Loading...
HomeMy Public PortalAboutLevinson_et_al Applied Energy 88(2011)4343-4357 Contents lists available at ScienceDirect Energy Applied Energy ELSEVIER journal homepage: www.elsevier.com/locate/apenergy Potential benefits of solar reflective car shells: Cooler cabins, fuel savings and emission reductions Ronnen Levinson*, Heng Pan, George Ban-Weiss, Pablo Rosado, Riccardo Paolini 1, Hashem Akbari 2 Heat Island Group,Environmental Energy Technologies Division,Lawrence Berkeley National Laboratory, 1 Cyclotron Road,Berkeley,CA 94720,USA A R T I C L E I N F O A B S T R A C T Article history: Vehicle thermal loads and air conditioning ancillary loads are strongly influenced by the absorption of Received 15 February 2011 solar energy.The adoption of solar reflective coatings for opaque surfaces of the vehicle shell can decrease Received in revised form 29 April 2011 the"soak"temperature of the air in the cabin of a vehicle parked in the sun,potentially reducing the vehi- Accepted 2 May 2011 cle's ancillary load and improving its fuel economy by permitting the use of a smaller air conditioner.An Available online 29 July 2011 experimental comparison of otherwise identical black and silver compact sedans indicated that increas- ing the solar reflectance(p)of the car's shell by about 0.5 lowered the soak temperature of breath-level Keywords: air by about 5-6°C.Thermal analysis predicts that the air conditioning capacity required to cool the cabin Cool colored car air in the silver car to 25°C within 30 min is 13%less than that required in the black car.Assuming that Solar reflective shell Vehicle air conditioning potential reductions in AC capacity and engine ancillary load scale linearly with increase in shell solar Vehicle fuel economy reflectance,ADVISOR simulations of the SCO3 driving cycle indicate that substituting a typical cool-col- Vehicle emission reduction ored shell (p=0.35) for a black shell (p=0.05)would reduce fuel consumption by 0.12 L per 100 km ADVISOR (1.1%),increasing fuel economy by 0.10 km L-1 [0.24 mpg](1.1%).It would also decrease carbon dioxide (CO2)emissions by 2.7 g km-1 (1.1%), nitrogen oxide(NOx) emissions by 5.4 mg km-1 (0.44%), carbon monoxide (CO) emissions by 17 mg km-' (0.43%), and hydrocarbon (HC) emissions by 4.1 mg km-1 (0.37%).Selecting a typical white or silver shell(p=0.60)instead of a black shell would lower fuel con- sumption by 0.21 L per 100 km(1.9%),raising fuel economy by 0.19 km L-1 [0.44 mpg](2.0%).It would also decrease CO2 emissions by 4.9 g km-1(1.9%),NOx emissions by 9.9 mg km-1(0.80%),CO emissions by 31 mg km-1 (0.79%),and HC emissions by 7.4 mg km-1(0.67%).Our simulations may underestimate emission reductions because emissions in standardized driving cycles are typically lower than those in real-world driving. ©2011 Elsevier Ltd.All rights reserved. 1.Introduction cooling load lowers the required cooling capacity, reducing ancillary load, improving fuel economy, and decreasing tailpipe Over 95% of the cars and small trucks sold in California have emissions. air conditioning [1,2]. Use of air conditioning (AC) in cars has The current study focuses on the decrease in soak temperature, been estimated to increase carbon monoxide (CO) emissions reduction in AC capacity,and improvement in fuel economy attain- by 0.99 g km-1 (71%), increase nitrogen oxide (NO„) emissions able through the use of solar reflective shells.Here"shell"refers to by 0.12 g km-1 (81%), and reduce fuel economy by 2.0 km L-1 the opaque elements of the car's envelope, such as its roof and [4.6 mpg] (22%) [3]. Air conditioning is the major ancillary load doors. First,we experimentally characterize component tempera- for a light-duty vehicle. The AC is sized to cool the cabin air tures and cooling demands in a pair of otherwise identical dark from its "hot soak" condition (i.e., vehicle parked in the sun, and light colored vehicles, the former with low solar reflectance facing the equator, on a summer afternoon) to a comfortable and the latter with high solar reflectance. Second, we employ a quasi-steady temperature, such as 25°C. Reducing the peak thermal model to predict the AC capacity required to cool each vehicle to a comfortable final cabin air temperature. Third, we use the AVL ADVISOR vehicle simulation tool to estimate the * Corresponding author.Tel.:+1 510 486 7494. dependence on ancillary load of the fuel consumption and pollu- E-mail address:RML27 @cornell.edu(R.Levinson). tant emissions of a comparable prototype vehicle in various Stan- ' Present address: Department of Building Environment Science &Technology, dard drive cycles. Finally, we calculate the fuel savings and Politecnico di Milano,Milano,Italy. 2 Present address: Department of Building,Civil,and Environmental Engineering, emission reductions attainable by using a cool shell to reduce Concordia University,Montreal,Canada. ancillary load. 0306-2619/$-see front matter©2011 Elsevier Ltd.All rights reserved. doi:10.1016/j.apenergy.2011.05.006 4344 R Levinson et al./Applied Energy 88(2011)4343-4357 2.Literature review Energy Laboratory (NREL) concluded that transient CFD tools are best suited for this task[19]. Extensive research over the past two decades has focused on Bharathan et al.[20]provide a definitive overview of the models reducing air conditioner ancillary loads.Most studies consider ca- that have been adopted or developed by NREL to simulate environ- bin air temperature,AC cooling load,and/or occupant comfort. mental loads, thermal comfort, and AC fuel use. We summarize some findings below. 