flow rate to maintain a high heat
transfer coefficient, which in tum increased condenser plate width to decrease condenser
pressure drop. A more detailed analysis of the system is provided in Appendix A.
4.3 Effect of water loop pressure drop In the foregoing analyses, most of the water pumping power was needed to overcome ~P
in indoor and outdoor piping systems, as shown in Table 4.1. Values were selected arbitrarily
and held constant throughout the analysis, only the (relatively small) heat exchanger ~P'S
changed as different geometries were evaluated. The question arises: how sensitive are these
optimal heat exchanger designs to our assumptions about indoor and outdoor pipe lengths and
diameters? To answer this question, we examined an extreme case where there is no pressure
drop in the water lines, and pumping power is only needed to pump water to overcome pressure
drop in the condenser and evaporator. To determine how the minimum-TEWI system would be
designed under this assumption, pressure drop in both liquid lines was set to zero. The first
column of Table 4.2 lists the optimized chiller from Section 3. Next, the water loop pressure
17
a e 'pttma c 1 er WIt Igl er T bl 4 1 O' I h'll . hh' h T
COP
mlot [glkW]
Woomp [W/kW]
Wpumping [W/kW]
Wbtower [W/kW]
W rood fan [W/kW]
WroidlOOP [W/kW]
WhotlOOP [W/kW]
Tav2 [0C]
w [mm]
br [mm] ~ bw [mm] rIl t:: .g L [m] 0 No --U
LlPref [kPa]
LlPw cond [kPa]
LlPwlooo [kPa]
Tav2 [0C]
w [mm]
.... br [mm]
~ bw [mm] !5 L [m] ra- Np > --~
LlPref [kPa]
LlPwevao [kPa] LlPw loop [kPa] ..
- value IS at mmlmum search bound + value is at maximum search bound
Initial optimization (Section 3)
3.96 47
199
46
20
18
5
3
34 1.9
0.4-4.8 3.8
200+ 355 8.6 29
5 19.0 0.4-0.6
0.69 200+ 2.9 9.2 64
evap Min-TEWI (T evao=9°C)
4.38 77
160
63
20
18
21
3
33 2.7 0.4-4.0 5.0
200+ 245 6.6 29
8.7 22.2 0.4-0.8 1.2
200+ 3.3 8.7 166
drops were set to 0 kPa. Finally, COP was re-maximized (within 2 decimal places) and charge
was then minimized. The last column of Table 4.2 shows the results of these new optimizations.
The new minimum-TEWI system decreases total TEWI by only 0.1 kgC02lkW per year
compared to the system just before re-optimization. The reoptimization resulted in heat
exchanger dimensions nearly identical to those in the initial optimization. Therefore, the chiller
that was optimized with the standard inputs has virtually the same TEWI as the chiller that was
optimized without water loop pressure drop. This is an important observation as the water pipe
lengths and diameters were specified arbitrarily for the initial optimization. Therefore, if these
18
..•.
"' "" -,'
values are incorrect, it would only affect the power consumption, but not the optimal strategy for
designing of the heat exchangers.
Table 4 2 Tradeoffs when ~Pw -0 loop-
TEWI [kgC02lkW-year] Indirect TEWI DirectTEWI COP
mtot [glkW]
Wcomp [WlkW]
Ww.COld [WlkW]
Ww •hot [WlkW]
Tav2 [0C]
m [glkW] W [mm]
~ br [mm] '" s:: .g bw [mm] s::
L [m] 0 U
Np --Mlref [kPa]
Mlwcond [kPa]
Tav~ [0C]
m [glkW] W [mm] ...
~ br [mm] 0 bw [mm] g.
L [m] > ~
No --Mlref [kPa]
Mlwevao [kPa] - value is at minimum search bound + value is at maximum search bound
Initial Re-optimized optimization results (Section 3) (MlWlooo= 0)
147 142 144 138 3 4
3.96 4.10 47 56
199 198
5 <0.1
3 <0.1
33.9 33.8 9.4 10.6 1.9 1.9 0.4- 0.4-4.8 5.1 3.8 4.0
200+ 200+ 355 358 8.6 8.0
5.09 5.10 8.3 15.3 19.0 41 0.4- 0.4-0.6 0.4-0.69 0.58 200+ 200+ 2.9 0.5 9.2 7.5
19
5. Conclusions Previous attempts to reduce TEWI have focused on increasing COP while slightly
decreasing refrigerant charge. Since many potential HCFC replacement refrigerants, like
hydrocarbons and ammonia, are either toxic or hazardous, this investigation approached TEWI
reduction through charge minimization by examining a compact chiller loop with secondary
water loops. The following are several key conclusions of this study.
.,' .....
1. Under standard chiller rating operating conditions, a system using commercially available compact brazed heat exchangers (CBEs) requires 61 % less charge than a conventional residential split system. However, due to the lower evaporating temperature necessary to supply water at the ARI standard chiller rating condition, compressor power is 5% higher. It was assumed that the sum of all indoor fan powers could be reduced to half of the blower power required for the conventional split system, due to the absence of ductwork. As a result, total energy consumption is slightly less than the conventional air/air split system, resulting in a decrease of 7% of total TEWI.
2. The minimum-TEWI chiller with flat plate heat exchangers had refrigerant-side plate spacing reduced to 0.4 mm in the evaporator to increase heat transfer coefficient and 200 plates (99 refrigerant circuits) decrease pressure drop. The plates were 59% longer than a conventional CBE (0.7 m vs. 0.4 m) but were 73% narrower (19 mm vs. 71 mm). The minimum-TEWI condenser also had 200 plates and refrigerant-side spacing of 0.4 mm. However, the condenser is not as sensitive to pressure drop, and therefore the plates were 3.7 m long and only 1.9 mm wide.
3. The minimum-TEWI chiller was able to reduce charge 82% compared to conventional split systems. Despite a lower saturation temperature of the compressor discharge pressure (40 DC to 38 DC for the split system and optimal chiller, respectively), compressor power was 0.5% higher. This is because of the difference in rating conditions; chillers require a lower evaporating temperature. However, since the indoor blower power was smaller, due to the absence of ducts, total power was 4% lower than the split system, resulting in 12% less total TEWI.