2.1. Technology performance 2.2.1.Environmental forcing Technologies to reduce AC load include solar reflective glazing, The GUI-driven MATLAB application Vehicle Solar Load Estima- solar reflective shells,ventilation,insulation and window shading tor(VSOLE),developed by NREL[21],calculates the solar radiation [4-8].Past research has identified the use of solar reflective glazing transmitted, absorbed, and reflected by glazing as a function of as an especially effective strategy for reducing cooling loads,since glazing properties and location,vehicle geometry and orientation, sunlight transmitted through glazing accounts for 70% of cabin time,and radiation source. heat gain in hot soak conditions [9]. For example, Rugh et al. [4] 2.2.2. Thermal modeling measured that solar reflective glazing in a Ford Explorer decreased The commercial CFD tool RadTherm can be used to simulate so- cabin air("breath")temperature by 2.7°C,lowered instrument pa lar heat load, interior and exterior convection, and conduction nel temperatures by 7.6°C,and reduced windshield temperatures through the envelope, while the commercial CFD tool Fluent can by 10.5°C.They also reported that this decrease in cabin air soak be used to simulate convective heat transfer and fluid flow in the temperature would permit an 11% reduction in AC compressor power.More generally,they estimated that AC compressor power cabin. could be decreased by about 4.1%per 1 °C reduction in cabin soak 2.2.3.AC performance temperature. Akabane et al.[10]experimentally determined that in a vehicle An NREL model uses transient analysis to optimize vehicle AC traveling 40 km h-1 (25 mph)on a hot day(outdoor air tempera-performance [22]. ture 38°C,horizontal solar irradiance 0.81 kW m-2),about 42%of the vehicle heat load resulted from transmission through the glaz 2.2.4.Fuel economy ing,with about 48%from conduction through the shell and about The AVL ADVISOR vehicle simulator originally developed by 10% from engine heat and air leaks. We note that load fractions NREL [23,24] can simulate the effect of vehicle ancillary load on may vary with window opacity and with the ratio of window area fuel consumption and pollutant emissions. to shell area. Rugh and Farrington [7] found that ventilation and window 2.2.5. Thermal comfort shading during soak can be effective in reducing AC loads. They NREL has applied two models from the University of California concluded that natural ventilation (achieved using appropriately at Berkeley-the Human Thermal Physiological Model and the Hu- placed inlets to allow for natural convection) can be almost as man Thermal Comfort Empirical Model-to evaluate thermal com effective as forced ventilation. fort in vehicles. Solar reflective shells have also been reported to reduce soak temperatures. Hoke and Greiner [6] used the RadTherm and 3.Theory UH3D modeling tools to simulate soak conditions for a sport utility vehicle (SUV) parked on a hot summer day in Phoenix,AZ. They 3.1. Cabin air temperature model concluded that each 0.1 increase in the solar reflectance p of the shell reduces the cabin air soak temperature by about 1 °C. For The cabin air heating rate,or rate at which the internal energy example,cabin air soak temperature in a vehicle with a white shell of the cabin air U(t)increases with time t,is (p=0.50)was predicted to be 4.6°C lower than that in a compara- dU dT ble vehicle with a black shell (p=0.05). Rugh and Farrington [7] d=mac„dt (1) measured for several vehicles the reduction in cabin air(breath) soak temperature versus increase in shell solar reflectance. where ma is the cabin air mass,cv is the specific heat of air at con- Increasing the shell solar reflectance of a Ford Explorer mid-size stant volume,and Taft)is the cabin air temperature.The cabin air is SUV by 0.44 lowered cabin air soak temperature by 2.1 °C assumed to be transparent to both sunlight and thermal radiation, (0.47°C reduction per 0.1 gain in shell solar reflectance), while but exchanges heat with the air conditioner and the cabin surface. increasing that of a Lincoln Navigator full-size SUV by 0.45 lowered If the air is well mixed, a simple model for the variation of cabin cabin air soak temperature by 5.6°C(1.2°C reduction per 0.1 gain air temperature with time is in shell solar reflectance). Increasing the solar reflectance of only dTa the roof of a Cadillac STS full-size sedan by 0.76 lowered cabin d=a[Tv(t)-Ta(t)]+fi[Ts(t)-Ta(t)], (2) air temperature by 1.2°C(0.15°C reduction per 0.1 gain in roof so- lar reflectance). where Tv(t)is the temperature of the air flowing into the cabin from A 3-D computational fluid dynamics (CFD) simulation by Han the AC vent, Ts(t) is the mean temperature of the cabin's surface, and Chen [8] estimated that increasing body insulation reduces and aand)3 are fitted constants. steady state thermal load,but raises air cabin temperature during As the cabin air is mechanically cooled, it may reach a quasi soaking and cooling. steady state in which Taft)asymptotically approaches a final value. In this condition, denoted by the superscript *, (dTa/dt)* 0 and thus 2.2.Modeling tools Ta Tv+a(TS-Ta)*. (3) Simulations of thermal load and thermal comfort in vehicles typically use either lumped-parameter models [5,11-13] or tran- We may need to lower the vent air temperature if the final cabin air sient CFD models [14-18]. A study by the National Renewable temperature Ta exceeds some design target Ta,such as 25°C.If the R.Levinson et al./Applied Energy 88(2011)4343-4357 4345 difference between the cabin surface temperature and the cabin air pollutant per unit distance traveled). If reductions in F and E are temperature (Ts—Ta) is insensitive to the vent air temperature Tv, each linearly proportional to reduction in P,then then AF H=yp 4 QH/COP (11) dTa ti 1. (4) dTv* That is,reducing Ty*by AT will lower Ta by approximately AT.Resiz- Table 1 General properties of test vehicles. ing the AC to yield a new vent air temperature Tv'(t)-Tv(t)—AT re- sults in a new cabin air temperature Ta(t)that can be computed by Make and model 2009 Honda Civic 4DR GX numerically integrating Cabin volume(m ) 2.57 Engine idle speed(RPM) 700 AC air flow rate(m3 s-1) 0.1 dr=a[T„(t)—Ta(t)]+l[TS(t)—Ta(t)] (5) Black Silver Odometer distance(mi)[km] 4300[6900] 6200[10,000] subject to the initial condition Ta(0)=Ta(0). Note that the second AC line high pressure(psi)[MPa] 165[1.13] 175[1.20] term on the right hand side of Eq. (5) is the same as that in Eq. AC line low pressure(psi)[MPa] 35[0.24] 40[0.28] (2)because we have assumed that T's(t)—Ta(t)=Ts(t)—Ta(t). 3.2.AC capacity model Table 2 Vehicle surface properties. In recirculation mode, the rates at which the original and re- sized air conditioners remove heat from the cabin air are Surface Area(m2) Solar reflectance Thermal emittance Roof 2.0 0.05(black) 0.83(black) gnc(t) =mcp[Ta(t)—Tv(t)] (6) 0.58(silver) 0.79(silver) Ceiling 2.0 0.41 n/a and Dashboard 0.6 0.06 n/a tf c(t) =thCp[Ta(t)—Tv(t)], (7) n/a Windshield 0.9 0.06 / ( Seat 2.4 0.38 respectively,where fit is the AC air mass flow rate and cp is the spe Door 3.0 0.11 n/a spe- cific heat of air at constant pressure.To meet peak cooling load,the capacity of the resized AC must be at least Q-max[gac(t)] =mcp max[Ta(t)—Tv(t)]. (8) (a) q 3 6 1 5 7 8 2 3.3.Fuel saving and emission reduction model �_ --r��;vr=-=~ 1.cabin air � � �[ �js 1 2. roof Consider two vehicles that differ only in shell solar reflectance p 1111 1,�jyl;�� ,d and required AC capacity Q The reduction in AC capacity attainable nil 3.windshield by substituting the high-reflectance shell (subscript "H”) for the t _I_. . 4.dashboard•low-reflectance shell(subscript"L")is (b) 2 —�—_ `5 5.ceiling 1 s r-will ‘11411--...- 11� 6.door AQH =QL—QH (9) 4,8 in+ ,�...:, _ 7.seat /.�-_'J~rte_,,,,,.• ,�� r and the reduction in vehicle ancillary power load P is i �� f _ �y~+, 8•vent air yr!�_� 41r APH =AQH/COP (10) i; / 6 7 where COP is the coefficient of performance of the AC system. Let F denote fuel consumption rate(volume of fuel per unit dis- Fig.2. Locations of the eight cabin temperature sensors(thermistors),shown in(a) tance traveled) and E represent pollutant emission rate (mass of top view and(b)side view. ■/.. i EMI ‘°111'141 i■ , ft i I ow- ME • '# Fig.1. Experimental vehicles parked facing south in Sacramento,CA on July 17,2010.Tower between vehicles(black car,solar reflectance 0.05,left; silver car,solar reflectance 0.58,right)supports a Davis Instruments Vantage Pro weather station(upper mount)and an Eppley Laboratory Precision Spectral Pyranometer(lower mount). 4346 R.Levinson et al./Applied Energy 88(2011)4343-4357 and and AEH =YEAQH/COP (12) where yF-dF/dP and yE-dEldP are constant coefficients. AEc=yEAQc/COP (15) (We will show that for the drive cycles simulated in this study, yE and yE are indeed nearly constant within the ancillary power load ranges considered. However, past studies have raised doubt respectively. about the degree to which standardized driving cycles represent vehicle emissions from real-world driving[25].Specifically,a dis- proportionate fraction of emissions occur from"off-cycle"driving Table 3 characterized by high speed and/or acceleration. Bevilacqua [3] Input parameters for the two vehicle prototypes simulated with ADVISOR.Simulation has shown that NO„ and CO emissions are almost doubled with results from these two cars were interpolated to match the 84 kW power rating of the Honda Civics used in our experiments. the operation of AC. Here the linearity assumption offers a very conservative estimate of pollutant reductions.) Prototype 1 Prototype 2 Finally,consider a cool colored vehicle(subscript"C")that dif- Vehicle type Compact Compact fers from the first two vehicles only in shell solar reflectance and Vehicle power rating(kW) 63 102 1601 required AC capacity.Since DOE-2 simulations indicate that reduc Vehicle mass(kg) Conventional Conventional configuration Conventional Conventional tion in a building's annual peak demand for cooling power is line- Fuel converter FC_S163_emis FC_SI102_emis any proportional to gain in roof solar reflectance[26],we assume that reduction in required AC capacity scales with increase in shell solar reflectance,such that Pc—PL Table 4 AQc=QL—Qc= x AQH• (13) US EPA[31]and ADVISOR fuel economy(mpg)estimates for the 2009 Honda Civic GX. PH-PL EPA sticker ADVISOR(0-4 kW ancillary load) It then follows that the rates of fuel savings and emission reduction attainable by substituting the cool colored shell for the low-reflec- City 24 31-27 tance shell are Highway 36 31-39 Combined city/highway 28 24-32 AFc=YFAQc/COP (14) (a) 50_ I - 100 (a) 100 20 45- —outside air temperature 90 90 black —outside air relative humidity —silver 40 _--= 80 v 80 black-silver 2 15 c.i- 35 -- �--J 70 T 70 2 30_-- - _ 60 :a E 60 C ' = 10 �- a) — co ii 25= 50 c E 50 - c ai £ 20 J 40 L 40 --- m 15 —, 30 30 0 cc 10- 20 m 20 - - 0 0 - 5= 10 10 - - i- 0 I I i I I i , 0 0 I�I II I I I -5 08:00 10:00 12:00 14:00 16:00 18:30 18:40 18:50 19:00 19:10 19:20 Local standard time(17 Jul 2010) Local standard time(16 Jul 2010) (b) 1200 ==== 3.0 (b) 100— a 20 —solar irradiance ===== --black ----= 90 i— —silver 1000 —wind speed ==== 2 5 -_ =om = of 80 �; black silver= 15 I ` 70 ■ — 800 �m ��mosomm 2.0 — _ _— v Fa 600 mm mimim ' 1.5 ig s 50 ■EE c wimm o co L 40 5 400 4 1.0 w 30 _ o R - o , • c N 200 ��� 0.5 > 20 i a - 0 —� 10 = 0 0.0 0 5 08:00 10:00 12:00 14:00 16:00 18:30 18:40 18:50 19:00 19:10 19:20 Local standard time(17 Jul 2010) Local standard time(16 Jul 2010) Fig. 3.Weather during soaking and cooling trials, including (a) outdoor air Fig.4. Comparisons of(a)cabin air temperature and(b)vent air temperature in temperature and humidity and(b)global horizontal solar irradiance and wind speed. each car during indoor HVAC calibration. R.Levinson et al./Applied Energy 88(2011)4343-4357 4347 4.Experiment(thermal study) (a) 100 . 