4. Charge could be reduced by another 28% at the standard operating conditions, at a 7% reduction in COP, compared the minimum-TEWI chiller design. Since TEWI of the optimal chiller is more than 99% due to energy use, total TEWI would increase nearly 7%, but still be less than the conventional split system. This option might be advantageous when using toxic or flammable refrigerants, or for manufacturing reasons because thick walls would be required for an evaporator channel that was 19 mm wide by 0.4 mm high. The additional charge reduction would also decrease the "optimal" condenser length from 3.8 m to 1.4 m.
5. Total TEWI could be decreased an additional 8% compared the "optimal" chiller by increasing evaporating temperature. This could be achieved by simultaneously decreasing superheat (requiring a low-side receiver) and increasing the cold water
20
delivery temperature from 6.7 °C to 9°C (requiring higher mass flow rate and therefore increased pumping power).
6. The optimization strategy is not very sensitive to the water loop assumptions. Using the heat exchanger geometry from the "minimum-TEWI" chiller, eliminating the water pressure drop decreased TEWI 3% by reducing the total pumping power (air and water) 20%. Re-optimizing the heat exchangers had virtually no effect on TEWI.
7. At standard operating conditions the minimum-TEWI chiller reduced TEWI 12% compared to conventional split systems, no improvement over a minimum-TEWI microchannel split system, as shown in Figure 5.1. However, if evaporating temperature were increased, or water pipe lengths decreased (or some combination), then the compact chiller loop appears to offer the most potential for reducing TEWI.
180.-------------------------------------------------.
160 'i:' ! 140
~ 120 Do
~ 100
CJ 80
~ 60
i w 40 I-
20
o
.Indirect Cl Direct
Qi E CD Iii en ~ c: "Iii c: i c: CD a::w :.E ~ Iii :.E ~~ olD
0 '6 ~
0 ~
§~ <oC) § c: § § l-x 0 .£: 0 W 0 w
&l w .- Co <0 t-;- t-;- t-;- Cl ::E UI c: ~ c: ~ c: ;S
~ ~ s ~
Figure 5.1 TEWI comparison for all options
21
· "~'
6. References
Andrade, M.A and C.W. Bullard, "Controlling Indoor Humidity Using Variable-Speed Compressors and Blowers." University of Illinois at Urbana-Champaign, ACRC TR-151, 1999.
American Society of Heating, Refrigeration and Air-conditioning Engineers, "Handbook of Fundamentals." ASH RAE, 1997.
ARI, 1989, Standard for Unitary Air-Conditioning and Air-Source Heat Pump Equipment, ARI-2101240. .
ARI, 1998, Water Chilling Packages Using the Vapor Compression Cycle, ARI-550/590.
Bogaert, R and A Bolcs, "Global Performance of a Prototype Brazed Plate Heat Exchanger in a Large Reynolds Number Range." Experimental Heat Transfer, vol. 8, pp 293-311, 1995.
Buonopane, RA and RA Troupe, "A Study of the Effects of Internal Rib and Channel Geometry in Rectangular Channels." AICHE Journal, vol. 15, no. 4, pp 585-596, 1969.
Cooper A, "Recover More Heat with Plate Heat Exchangers." The Chemical Engineer, pp 280-285, May 1974.
Dittus, F.W., and L.M.K. Boelter, University of California, Berkeley, Publications on Engineering, vol. 2, p. 443, 1930.
Dobson, M.K. and J.C. Chato, "Condensation in Smooth Horizontal Tubes." Journal of Heat Transfer, 120:2, pp. 193-213, 1998.
Focke W.W., J. Zachariades and I. Olivier, "The Effect of the Corrugation Inclination Angle on the Thermohydraulic Performance of Plate Heat Exchangers." International Journal of Heat and Mass Transfer, vol. 28, no. 8, pp. 1469-1479, 1985.
Hall, Scott. Grundfos Pumps Corporation, Fresno, CA Personal communication. July 6,2000.
Haseler, L.E. and D. Butterworth, "Boiling in Compact Heat Exchangers/ Industrial Practice and Problems." Proceedings of Convective Flow Boiling, Banff, Alberta, Canada, pp. 57-70, 1995
Kirkwood, AC. and C.W. Bullard, "Modeling, Design, and Testing of a Microchannel SplitSystem Air Conditioner." University of Illinois at Urbana-Champaign, ACRC TR-149, 1999.
Marvillet, Ch. "Welded Plate Heat Exchangers as Refrigerants Dry-Ex Evaporators." EUROTHERM Seminar No. 18 (Design and Operation of Heat Exchangers), Germany, pp 255-268, 1992.
Muley, A and RM. Manglik, "Experimental Study of Turbulent Flow Heat Transfer and Pressure Drop in a Plate Heat Exchanger with Chevron Plates." Journal of Heat Transfer, vol. 121, pp. 110-117, February 1999.
22
.. ~-
· .. ~.
Okada, K, M. Ono, T. Tomimura, T. Okuma, H. Kono and S. Ohtani, "Design and Heat Transfer Characteristics of New Plate Heat Exchanger." Heat Transfer-Japanese Research, vol. 1, no. 1, pp. 90-95, January-March 1972.
Panchal, C.B, D.L. Hillis, and A. Thomas, "Convective Boiling of Ammonia and Freon 22 in Plate Heat Exchangers." ASMElJSME Thermal Engineering Joint Conference, Hawaii, ASME Book IOOI58-B, vol. 2, pp. 261-268, 1983.
Rice, C.K "The Effect of Void Fraction Correlation and Heat Flux Assumption on Refrigerant Charge Inventory Predictions." ASH RAE Transactions, vol. 93, part 1, pp. 341-367, 1987.
Roetzel, W., KD. Sarit, and X. Luo, "Measurement of the Heat Transfer Coefficient in Plate Heat Exchangers Using a Temperature Oscillation Technique." International Journal of Heat and Mass Transfer, vol. 37, suppl. 1, pp. 325-331, 1994.
Rortgen, H.G., "Mathematical Modeling of Heat Transfer in Plate Heat Exchangers Using the Finite Element Method." Wiirme-und Stof:fiibertragung, vol 23, pp. 353-364, 1988.
Sand, J.R, S.K Fischer, V.D. Baxter, "Energy and Global Warming Impacts of HFC Refrigerants and Emerging Technologies." A report sponsored by the Alternative Fluorocarbons Environmental Acceptability Study (AFEAS) and the U.S. Department of Energy. Oak Ridge National Laboratory, Oak Ridge, Tennessee. 1997.
Shah, RK and W. W. Focke, "Plate Heat Exchangers and Their Design Theory." Heat Transfer Equipment Design, Hemisphere. Washington, D.C. pp. 227-254,1988.
de Souza, A.L. and M.M. Pimenta, "Prediction of Pressure Drop During Horizontal Two-Phase Flow of Pure and Mixed Refrigerants." ASME Conf. Cavitation and Multiphase Flow, S. Carolina, FED Vol. 210, pp. 161-71, 1995.
Stott, S.L, C.W. Bullard, and W.E. Dunn, "Experimental Analysis of a Minimum-TEWI Air Conditioner Prototype." University of Illinois at Urbana-Champaign, ACRC CR-21, 1999.
SWEP Refrigeration Inc., Compact Brazed Heat Exchangers for Refrigerant Applications: A Technical Handbookfrom SWEP. Bohemia, NY, 1992.
Talik, A.c., L.W. Swanson, L.S. Fletcher, and N.K Anand, "Heat Transfer and Pressure Drop Characteristics of a Plate Heat Exchanger." ASMElJSME Thermal Engineering Conference, vol. 4, pp 321-329, 1995.
Wattelet, J.P., J.C. Chato, A.L. Souza, and B.R Christoffersen, "Evaporative Characteristics of R-12, R-134a, and MP-39 at Low Mass Fluxes," ASH RAE Transactions, vol. 100, no. 1, pp. 603-615, 1994.
Yan, Y.Y. and T.P. Lin, "Evaporation Heat Transfer and Pressure Drop of Refrigerant R-134a in a Plate Heat Exchanger." Journal of Heat Transfer, vol. 121, pp. 118-127, 1999.
23
Yan, Y.Y., H.C. Lio and T.F. Lin, "Condensation Heat Transfer and Pressure Drop of Refrigerant R -134a in a Plate Heat Exchanger." International Journal of Heat and Mass Transfer, vol. 42, pp. 993-1006, 1999.
Yang, W.J. and D. Rundle, "Optimized Thermal Design of Plate and Spiral Type Heat Exchangers." ASME Heat Transfer Equipment, HTD-Vol. 282,1994.
24
, ... '
Appendix A. Baseline systems
A. 1 Baseline residential split system The simulation model was examined first to determine how charge is distributed
throughout a residential alc system (Andrade and Bullard 1999). The system used R410A and
had conventional copper tube/aluminum fin heat exchangers. As indicated in Figure A.l, the
condenser, evaporator and liquid line contain the greatest amounts of charge at the ARI 210/240-
B rating condition. Therefore, charge-reduction strategies focused on maximizing refrigerant
side surface-to-volume ratios in the heat exchangers, while minimizing lengths of connecting
lines. The 'other' category includes 29 glkW (of cooling capacity) charge in the accumulator,
compressor and refrigerant dissolved in oil, which are assumed to be unchanged for the case of a
chiller.
Figure A.l Alc split system charge distribution [glkW cooling capacity]
The conventional alc split system operated with evaporating and condensing temperatures
of 9° and 39°C, respectively. Total system energy use was calculated as follows, to facilitate
comparison with other alternatives. Condenser fan power was measured, and blower power was
set to the default value (365 W /1000 cfm) associated with the standard test procedure. The
compressor isentropic efficiency was 0.7 at this rating condition, so the same value was assumed
to apply to other systems (chillers) operating at their rating conditions.
25
c SPJIt sys em operatmg con 1 10 Table A.1 N l' t d't" ns at design point
variable value
qevap 10.6 kW
W comp 2051 W
Wblower 438 W
Wfan 196 W
Tevao 9°C
Tcond 39°C
~Tsup 5.6°C
~Tsub 8.7°C
A.2. Baseline chillers A.2.1 Geometry of commercially-available plate heat exchangers
Typical compact brazed plate heat exchangers (eBE) have chevron corrugations stamped
into the plates as shown in Figure A.2. For ease of manufacturing, many plates of the same
imprint design are stacked on top of one another creating many channels of the same width and
corrugation depth. While exact dimensions vary by manufacturer and model, most have w:b
ratios greater than 40, so hydraulic diameter approaches twice the corrugation depth (Dh,plate=
2·b).
Sec A-A /3=0°
Sec A-A /3=90°
~ ~
Figure A.2 Typical brazed plate heat exchanger
26
A.2.2 Correlation selection for plate heat exchangers
One of the major tasks of creating a system simulation model was to determine how to
model the plate heat exchangers. There is much published literature available for liquid-liquid
heat transfer in brazed plate heat exchangers (CBEs), but the openly available correlations for
condensation and evaporation is very limited. A computer program created by SWEP (1996),
simulating their actual CBEs, was run multiple times to provide a baseline to help select which
correlation to use.
The SWEP program includes many variables such as refrigerant and water Reynolds
numbers, film coefficients, pressure drops and property data. However, not all the parameters in
the heat transfer and pressure drop correlations are provided so they had to be estimated using a
"data set" created by running the program over a wide range of heat exchanger operating
conditions. The SWEP calculation procedure determines Re as
Re = mchannel ·2 I-lbulk • w
(Al)
The SWEP program reports Reynolds number, mass flowrate and viscosity. The
refrigerant-side Reynolds number is based on J.1v. Therefore, from each of the program runs it
was possible to use Eq. Al to determine the internal plate width (w). These numbers were then
averaged for each CBE simulated, including both water and refrigerant sides, with results
provided in Table A.2. The average values are all within 2mm the value listed for the external
plate width as provided on the SWEP dimension sheet.
The plate pressing depth (b) is required for all correlations, as the hydraulic diameter is
defined as twice the pressing depth (Dh=2·b). The value of b for each CBE is not explicitly
provided by SWEP so it must be calculated from other variables. The plate corrugations are
sinusoidal in shape, thus the pressing depth is the amplitude of the sine wave. Since the average
value of a sine wave is its amplitude, a sinusoidal channel of amplitude b would have the same
volume as two flat plates (of the same width and length) separated by a distance b, as indicated
in Figure A.3. This allows for the calculation of b to be determined by Eq. A2 since all other
variables are now known (Vol. and L from SWEP dimension sheet, w from Eq AI). The
average value of b for each geometry is provided in Table A2.