90 _ =black cabin surface_ 4.1. Overview 90 - -cabin air _ 80 vent air 80 _ = cabin air-vent air= 70 A pair of otherwise identical light duty vehicles, one with a c3 70 _ cabin surface-cabin air 60 v black shell and the other with a silver shell, were instru- 2 60 - = T 50 CD mented with surface and air temperature sensors. AC perfor- c mance was calibrated with an indoor heating and cooling 50 = 40 trial. The vehicles were then parked outdoors on a sunny sum- a 40 = 30 mer day and subjected to a series of five soaking and cooling E 30 _ ` 20 0 trials. 1 20 = ` 6 le 1/4 10 10 - _,.L NEE I_ l 0 4.2. Vehicles 0 --- , , -10 Two 2009 Honda Civic 4DR GX compact sedans,one black and 08:00 10:00 12:00 14:00 16:00 one silver,were loaned by California's Department of General Ser- standard time(17 Jul 2010) vices(Fig. 1).Apart from shell color,the vehicles were essentially cm=identical, with only minor differences in odometer distance and (b) 100 cabin surface 90 AC line pressures(Table 1). 90 --- cabin air 80 il The air mass one global horizontal solar reflectance [27,28] of 80 vent air cabin surface-cabin 70 each exterior surface(roof)and interior surface(ceiling,dashboard, U 70 ca bin birLair-vent.air _ 60 windshield, seat and door) was measured with a solar spectrum v reflectometer (Devices & Services SSR-ER, version 6; Dallas, TX). a) 60 --_ / 50 The hemispherical thermal emittance of each roof and windshield is 50 t 40 c was measured with an emissometer(Devices&Services AE1; Dal- m a 40 / 30 las,TX). The solar reflectances of the black and silver roofs were E / ` o 0.05 and 0.58, respectively, while their thermal emittances were H 30 `1 �i 20 0.83 and 0.79(Table 2). 20 ` 6 i' ki 10 10 �'',`'— [ 0 (a) 90 0 -10 black 08:00 10:00 12:00 14:00 16:00 80 , standard time(17 Jul 2010) 70 C=== ■��� Fig.6. Cabin surface,cabin air and vent air temperatures measured during soaking U60 C IS and cooling trials in (a) the black car and (b) the silver car. Also shown are 3 50 = / differences between cabin surface and cabin air temperature and between cabin air and vent air temperature. L 30 – i =" —root 1 4.3.Instrumentation —dashboard 20 --ceiling windshield The roof,ceiling,dashboard,windshield,seat,door,vent air and 10 —seat cabin air temperatures in each car were measured with thermis- 0 C====== cabin air tors (Omega SA1-TH-44006-40-T [surfaces], Omega SA1-TH- 08:00 10:00 12:00 14:00 16:00 44006-120-T[air]; Stamford,CT)and recorded at 1 Hz with a por- Local standard time(17 Jul 2010) table data logger (Omega OM-DAQPRO-5300; Stamford, CT). The vent air thermistor was suspended in front of a central HVAC out- (b) 90 let,while the cabin air thermistor was suspended at breath level silver midway between the front seat headrests. Top and side views of 80 eases the eight temperature measurement points in each vehicle are 70 — l� shown in Fig.2. _ 77 Each vent air and cabin air thermistor was wrapped in alumi- 60 e� num foil(low solar absorptance; low thermal emittance)to mini- 50 VAI5 \_ mize both solar absorptance and radiative coupling to the cabin.3 Interior surface thermistors (ceiling, dashboard, windshield, seat, a 40 \ and door)were wrapped in foil and secured with clear adhesive tape. E 30 =root Clear tape over foil yields high solar reflectance and high thermal emittance, minimizing solar absorptance while retaining radiative 20 __� dashboard �■■■ ceiling ■■■■■■■■■•■ windshield 10 ∎∎∎ seat —door 3 Increasing the thermal emittance of the sensor by replacing the foil with white 0 cabin air paint(low solar absorptance,high thermal emittance)would tend to increase,rather 08:00 10:00 12:00 14:00 16:00 than decrease,the apparent cabin air temperature by radiatively coupling the sensor to warm cabin surfaces.To illustrate,we note that the windshield in the black car has Local standard time(17 Jul 2010) low solar absorptance and high thermal emittance(as would a white-coated sensor), but runs about 4°C warmer than the cabin air during the soak and about 13°C Fig.5. Roof,dashboard,ceiling,windshield,seat,door and cabin air temperatures warmer than the cabin air during cooldown.These elevated windshield temperatures measured during soaking and cooling trials in(a)the black car and(b)the silver car. result from radiative coupling to the hot dashboard and ceiling. 4348 R.Levinson et al./Applied Energy 88(2011)4343-4357 coupling to the cabin. Roof thermistors were affixed with reflec- 0.05 was used on the black roof(p=0.05),and a light-colored tape of tance-matched opaque adhesive tape.Black tape of solar reflectance solar reflectance 0.62 was used on the silver roof(p=0.58). (a) 100 --black �_I � I - - 50 (b) 100 _black I _ 50 90 —silver = 45 90 —silver 45 ci 80 —black-silver 40 v 80 —black-silver 40 70 35 U ` 70 35 v ' 60 = 30 d m• 60 30 a 2 _ a 50 25 d o. 50 25 40 m 40 - 20 g m 40 1 20 W c 30 _ 15 6 F 30 It 15 6 o w 20 10 E 20 (`___- 10 � 10 5 10 -I "- 5 08:00 10:00 12:00 14:00 16:00 08:00 10:00 12:00 14:00 16:00 Local standard time(17 Jul 2010) Local standard time(17 Jul 2010) (c) 100— 20 (d) 100-- -20 —black - --4- black T - P- 90 —silver-, r 18 v 90 silver f : 18 m 80 black-silver----- 16 - 80 black-silver 16 `c 70 / (III 14 v r 70 — — 14 -, r U co c R — ,r 0 60 / \ _� 12 m o 60 l 12 E 50 ay 10 m d 50 ` \_ 10 v -a 40 - - 8 g -a 40 — — 'V 8 w 0 30 - 6 0 12 30 6 0 20 4 c 20 ■ _ _ 4 ci 10 2 10 2 0 — b : : : : , : : : , : : , 0 0 0 08:00 10:00 12:00 14:00 16:00 08:00 10:00 12:00 14:00 16:00 Local standard time(17 Jul 2010) Local standard time(17 Jul 2010) (e) 100 black ======== 50 (f) 100 —bla 90 —silver - 45 90 —silver -- 45 cj 80 black-silver 40 t j 80 —black-silver - 40 - 70 - - 35 v L 70 .— - 35 v c. Y 60 30 " 60 — 30 d d 50 4171A 25 c a 50 25 L w 40 : 20 d 40 == 20 V 30= / 15 0 30 — 15 m• 20:- I 10 0 20 __� 10 10--1—1—. 