27
Vol channel = w· b· L (A. 2)
To check the results of the calculation of plate spacing, Eq. B.3 was used to calculate
channel velocity. The SWEP program output includes both refrigerant and water channel
velocities, rounded to one decimal place. Figures A.4 and A.5 show how the calculated values
compare to the SWEP values for the refrigerant and water-side velocities. On each graph, the
two lines indicate the bounds of where the calculated values should lie to fall within the rounding
to one decimal place.
mchannel = p . VChannel • W • b
~ ==> ---=-b.=X __ _ ~
Figure A.3 Plate spacing
a e PJ a e geomerry T bl A 2 SWEP Itt
w [m] b [mm] L [m] Vol. [cm3]
calculated Eq. A.l Eq.A.2
dimension dimension by
B15
B25
B45
.!.!
0.071
0.115
0.242
2
1.8
1.6
1.4
o 615
C 625
A 645
3 1.2 >
1
0.8
0.6
0.4
0.2
1.65
1.75
1.69
~ i
sheet
0.432
0.479
0.559
li C
oAC
o~
o~~~~~~~~~~~~
o 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2
Vswep
sheet
51.0
96.3
186.9
Figure A.4 Refrigerant-side velocity comparison
28
(A.3)
.. ~.
0.6
o 615
0.5 D 625
l>. 645
0.4
u
>B 0.3
0.2 .M---'rounding bounds
0.1
0.1 0.2 0.3 0.4 0.5 0.6
VSwep
Figure A.5 Water-side velocity comparison
At lower velocities, the calculated values match reasonably well to the SWEP data within
the one decimal place roundoff. At higher velocities Eq. A.3 tends to overpredict the SWEP
velocity. This most likely indicates that the average plate spacing is slightly greater than the
value obtained from Eq. A.2. Therefore, either there is a problem with the assumption of using
an average value of plate spacing for Eq. A.3, or the actual length of the channel is less than the
port-to-port value reported on the SWEP dimension sheet. However since the differences in
velocity are slight, the correlation should still closely match the SWEP data. Since most of the
heat transfer and pressure drop correlations are in the forms of equations A.4 and A.5, if there are
several good candidates, the one chosen should slightly overpredict h and ~P to account for the
actual pressing depth being larger than the calculated value.
A. 2. 2. 1 Single-phase correlations
A.2.2.1.1 Heat transfer Twelve different correlations were tested for each of the geometries and compared to
(A.4)
(A.5)
values from the SWEP program. Table A.3 shows the geometry and testing range of the twelve
29
correlations. Figure A.6 compares the 12 candidate correlations, using them to predict
performance of the B15 CBE parameter values from Table A.2. The procedure was repeated for
the other CBEs, with only slight differences in h, indicating that any slight difference between
actual and calculated plate spacing should not affect correlation selection. The dashed lines
indicate that the correlation has been extrapolated outside the Reynolds number range reported in
the correlation.
12000
10000
¥ (\Ie 8000
~ 6000
~
4000
2000
-{]- hS&T
--1:r hsogaert
-* heooper
~hFocke
~hMarriot
--¢-hMuley
-e-hOkada
_ .• - hRoetzel
-.-hRortgen
o Swep (all PHEs) -H-~alik
--- extrapolated ~hYan
outside Re.range -il-hY&R
o~~~~--~~~~--~~~~--~~~
o 200 400 600 800 1000 1200 1400
Rew
Figure A.6 Single phase heat transfer correlations
The correlation most closely matching the SWEP data is the Buonopane and Troupe
(B&T), however the correlation was developed for Re>3000. Therefore, of the correlations
within the proper Re range, the Marriott correlation (reported in Shah and Focke, 1988) is the
one that best corresponds to the SWEP data set and is provided in Eq A.6.
{0.729 Rel!3 Pr l/3
Nu . = Marnott 0.380 Re2l3 Pr l!3
Re ::; 7
Re>7
30
(A.6)
a e orre atlon range T bl A 3 C I'
Correlation
SWEPdata
Buonopane and Troupe (B&T)
Bogaert
Cooper
Focke
Marriott (from Shah and Focke)
Muley
Okada
Roetzel
Rortgen
Talik
Yan and Lin
Yang and Rundle (Y&R)
NIP: not provIded N/C: no correlation
geometry
Corrugated chevron plate
Triangular herringbone plate
Corrugated chevron plate (13=68°)
half corrugated plate
Corrugated chevron plate (8:-60°) Corrugated chevron plate (13=60°)
Corrugated chevron plate (13=60°) Corrugated chevron plate (8=60°) Corrugated chevron plate (13=70°)
plate (details not provided)
Corrugated chevron plate (13=60°)
Corrugated chevron plate (13=60°)
Corrugated plate
A2.2.1.2 Pressure drop
w [m] b
[mm]
varies varies
NIP NIP
0.113 2.0
NIP 2.5
NIP 5.0
0.354 2.9
0.163 2.5
0.100 8.5
0.071 2.0
0.260 3.6
0.346 2.3
0.120 2.9
0.163 3.2
L [m] Re range
for hand M>
varies Re < 1350
NIP 3000 < Re < 30,000
0.236 h:all M>: Re< 10
NIP NIP
NIP h: 20 < Re < 16,000 M>: 90 < Re < 16.000
0.904 all
0.392 600 < Re < 10,000
0.358 h: 500 < Re < 10,000
M>: N/C
0.177 h: 400 < Re < 2,000
M>: N/C
NIP h: NIP
M>: NIC
0.946 1500 < Re < 6000
0.450 h: 250 < Re < 2250
M>: N/C
0.406 Re>400
Some of the sources that provided heat transfer correlations did not include pressure drop
correlations too. Therefore, there are only 8 options to choose from, instead of 12. The above
procedure was repeated to compare correlation pressure drops and again there was no
appreciable difference in the values when simulating the various geometries. Therefore, Figure
A7 shows a comparison of all values calculated by the SWEP program to obtained using the
correlations when simulating the B15 geometry. Dashed lines indicate when the correlation has
been extrapolated outside the Reynolds number range used in its development. The Focke
correlation most closely matches the SWEP data and therefore is selected to be used in the
charge optimization program (Equations A7 and A8).
31
35
30
25
Ii D.. ~ 20 ~
D.. <l
15
10
5
o Swep (all PHEs)
._-- extrapolated
outside Re range
/ /
/
I
Ii /
Jioa / 0
//
A
I I
I I
I
p. /
I I
" }. -M- ~P cooper
"""*'" ~P Marriot
-e- ~PTalik
",/ e---e ~£1~~~i=f;~~:~:;-:;e1:-=::;,:=~.e!.~--=---~e~-~-....I--=e:"'--..J..-_---1.._,--J -.-~P Y &R (off chart)
0 0 200 400 600 800 1 000 1200 1400
Rew
Figure A.7 Single phase pressure drop correlations
A.2.2.2 Evaporation
A.2.2.2.1 Heat transfer
f = {5.03 + 755/Re 26.8 Re·O.209
Re<400
Re~400
There is not much discussion of evaporation in plate heat exchangers in the open
(A7)
(A8)
literature. Making things more difficult is the fact that most of the literature available on
evaporation in eBEs does not actually provide a correlation (Panchal 1983, MarviIIet 1991 and
Haseler 1995). Figure A8 shows a comparison of the SWEP values to the one correlation
available extrapolated into the SWEP data range (Yan and Lin, February, 1999) plotted against
the vapor Reynolds number. While the correlation was developed for R134a, it was applied to
R410A to match the SWEP data set runs. For the calculations, values such as liquid Prandtl
number, boiling number, and densities could be determined from the SWEP output program.
32
These values were used to calculate an integrated average two-phase heat transfer coefficient
from Yan's correlation, developed for vapor Reynolds numbers above 25,000. The evaporation
heat transfer correlation is given in Equations A.9 through A. 13.
9000 open-Swep
closed-Yan 8000 ••
~~ I!.
o t.. ~oD S2' 7000 <fI m.~ /J. t..
N I 00 t.. fIt. ~ .E
t.. Lbt..i~ tlJCbo 0
== 6000 h -...... o Ijj cP ~oA .c:
t.. Co ••• I.c:N IQ o ., -. --5000 .. 4.. 0 .1 o 815
4000 .- o 825
• t.. 845 I-
3000 2000 6000 10000 14000 18000
Revapor
Figure A.8 Evaporation heat transfer correlation
While the correlation does not match very well, especially at lower vapor Reynolds
numbers, the lack of choices forces the selection of this correlation. In the current model, the
vapor Reynolds number is between 4500 and 7000, depending on the number of plates.
(A. 10)
N = 1 926 Pr 113 Bo 0.3 Re 0.5 [(1 - x) + X(ELJO.5] uevap • I eq I Pg
(A.ll)
(A. 12)
33
" ,."
(A. 13)
A.2.2.2.2 Pressure Drop The above procedure was repeated for the pressure drop correlation provided by Yan
(Figure A.9). Again, at low Reynolds numbers there is a very large discrepancy between the
correlation and the SWEP data, especially considering that the SWEP program provides the total
pressure drop, and the Yan correlation is used for the two-phase zone only. However the lack of
options forces the selection of this correlation. However, since heat transfer is underpredicted
and pressure drop is overpredicted by the Yan correlations, the COP of an actual system will
higher than the calculated value, discussed in A.2.3 Typical CBE simulation results. Yan 's
evaporation pressure drop correlation is provided in Equations B.14 through B.16, and is
integrated over the quality range.
.... I'll Q. JIC ..... Q. <I
40r---~--.---~---r--~---,,---r---.
35
30
25
20
15
10
5
open-Swep (~P evap)
closed-Van (~Pfric, 2ph)
...... rI' ~~. -. U
.~ •• II
• • dlto
~~%o oCto%-~
cto
.,. ... "If
-ill -4. [] 0
[]~o " ~
[]
8
o 815
[] 825
to 845
-
o~--~--~--~--~--~----~--~~ 2000 6000 10000 14000 18000
Revapor
Figure A.9 Evaporation pressure drop correlation
L 0 2
LlP =f ---evap,x x Db 2 Px
34
(A. 14)
· ...
fx = !6.947X105 Re~609 [(1- x) + x~I/Pg t.5]
31.21 Re~~.45443[(I - x)+ X~I/pJO.5]
A.2.2.3 Condensation
A.2.2.3.1 Heat transfer
Reeq < 6000
Reeq ~ 6000
(A15)
(A16)
There is even less literature available for condensation than for evaporation. Fortunately,
Yan, Lio and Lin studied condensation of R134a in CBEs (1999). The analysis from above was
repeated for condensation heat transfer of R410A with the results shown in Figure A 10. The
scale is smaller than for the evaporation comparison, so while it may appear that the correlation
does not match as well as, it is actually better for condensation. This correlation is provided in
Equations A17 and A18.
4000r-r-~~'-~~~u~~~~~~~r-r-~~ open-Swep 0 0 0
3500
3000
1500
closed-Van B DO 0 0
o 0 0
DO BOo
o
~oBo 0
o o
•• ~o .: •• .. .". y ...
• • • ••
o 815 o 825 l1 845
1000 2000 3000 4000 5000 6000 7000 8000
Rev
Figure A10 Condensation heat transfer correlation
35
h = f NUeondkl d eond D x
h
(A. 17)
(A. 18)
A.2.2.3.2 Pressure drop Figure A.II shows the correlation comparison for condensation pressure drop. This
correlation matches the SWEP data much better than did the evaporation correlation, and is
provided in Equations A.19 and A.21, and is integrated over the full quality range.
0.3 ~
open-Swep closed-Yan
0.25
0.2 0 o 0 • I • ! ~ 0.15 • d • n • 8
0.1 OOO~"OO , ~
fi • •• o 815 c;; .,. 0.05 , o 825 • /'. t. 845
0 0 1000 2000 3000 4000 5000 6000 7000 8000
Rev
Figure A.II Condensation pressure drop correlation
L G2
~Peond,x = feond ---,x D 2
h Px
( JO.8
f = 94 75 Re-O.5 BoO.5 Peond Re-O.0467 eond,x . 1 P . eq
ent
" Bo=-q-G· i fg
36
(A. 19)
(A. 20)
(A.21)
"~'
A.2.3 Typical CBE simulation results
To provide a basis for comparison with existing chillers, the model was adapted to
simulate a system characterized by the design variables and operating assumptions listed in
Tables B.l and B.2. Then the heat exchanger geometry was modified to match a typical eBE
(SWEP B15) for both the condenser and evaporator. To model the heat exchangers, correlations
were chosen as described in Section A.2.2.