5 10 ��_ 5 0 0 0 0 08:00 10:00 12:00 14:00 16:00 08:00 10:00 12:00 14:00 16:00 Local standard time(17 Jul 2010) Local standard time(17 Jul 2010) (g) 100 50 (h) 100 20 90 #—black 45 90 black i I i 18 —silver a silver ;i- 80 —black-silver - 40 ., 80 —black-silver 16 2 70 - 35 v • 70 _14 ci 1 60 — —�— 30 v m 60 r 12 - O 50 25 E 50 —- — _-10 v 8 - cu 40 20 c E 40 / %\,s (' 8 c 30=_� 15 c 30 6 20— ` 10 v 20 ` \ k"�4 14 114 10 — 5 10 �(� =2 08:00 10:00 12:00 14:00 16:00 08:00 10:00 12:00 14:00 16:00 Local standard time(17 Jul 2010) Local standard time(17 Jul 2010) Fig.7. Comparisons of(a)roof,(b)ceiling,(c)dashboard,(d)windshield,(e)seat,(f)door,(g)vent air and(h)cabin air temperatures measured during soaking and cooling trials. R.Levinson et al./Applied Energy 88(2011)4343-4357 4349 (a) 0.3 -black mmmirommimommii 0.25 (a) 50 ■ cabin air(black) ======i mi U 45 •• cabin air(silver) ■ cabin air: -silver �mmini�mioi�, _ y=0.70x-10.8 Y 0 2 -black .silver :mmi:�:iim 0.20 ca 40 • vent air(black) ■■■■■■■■ 2>0.99 ■_____i■_ c vent air(silver) ■■■■■■■■ R CD =■�===I 0 35 R 0.1 0.15 o -cabin air(fit) a) c..) 30 vent air(fit) EMME■r! "45 25 - I ■■■rte■ .0 0.0 0.10 c @ _ c ` � .�� L .. 3 20 .... to -0.1 0.05 15 :°■■■ _. ' ° 10 E vent air: U 0.2 J iiiliiiiiiiiii 0.00 E - y=0.56x-15.8 i- 5 R2=0.95 0.3 0.05 0 08:00 10:00 12:00 14:00 16:00 40 45 50 55 60 65 70 Local standard time(17 Jul 2010) Cabin air temperature after soaking(°C) (b) 4.0 _ 1.8 (b) 100 F after soaking black after soaking: black i • after soaking )(silver 3.5 - -i 1.6 v 90 9( ) -silver -� 0.90x+4.8 -black-silver - 1.4 d 80 •• after cooling(black) R2=0.87 L 3.0 = _ - - r after cooling(silver) _ - 1.2 � c� 70 �� m _ - v d after soaking(fit) p 2.5 = 1.0 £ 60 after cooling(fit); ■� ���° L a) c 2.0 - 0.8 °' 50 MMMMMM m °0 1 5 0.6 u 40 ■ . V � _ Q 1.0 _ 0.4 o N 30 -_, - C L;' 0.2 = 20 C 1 after cooling: 0.5 m ■ y=0.72x-5.1 - I - 0.0 c� 10 ■ R2-0.93 0.0 , , , , , ! , , -0.2 0 �C 08:00 10:00 12:00 14:00 16:00 40 45 50 55 60 65 70 Local standard time(17 Jul 2010) Cabin air temperature after soaking(°C) Fig.8. Comparisons of(a)cabin air heating rates measured during soaking and Fig.9.Variations with cabin air final soak temperature of(a)cabin air and vent air cooling trials and(b)AC cooling rates measured during cooling trials. final cooldown temperatures and(b)cabin surface final soak and final cooldown temperatures. Soak and cooldown intervals were approximately 60 and 30 min, respectively. A weather station (Davis Instruments VantagePro2; Hayward, CA) mounted between the vehicles at a height of 2 m recorded (Eppley Laboratory Precision Spectral Pyranometer; Newport, RI) 1 min averages of outside air temperature, relative humidity, to check the solar irradiance reported by the weather station's global horizontal solar irradiance, and wind speed (Fig. 1). Solar silicon radiometer.The first class pyranometer shared a datalogger irradiance was also measured with a first class pyranometer channel with the black car's vent air thermistor. During daytime Table 5 Cooling trial measurements.Temperature differences are black car-silver car. Cycle Cooll Cool2 Cool3 Cool4 Cool5 Start(LST) 09:32 11:02 12:33 14:02 15:32 End(LST) 09:59 11:32 13:01 14:32 16:00 Duration(min) 28 30 28 30 28 Mean outdoor air temperature(°C) 25.0 28.8 32.9 35.8 37.5 Mean solar irradiance(kW m-2) 0.83 0.98 1.00 0.86 0.64 Black cabin air temperature after soaking(°C) 47.7 55.9 61.8 64.4 63.6 Silver cabin air temperature after soaking(°C) 43.3 50.7 56.1 58.0 57.3 Cabin air temperature difference after soaking(°C) 4.4 5.2 5.7 6.4 6.4 Black cabin air temperature after cooling(SC) 22.1 27.7 32.0 343 33.7 Silver cabin air temperature after cooling(°C) 19.6 24.9 28.2 29.9 29.5 Cabin air temperature difference after cooling(CC) 2.5 2.8 3.8 4.4 4.2 Black cabin surface temperature after soaking(SC) 48.5 533 58.0 62.2 66.0 Silver cabin surface temperature after soaking(°C) 45.3 48.4 52.5 56.4 59.9 Cabin surface temperature difference after soaking(°C) 3.2 4.9 5.5 5.8 6.1 Black cabin surface temperature after cooling(°C) 28.7 33.7 383 40.9 42.7 Silver cabin surface temperature after cooling(CC) 27.4 30.5 34.2 36.5 38.3 Cabin surface temperature difference after cooling(°C) 1.3 3.2 4.1 4.4 4.4 Black vent air temperature after cooling(SC) 9.6 14.2 17.4 20.5 20.2 Silver vent air temperature after cooling(CC) 9.1 13.0 15.4 16.8 17.0 Vent air temperature difference after cooling(°C) 0.5 1.2 2.0 3.7 3.2 4350 R Levinson et al./Applied Energy 88(2011)4343-4357 trials,the shared channel recorded vent air temperature while the temperature in each vehicle to about 60°C in 16 min.Each cabin's vehicle was being cooled,and solar irradiance at other times. air temperature was then reduced to about 20°C after 20 min of maximum cooling (lowest HVAC temperature setting, top fan 4.4.AC calibration(16 July 2010) speed,recirculation mode).The cabin air and vent air temperatures during the cooling cycle in the black car were compared to those in On the evening of 16 July 2010,each vehicle was parked under a the silver car to verify that the AC systems performed similarly. carport to shield it from sunlight. All windows were closed. At 18:38 LST, maximum heating(highest HVAC temperature setting, top fan speed,recirculation mode)was used to raise the cabin air 4.5.Soaking and cooling(17 July 2010) At 08:00 LST on the following day(17 July 2010),the vehicles Table 6 were removed from the carport and parked outdoors,side by side, Characteristics of each cooling trial,including fit parameters a and/3;coefficient of facing due south(Fig. 1).All windows were closed. determination R2;measured final cabin air temperature T;;and AC cooling capacity Q The weather was warm and sunny,with the outside air temper- needed to attain a final cabin air temperature of 25 T. ature rising steadily from 21 °C at 08:00 LST to 38°C at 16:00 LST. Trial a(s-1) /1(s-1) R2 r;(°C) Q(kW) Global horizontal solar irradiance reached about 1.0 kW m-2 