Typical eBEs offer great flexibility by allowing the designer to select the number of
plates for the specific application. Most typical designs have between 10 and 80 plates, with the
maximum around 200. Analysis of eBEs examined multiple combinations of plates. Systems
with 40 plates in both the evaporator and condenser, as well as 60 in each are presented here.
The system with 40 plates in each had 79 glkW and a eop of 3.69 (although it is probably
higher due to the underprediction of hand overprediction of M>cond as discussed in Section
A.2.2). The compressor consumed 2266 W by condensing at 38°e and evaporating at 5°e.
Each heat exchanger area was 1.2 m2, compared to 1.6 m2 and 1.1 m2 for the conventional split
system condenser and evaporator respectively.
In order to increase eop, the number of plates for each heat exchanger was increased to
60. This system increased charge to 100 glkW and eop to 3.82 by reducing compressor power
to 2196 Wand decreasing water pumping power for the hot and cold loops by at least 10 W
each. Table A.4 compares the two eBE simulations to the conventional split system design.
The biggest difference between the split system and eBE system operation is the
evaporating temperature. The refrigerant-air heat copper tube/aluminum fin evaporator in the
split system evaporates at 9 °e, while the refrigerant-water compact corrugated brazed plate
evaporator evaporates around 5 °e in order to cool every room in the house. Despite the higher
energy consumption for the 60 plate eBE system, total TEWI is slightly lower due to the 61 %
reduction in charge. Figures B.ll through B.13 compare the split system to the two eBE
systems.
37
....
. "~.
300 • Compressor [JBlower • Cond. Fan [JCold Loop • Hot Loop
250
'>: -"1200
B ~15O
III ;= ..... ... ; 100 0 a.
50
0
Com.entional split 60x60 Plate CBE 4Ox40 Plate CBE system
Figure A.12 Split system and CBE power
300 mCondenser [J E ISporator • Liquid Line [J Suction Line • Discharge Line • Other
250
200
{150 ..... E
100
50
0
Conventional split 60x60 Plate CBE 40x40 Plate CBE system
Figure A. 13 Split system and CBE charge distribution
38
a e ipllt syste T bI A 4 S r m1CBE d I mo e companson
ale split system CBE CBE
(40x40 plates) (60x60 plates) TEWI [kgC02/year ]
Indirect TEWI Direct TEWI COP mtot [glkW]
Wcomp [W/kW]
Wblower [W/kW]
Wfan [W/kW]
WW,COld [W/kW]
Ww.h01 [W/kW]
Tavg [0C]
AHT [m2]
mass [glkW]
D [mm]
w [mm] ... br [mm] <!) <Il
bw [mm] s:: <!)
"0 L [m] s:: 0 u No --
# ref circuits
~ref [kPa]
~wcond [kPa]
~wlooo [kPa] LMTD [0C]
Tavg [0C]
AHT [m2]
mass [glkW]
D [mm] w [mm]
... br [mm] 0 .... "" bw [mm] ... 0 0.. L [m] "" ;> ~ No --
# ref circuits
~ [kPa]
~wevao [kPa]
~wloop [kPa] LMTD [0C]
.. - value IS at minImum search bound + value is at maximum search bound
168 159 156 150 154 149 18 5 7
3.77 3.69 3.82 258 79 100
198 215 208
40 20 20
18 18 18
-- 6 5
-- 5 4
39 38 37 1.6 1.2 1.8 144 30 43 9.1 3.3 3.3 -- 71.4 71.4 -- 1.6 1.6 -- 1.6 1.6
37.1 0.43 0.43
-- 40 60 1.5 19 29
104 3.6 2 -- 25 12 -- 29 29
5.6 6.0 5.5
9 5 5.2
1.6 1.2 1.8 57 19 27
9.2 3.3 3.3
-- 71.4 71.4 -- 1.6 1.6
-- 1.6 1.6 6.3 0.43 0.43 -- 40 60 6 19 29
5.5 14.1 9.8 -- 25 11 -- 64 64
11.0 3.6 3.4
39
180,--------------------------------------,
160
'i:' 140
m ~ 120 CI) Q.
~ 100
(.) 80
~ ...... 60 3: UJ I- 40
20
o Conl.entional split 60x60 Plate CBE 40x40 Plate CBE
system
Figure A.14 Split system and TEWI comparison
40
Appendix B. Details of ideal chiller optimization
B. 1 Model development B.1.1 Design operating conditions
The model simulated a system with 10.6 W of cooling capacity while operating at the
ARI 550/590 standard test condition for water temperatures and flowrates (Table B.l). Compact
brazed heat exchanger (CBE) manufacturers recommend 6-8 °C of superheat to minimize
adverse effect of maldistribution on TXV operation (SWEP 1992). Therefore, in the
optimizations, superheat was specified as 7°C. To minimize the amount of liquid in the
condenser while ensuring full condensation, the amount of subcooling after the liquid line was
set equal to 2 °C at this standard rating condition.
Tabl BID . e estgn operatmg con d' 'ons ttl
variable value
qevap 10.6 kW
TWe•out 6.7°C
TWc.in 29.4 °C
"w.e 1.73 m3/hr
"we 2.17 m3/hr
dT,up 7°C
dT,ub 2°C
Tamb 27°C
B.1.2 Assumed model inputs
Since a compact water chiller used for home air conditioning would require secondary
loops, the fan power required for the each loop was considered in the COP calculation. It was
assumed that the required air flow rates would be the same as those of the traditional 10.6 W
split system (Table B.2). Due to the absence of ductwork, however, the air side pressure drops
would be lower, so the fan power requirements were assumed to identical outdoors, but indoors
were halved. Sensitivity analysis also examined the optimal system when blower power was
equal to the split system's (40 W/kW). This did not affect the geometry of the optimal heat
exchangers, only the total energy consumption, which affected COP and TEWI. Therefore, all
results presented are for the condition where blower power was 20 W/kW.
41
"~'
The additional water pumping power was calculated by Equation B.l, and assuming
TJpump= 0.60 (Hall 2000). The water-side pressure drop included the pressure drop across the heat
exchangers and the remainder of the water loop. It was assumed the cold water loop would be at
least 100 meters to distribute the chilled water to every room in the house. To account for the
pressure drop across the individual heat exchangers the entire loop was modeled as 25.4 mm ID
pipe 150 meters long. The hot water loop can be much shorter than the cold water loop, however
there would be water-side pressure drop across the air-water heat exchanger outside. Therefore,
to account for the tubing and heat exchanger, the hot water loop was simulated as 50 m long 25.4
mm ID pipe.
V(dP w + dPlOOP ) Wpump = ------'--
TJpump
(B.l)
To calculate compressor power, an isentropic compressor efficiency of 0.7 was assumed
for all cases, the same value calculated by Andrade for the ale split system at its standard test
condition. To calculate the heat rejected from the compressor, VA was set to 15.8 W/K. This
value was taken from Andrade, for the same 10.6 W compressor with R410A used in the
conventional split system.
T bl B 2 M d I . t a e o e mpu s variable value
Wrru].c 18 W/kW
"'blower,e 20 W/kW
11 pump 0.6
l1s,comp 0.7
UA"omp 15.8 W/K
Lsuction 0.5 m
Ldischarge 0.5 m
Lliquid 0.2 m
Dsuction 10.7 mm
Ddischarge 12.7 mm
Dliquid 3.0 mm
Lloop.cold 150 m
Lloop,hot 50 m
Dloop.h&c 0.025 m
42
It should be noted that the alc split system compressor utilized 198 W/kW, plus 58 W/kW
for fan and blower, for a total power of 256 W/kW. In contrast, the chiller will use 20 W/kW
less for the blower, but it will need about 8-9 W/kW for water pumping. Since the standard
design conditions for a chiller will cause its compressor to operate at different suction and
discharge pressures than the conventional split system. The objective of the current investigation
is to determine chiller heat exchanger geometry to decrease total power and charge to decrease
TEWI.
Figure B.l shows one possible layout of the refrigerant loop. To reduce system charge,
connecting line lengths should be minimized and were estimated accordingly. The values shown
in Figure B.l seemed reasonable. The compressor suction and discharge lines contain only
vapor and are assumed to be quite short (0.5m each) but longer tubes would have very little
impact on charge. When standard diameter tubes are assumed, the pressure drop in each line is
around lkPa, which is negligible compared to 355 kPa for the condenser and 3 kPa for the
evaporator. Since pressure drops are negligible, the diameters of both vapor lines were
arbitrarily set to the same values as found in the alc split system (10.7 and 12.7 mm respectively
for the suction and discharge lines). There is more charge in the liquid line then either vapor
line, however, liquid line diameter minimally affects the optimal heat exchanger geometry.
Therefore, liquid line diameter was decreased to 3.0 mm to reduce the internal volume.
~,COld Ww,hot
Evaporator Condenser coil
Figure B.l Chiller loop layout (10.6 kW unit)
43
B.1.3 Search variables
Table B.3 lists the condenser and evaporator parameters that were allowed to vary during
the search for minimum-charge configurations. Commercially available brazed plate heat
exchangers have plate spacings (b) between 1.5 and 3 mm. However, microchannel heat
exchangers have channel sizes of 0.4 mm or less, implying that it could conceivably be possible
to achieve plate spacings of similar magnitude. Therefore, refrigerant and water-side plate
spacings varied independently, with lower limits set at 0.4 mm. The lower limit for plate width
(w) was set to 0.4 mm as well, which would then create square microchannel ports.
There is a wide range of plate lengths (L) available commercially, with the shortest being
110 mm. To allow for even shorter plates, the lower limit was set to 50 mm. At least 4 plates
are necessary for plate heat exchangers (allowing for one refrigerant channel and two water
channels) but many current designs use 10 to 80 plates. Again, to increase the search range, the
maximum number of plates (Np) was limited to 200, corresponding to 99 refrigerant channels
and 100 water channels.
Table B.3 Search variable constraints lower bound upper bound
search standard variable standard
search CBEs CBEs
0.4 1.4 br [mm] 2.9 --0.4 1.4 bw [mm] 2.9 --0.4 71.4 w [mm] 350 --
50.0 110 L [mm] 900 --4 10 Np 80 200
B.1.4 Model correlations
The heat exchangers were modeled as having many parallel channels of alternating
refrigerant and water as shown in Figure B.2. As geometry was allowed to vary considerably, it
became obvious that the optimally designed heat exchangers had geometries lying well outside
the applicable range of typical heat transfer and pressure drop correlations for chevron plate (See
Appendix B). Therefore, the heat exchangers were modeled simply as smooth flat plates using
the Dittus-Boelter (1940) correlation for single-phase heat transfer, and a Darcy friction factor
for pressure drop in the single-phase regions (ASHRAE, 1997). Correlations from Wattelet et al.
44
'.~.
(1994) and Dobson et at. (1994) were used for evaporation and condensation heat transfer,
respectively, and the de Souza et al (1995) correlation calculated two-phase pressure drop. All
of these correlations were developed for flow in smooth tubes, and applied to rectangular
channels using the hydraulic diameter, calculated according to Equation A.2. The optimizations
had heat flux and mass flux values within the range of the correlations and are only extrapolated
on diameter.
2·w·b Dh =--
w+b
1E---w--~
Figure B.2 Heat exchanger layout for current investigation
B.2 Chiller optimization
(A.2)
The alc split system has refrigerant-air heat exchangers allowing for evaporating and
condensing temperatures around 9 and 39 °e, respectively. Air exiting the evaporator then is
routed throughout the ductwork of the house, requiring high blower power (40 W IkW). The
current investigation has compact refrigerant-water heat exchangers with cold water pumped
throughout the house and passing through several water-air heat exchangers (one in each room).