black_cooll 0.017 0.030 0.86 22.1 2.60 shortly after noon, and wind speed ranged from about 0.5 to black_cool2 0.012 0.024 0.90 27.7 3.05 1.3 m s-1(Fig.3).Solar irradiances measured with the silicon radi- black_coo13 0.010 0.023 0.88 32.0 3.66 ometer closely matched those measured with the first class black_cool4 0.012 0.024 0.93 34.3 3.83 black_cool5 0.015 0.022 0.93 33.7 3.64 pyranometer. silver_cooll 0.018 0.025 0.95 19.6 2.35 From 08:30 to 16:00 LST,each parked car was run through five silver_cool2 0.012 0.024 0.86 24.9 2.61 rounds of soaking and cooling in which an approximately 60 min silver_cool3 0.011 0.023 0.90 28.2 3.03 soak(HVAC off)was followed by about 30 min of maximum cool- silver_coo14 0.011 0.023 0.89 29.9 3.34 ing. The soaking and cooling intervals were closely synchronized silver cools 0.015 0.022 0.90 29.5 3.25 car-to-car. a 0.2 30 I_i-1-1 4_i,1--i_L O in' black -measured (a) U black r -- r _, - ✓ 0.1 fi o 25 - - 3 2 0.0 20 ri--- - a 15 7 L - 0.1 - E _ I a) Q I 4- •6 0.2 _ E -I.J ,R 10 -black_cooll(09:32-09:59 LST,28 min,25°C,0.83-k.1Whi c -- black_cool2(11:02-11:32 LST,30 min,29°C,0.98 kW/m')- 3 -0.3 - c 5 - -black_cool3(12:33-13:01 LST,28 min,33°C,1.00 kW/m')- m u - - ,. - Q - black_cool4(14:02-14:32 LST,30 min,36°C,0.86 kW/m')- v -_black_cool4(14:02-14:32 LST,30 min,36°C,0.86 kW/m1 _ @ _ -black_cool5(15:32-16:00 LST,28 min,37°C,0.64 kW/m') -0.4 . 1 . . i . . . . , , , , , , , , , , 1 , „ v 0 0 5 10 15 20 25 30 0 5 10 15 20 25 30 Cooling time(min) Cooling time(min) (b) silver rreasured O V 30 _silver MI- L_:_ .-,_:__ b 0.2 b N U• 0.1 0 25 ■iiinon■mm -- 28 u 0.0 15 20 - a' -0.1 15 ■■■■■■�■■■■■■■■■■■ a _ o ■■■ ■�■■ 1111■ ■■ C ! 6 - I ■■■ ■ ■■ ■■■■■-■■ a a -. ■■■■•■••■■••••■•■ 73 0.2 !I E 10 _ -silver_cooll(09:32-09:58 LST 27 min,25°C,0.83 kW/m')- C - 6 -- -silver_cool2(11:02-11:32 LST,30 min,29°C,0.98 kW/m')- -0.3 •c 5 silver_cool3(12:33-13:00 LST,28 min,33°C,1.00 kW/m')- 6 a silver_cool4(14:02-14:32 LST,30 min,36°C,0.86 kW/m°)- silver_cool4(14:02-14:32 LST,30 min,36°C,0.86 kW/m') V 0 silver cool5(15:32-15:59 LST,28 min,37°C,0.64 kW/m') -0.4 0 5 10 15 20 25 30 0 5 10 15 20 25 30 Cooling time(min) Cooling time(min) Fig. 10. Measured and fitted rates of change of cabin air temperature Fig.11. Reduction in cabin air temperature versus cooling time in each of five trials, versus cooling time in Trial 4, shown for (a) the black car and (b) the shown for(a)the black car and(b)the silver car.Cooling trial interval,duration, silver car. mean outside air temperature and mean solar irradiance are listed in parentheses. R.Levinson et al./Applied Energy 88(2011)4343-4357 4351 5.Simulations(fuel savings and emission reductions) Civic GX(84 kW).Note that while the Honda Civic GX is fueled by natural gas, our ADVISOR simulations represent an equal-power We used the vehicle simulation tool AVL ADVISOR (version vehicle fueled by gasoline. Table 3 presents additional details of 2004.04.09 SP1)to relate rates of fuel consumption,nitrogen oxide the ADVISOR simulations. (NO„)emission,carbon monoxide(CO)emission,and hydrocarbon We estimate CO2 emission reduction from fuel savings at the (HC) emission to ancillary power load. ADVISOR was first devel- rate of 2321 g CO2 per L of gasoline[30]. oped in November 1994 by the National Renewable Energy Labora- We compared ADVISOR fuel economy predictions to EPA gaso- tory.It was designed as an analysis tool to help the US Department line gallon equivalent estimates for the 2009 Honda Civic GX[31]. of Energy (DOE) quantify the potential of hybrid vehicles to save EPA estimates are derived from a formula that weights results from fuel and reduce emissions.ADVISOR simulates vehicle powertrains five different drive cycles[32].ADVISOR simulations of the EPA Ur- and power flows among its components [23,24]. Fuel use and ban Dynamometer Driving Schedule(UDDS)and the EPA Highway tailpipe emissions can be simulated in a variety of standard driving Fuel Economy Test(HWFET)were used to estimate city and high- cycles. In ADVISOR, AC power load is added as an accessory way fuel economies,respectively.A"combined"fuel economy was mechanical load. computed as a weighted average(55%UDDS,45%HWFET). In this study we focus on ADVISOR simulations of the EPA Speed Table 4 presents results for ancillary loads ranging from 0 to Correction (SC03) driving cycle, a transient test cycle with an 4 kW. The EPA city, highway, and combined fuel economies lie average speed of 34.8 km h-1 (21.6 mph) and a maximum speed within the range of our simulation results.We also note that ADVI- of 88.2 km h-1 (54.8 mph). We also show results for the EPA SOR predicts an SC03 drive cycle fuel economy of 19-26 mpg,sim- Urban Dynamometer Driving Schedule (UDDS) (average speed= ilar to that of the UDDS urban drive cycle. 31.5 km h-1=19.6 mph),and the EPA Highway Fuel Economy Test (HWFET) driving cycle (average speed=77.7 km h-1=48.3 mph) [29].Simulations were performed with ancillary power load rang- 6.Results ing from 0 to 4 kW at a resolution of 0.2 kW.Since we did not have access to an ADVISOR vehicle prototype for the Honda Civic,each 6.1.AC calibration(16 July 2010) ADVISOR simulation was run for two available prototypes:one with a 63 kW engine,and the other with a 102 kW engine.Results were The heater in the silver car was slightly more powerful than that then interpolated to match the engine power rating of the Honda in the black car, yielding 2-3°C higher peak values of vent air a U 12 (a) 70 black ,,,,mn, —measured O " 11 black - - 65 —fit a 10 i- U 60 - ■■•■N•■••••■■ —after resizing AC E -- y 9 .- - a) cooldown target(25°C) •- - 55 c 8 - -, • 6 E 45 i=_ 5 1l 40 `_ E co w 4 1 —black_cooll(09:32-09:59 LST,28 min,25°C,0.83 kW/m2) c 35 - 3 —black_cool2(11:02-11:32 LST,30 min,29°C,0.98 kW/m')- R 30 �� ���� , ■ N 2 —black_cool3(12:33-13:01 LST,28 min,33°C,1.00 kW/m�}L U N a 1 black_cool4(14:02-14:32 02-14:32 LST,30 min,36°C,0.86 kW/m')L black_cool4(14:02-14:32 LST,30 min,36°C,0.86 kW/m2) m O —black_cool5(15:32-16:00 LST,28 min,37°C,0.64 kw/me)!