The total length of the chilled water loop was assumed to be 150 m and was considered in the
water pumping power calculation. It was also assumed that the sum of all the individual fan
powers was half the split system evaporator blower power (20 WIkW). The optimum chiller
45
....
evaporated at 5 DC and condensed at 34 DC. The reduced temperatures resulted in a compressor
power of 199 W/kW, versus 198 W/kW for the split system. However, the additional water
pumping power adds an extra 8 W/kW to the total power, while saving 20 W/kW with the
blower power (Figure B.3). As a result, COP increased from 3.77 to 3.96. Table BA provides a
detailed comparison of the different models.
COP could not be increased above 3.96 because with the specified 7 DC superheat and the
evaporating temperature of 5 DC, the evaporator outlet was 12.1 DC, within 0.1 DC of the cold
water inlet temperature. The average condensing temperature could not be lowered below 34 DC
because the pressure drop and subcooling resulted in a condenser outlet temperature of 29.5 DC,
0.1 DC above the hot water inlet temperature.
300 II!! Compressor [J Blower • Cond. Fan • Cold Loop [J Hot Loop
250
>: ... . ~ 200 Q.
B ~ 150 -CJ
3= ...... ... ~ 100 0 a.
50
0 ConlA3ntional split Min-TEWI chiller Min-TEWI chiller
system (high Tevap)
Figure B.3 Split system and optimum chiller power
One means of reducing total power is to increase the evaporating temperature by
increasing the chilled water outlet temperature. To determine these effects, the model was run
again by holding the return temperature constant at 12.2 DC, and pumping more water to achieve
the same amount of heat transfer while having an outlet temperature of 9 DC instead of 6.7 DC.
The maximum COP with the decreased water temperature glide was 4.38, instead of 3.96
for the initial optimizations. This was accomplished by increasing evaporating temperature from
46
· .,'
5.0 °C to 8.7 °C. However, refrigerant mass f10wrate increased slightly and condenser plate
width increased from 1.9 mm to 2.7 mm to decrease pressure drop. The wider plates decreased
heat transfer coefficient and forced the average condensing temperature to decrease to 33.2 °C
while increasing total area to 3.0 mZ to rely less on LMTD and U, and more on A (Table B.4).
The higher evaporating temperature decreases compressor power from 199 to 160 W/kW
(Figure B.3), however the increased water f10wrate in the evaporator increased the water
pumping power from 5 to 21 W/kW. The result is a system using the least energy per year,
while having more than twice as much charge as the optimum systems at the standard operating
conditions. However, as shown in Figure B.4, there are other options for the higher-temperature
water systems that do not decrease power as much as the optimum, but still offer reduced power
compared to the standard-input system.
245
240 Do Optimized chiller
Do ° Tw=9°C
235 Do Do closed-optimized Do open-tradeoffs 230 Do
'l:' Do ~225 ~ 3:: Do ~ 8 " ';:220 0 CI ... 0 ~ 215 0 w 0
210 ~ 0,
205 ° 0
° 200 • 195
0 20 40 60 80 100 charge [gJkW capacity]
Figure B.4 Energy-charge comparison
The optimized chiller contains 47 grams of refrigerant per kW cooling capacity,
compared to 258 glkW for the split system and 77 glkW for the water chiller with increased
water temperature. Figure B.5 compares the charge distribution for the split system to the
optimized chiller.
The direct portion of TEWI (refrigerant leakage) is only 3 kgCOz/kW per year for the
optimized chiller, compared to 18 kgCOz/kW per year for the split system and 5 kgCOz/kW per
year for the higher temperature chiller. Since the optimized chiller consumes a total of 246
47
W IkW versus 256 W for the alc split system, the indirect portion of TEWI (C02 emissions from
electricity generation) was 144 kgCOz/kW per year versus 150 kgC02lkW per year for the alc
system. Since the optimized high temperature water chiller consumes only 222 WlkW, its
indirect portion of TEWI is the lowest, at 130 kgC02lkW per year. Figure B.6 shows the TEWI
comparison for each design.
300
.. Condenser c Evaporator • Liquid Line c Suction Une • Discharge Une • Other
250
200
~ 150 Q ...... E
100
50
0 Com.entional split Min-TEWI chiller Min-TEWI chiller
system (high T ov ap)
Figure B.5 Split sytem and optimized chiller charge distribution
180,-------------------------------,
160
..::" 140 ca cu >-120 b Q.
~ 100
~ 80
~ i 60 w I- 40
20
o
• Indirect c Direct
Conventional split Min-TEWI chiller Min-TEWI chiller system (high T .. ap)
Figure B.6 Split system and optimized chiller TEWI comparison
48
a e o e T bI B 4 M d 1
TEWI [kgC02IkW-year] Indirect TEWI Direct TEWI COP mtot [glkW]
Wcomp [WIkW]
Vvblower [WIkW]
Vvran [WIkW]
VvW,COld [WIkW]
Wwhot [W/kW]
Tavg [DC]
AHT [m2]
mass [glkW] D [mm]
w [mm]
t> br [mm] '" bw [mm] s:: CI)
"0 L [m] s:: 0
U Np --# ref circuits
LlPref [kPa]
LlPwcond [kPa]
LlPw 1000 [kPa] LMTD [DC]
Tav2 [DC]
AHT [m2]
mass [glkW] D [mm]
w [mm] .... br [mm] ~ bw [mm] .... 0 ~ L [m] > ~ No --
# ref circuits
LlP ref [kPa]
LlPwevan [kPa] LlP w 1000 [kPa] LMTD [DC]
- value is at minimum search bound + value is at maximum search bound
Conventional split system
168 150 18
3.77 258
198
40
18
----39 1.6 152 9.1 ------
37.1 --1.5 104 ----
5.6
9 1.1 57 9.2 ------
6.3 --6
5.5 ----
11.0
49
, ..... .
companson min-TEWI Optimized
chiller (T evan=9°C) 147 135 144 130 3 5
3.96 4.38 47 77
199 160
20 20
18 18
5 21
3 3
34 33 1.7 3.0 9 18
0.7 0.7 1.9 2.7
0.4- 0.4-4.8 4.0 3.8 5.0
200+ 200+ 99 99
355 245 8.6 6.6 29 29 2.2 1.5
5 8.7 2.6 5.4 8 29
0.8 0.8 19.0 22.2 0.4- 0.4-0.6 0.8
0.69 1.2 200+ 200+
99 99 2.9 3.3 9.2 8.7 64 166 3.6 1.2