- 20 0 5 10 15 20 25 30 0 5 10 15 20 25 30 Cooling time(min) Cooling time(min) b 70 (b) ci 12 silver I I I ( ) 65 silver measured' r 12. 11 r- -- Cam= —fit = 10 U 60 after resizing AC d g m 55 C C.= cooldown target(25°C) c7 3 c 8 ! L 1 LL____ as 50 _ °- 45 E d 5 i 40 u 4 —silver_cooll(09:32-09:58 LST,27 min,25°C,0.83 kW/m2) E 35 w Q 3 —silver_cool2(11:02-11:32 LST,30 min,29°C,0.98 kW/m2) s3 30 _ E 2 —silver_cool3(12:33-13:00 LST,28 min,33°C,1.00 kW/m2) U C���� ■=1■ a 1 silver_cool4(14:02-14:32 LST,30 min,36°C,0.86 kW/m2) 25 silver-cool4(14:02-14:32 LST,30 min,36°C,0.86 kW/m2) O —silver_cool5(15:32-15:59 LST,28 min,37°C,0.64 kW/m2) 20 • • v 0 5 10 15 20 25 30 0 5 10 15 20 25 30 Cooling time(min) Cooling time(min) Fig.13. Measured and fitted cabin air temperatures versus cooling time in Trial 4, Fig. 12. Difference between cabin surface and cabin air temperatures versus shown for(a)the black car and(b)the silver car.Each graph also shows the cabin air cooling time in each of five trials,shown for(a)the black car and(b)the silver temperature time series predicted after the AC is resized to attain a target final car. cabin air temperature of 25°C. 4352 R Levinson et al./Applied Energy 88(2011)4343-4357 temperature and cabin air temperature.However,the AC systems Fig.8 compares the cabin air heating rate dUldt and AC cooling performed comparably: after just 2 min of cooling, the vent air rate qAC in the black car to those in the silver car.While cooling,the temperatures matched to within I °C,and the cabin air tempera- difference in dUldt is roughly centered about zero and less than tures agreed to within 0.5°C(Fig.4). 0.03 kW in magnitude.The difference in qAc is much larger,peak- ing around 0.3 kW.During the cooling cycle,qAc is one to two or- 6.2.Soaking and cooling(17 July 2010) ders of magnitude larger than dU/dt, suggesting that most of the heat removed by the AC comes from the cabin surface,rather than 6.2.1. Temperatures profiles within each car the cabin air. Fig. 5 shows the evolution of the exterior surface, interior sur- face,and cabin air temperatures in each car over the course of its 6.2.3. Cooldown temperature versus soak temperature five soaking and cooling cycles.The following remarks will focus The five cooling cycles are denoted"cooll"through"cools".Ta- on the middle three soaking and cooling cycles, which span ble 5 summarizes the properties of each cooling cycle,including its 10:00-14:30 LST and are centered about solar noon(12:15 LST). start and end time,duration,primary weather conditions,cabin air While soaking,the warmest surfaces of the black car are usually and surfaces temperatures after soaking and after cooling,and vent its black roof(solar absorptance A=1 -p=0.95)and black dash- board (also A=0.95), both of which are directly heated by the sun.Its next warmest surfaces are the ceiling(conductively heated (a) 4.5 by the roof and radiatively heated by the dashboard and wind- black measured i shield), followed very closely by the windshield (radiatively and 4.0 after resizing ACI_ convectively heated by the dashboard).The seat,which is heated ^ 3.5 primarily by radiative exchange with the ceiling,is markedly cool- er.The coolest interior surface in the black car is the door,which d 3.0 from geometric considerations can be expected to receive less than L- 2.5 half its thermal radiation from the ceiling. c 2.0 ;.,`,",. 0 The warmest warmest surfaces of the silver car while soaking are its o ✓'~ '^',,,.ri/'"vJ^Mo,---, black dashboard (A=0.95), which is directly heated by the sun, u 1.5 Cn and its windshield, which is radiatively and convectively heated Q III 1 0 by the dashboard.The next warmest surfaces are its silver roof(di- p_ rectly heated by the sun,but absorbing only 42%of sunlight)and 0.5 its ceiling(conductively heated by the roof and radiatively heated black_cool4(14:02-14:32 LST,30 min,36°C,0.86 kW/m2) by the dashboard and windshield).As in the black car,the seat is 0.0 markedly cooler,and the door is coolest. 0 5 10 15 20 25 30 We note that in each car, abnormally high door temperatures Cooling time(min) are observed during the first and last soaking cycles, and abnor- mally high seat temperatures are seen in the last soaking cycle. (b) 4.5 _ silver M measured This is simply due to direct solar illumination of the door exterior 4.0 - ' after resizing TAC T and seat surface,which does not occur at other times. Air conditioning rapidly cools the cabin air and all interior sur- 3.5 faces in each vehicle.Dashboard and windshield temperatures re- Q11111111111 main well above the cabin air temperature because the dashboard 1-3- 2.5 ::r::r=:r:=:=G is still heated by the sun and the windshield is radiatively coupled a, pppp�ppp to the dashboard.Air conditioning has little effect on roof surface •E 2.0 temperature, indicating that the conductive heat flow through Z - the lined ceiling is small compared to the roofs solar heat gain. 1.5 _—.— ,___, - - u ��.ii. We approximate each car's cabin surface temperature Ts as the 4 1.0 • area-weighted average of its ceiling, dashboard, windshield, seat 0.5 and door surface temperatures.This estimate of mean interior sur- silver_cool4(14:02-14:32 LST,30 min,36°C,0.86 kW/m') face temperature neglects unmonitored surfaces, including the 0.0 floor, rear window and side windows. Fig. 6 shows the evolution 0 5 10 15 20 25 30 of T5, Ta and Tv in each car, as well as that of Ta-T„and T5-Ta. Cooling time(min) (Since the vent air temperature is relevant only when the AC is on, zero values drawn for Tv and Ta—Tv during the soak cycles Fig.14. Measured AC cooling rate in Trial 4 shown for(a)the black car and(b)the should be ignored.) silver car.Each graph also shows the cooling rate predicted after the AC is resized to attain a target final cabin air temperature of 25°C. 6.2.2.Black car versus silver car Fig. 7 compares the roof, ceiling, dashboard, windshield, seat, door,vent air and cabin air temperatures in the black car to those Table 7 Coefficients of proportionality y relating changes in rates of fuel consumption F,NOx in the silver car.As expected,the greatest temperature difference is emission ENO„CO emission Eco and HC emission EHc in each of three drive cycles to observed at the roof,where the black car was up to 25°C warmer change in ancillary power load. than the silver car.While soaking,the ceiling temperature differ coefficient UDDS SCO3 HWFET ence(black-silver)peaked at 11 °C,while the dashboard tempera- ture difference was less than 5°C and the windshield temperature Ye(L per 100 km per kW) 0.884 0.830 0.403 difference was less than 2°C.The seat and door temperature differ- Ye,NO,(mg km kw� ) 33 39 24 Ys,co(mg km kW-1) 60 123 29 ences reached 7°C and 5°C,respectively.The vent air and cabin air YE,HC(mg km-1 kw ') 22 29 10 temperature differences each peaked around 5-6°C. R.Levinson et al./Applied Energy 88(2011)4343-4357 4353 air temperatures after cooling. Each temperature is reported first at the start of each 30 min cycle and slow near its end.For example, for the black car,then for the silver car,followed by the black-sil- during the fourth cooling cycle, the cabin air temperature in the ver temperature difference. black car falls 16°C in the first two minutes(8°C min-1),another The cabin air and vent air temperatures attained after—30 min 11 °C in the next 18 min(0.6°C min-1), and just 2°C in the final of cooling each strongly and linearly correlate to the cabin air tem- 10 min (0.2°C min-1). This indicates that Ta asymptotically ap- perature reached after —60 min of soaking, with coefficient of proaches a quasi-steady value Ta toward the end of the cooling determination R2>0.99 for the former and R2=0.95 for the latter cycle. (Fig.9a).The(area weighted mean)cabin surface temperatures at- Fig. 12 shows the variation with cooling time of Ts—Ta in each tained after soaking and after cooling also linearly correlate to the car.In the three middle cooling cycles(coo12,cool3 and cool4),this cabin air soak temperature, with R2=0.87 and R2=0.93, respec- temperature difference varies little after the first two minutes of tively (Fig. 9b). The same linear relationships work equally well cooling.For example,during the final 28 min of the fourth cooling for both cars.This indicates that under the strictly controlled con- cycle,Ts—Ta decreases by 1.2°C in the black car and 0.5°C in the ditions of these experiments,cabin air soak temperature captures silver car,while the vent air temperatures each fall by about 7°C. the thermal history of the soaking interval sufficiently well to pre- This indicates that Ts—Ta depends only weakly on Tv. dict cabin air and cabin surface temperatures after cooling. 6.2.5.Resizing AC to attain 25°C final cabin air temperature 6.2.4.Applicability of cabin air temperature model The black car attained a final cabin air temperature below 25°C The validity of the cabin air temperature model in Eq. (2)was in the first cooling cycle,and the silver car did so in the first and tested by regressing the rate of change of the cabin air temperature, second cooling cycles. Otherwise, neither vehicle's cabin air tem- dTaldt, to the temperature differences Tv—Ta and TS—Ta. Regres- perature was reduced to 25°C or lower after approximately sion coefficients a and/3 for each car and cooling cycle are presented 30 min of maximum cooling.For example,at the end of the fourth in Table 6,along with each fit's coefficient of determination R2.Val- cooling cycle, the cabin air temperatures in the black and silver ues of R2 were fairly high,ranging from 0.86 to 0.93 for the black car cars were 34.3°C and 29.9°C, respectively(Table 6). and 0.86 to 0.95 for the silver car.Fig.10 shows the measured and To attain a lower final cabin air temperature Ta,Eq.(4)indicates fitted values of dTaldt for the fourth cooling cycle in each car. that the vent air temperature Tv(t)must be decreased by the differ- Fig. 11 shows the variation with cooling time of the cabin air ence AT between the cabin air final temperature Ta(approximated temperature reduction Ta(0)—Ta(t)in each vehicle.Cooling is rapid by the cabin air temperature measured at the end of the cooling (a) 30 - - _e (b) 14 f ‘t:/ • 25 - 9g/ 0 12 O 20 a' •5 10 9r,. N co a 8 ISIN °' 15 x O w / To Z 6 0 10 - ,ry s 0 4 ry co Li 5 'o LL 2 h / 0 ' _ 0 1 2 3 4 0 0 1 2 3 4 Ancillary load of cool car(kW) Ancillary load of cool car(kW) (c) 14 (d) 14 / % I / 12 / e 12 - iv C ` C .9 10 - // 2 10 Vii" .5 �' .0 / m = c P = Q! 2 8 0 13 8 u. To C C 0 4 0 4 ,'ry V U to co u 2 43 LL 2 - 0 0 r/ 0 1 2 3 4 0 1 2 3 4 Ancillary load of cool car(kW) Ancillary load of cool car(kW) Fig.15. Fractional reductions in rates of(a)fuel consumption,(b)NO,emission,(c)CO emission and(d)HC emission as a function of ancillary loads of the standard(black) and cool(nonblack)cars for the SCO3 driving cycle.Each curve represents a different value for ancillary load of the standard car.Results represent a vehicle with a power rating of 84 kW. 4354 R.Levinson et al./Applied Energy 88(2011)4343-4357 cycle)and Ta T .For example,in the fourth cooling cycle AT would be Table 8 9.3°C for the black car and 4.9°C for the silver car if Ta=25°C. Variations with shell solar reflectance of rates of fuel consumption, fuel savings, New vent air temperature profiles T,(t) Tv(t)-AT were com Pollutant emission and emission reduction for a compact sedan (engine power 84 kW). Results are presented for three different drive cycles simulated using puted for each car and cooling cycle based on the value of AT re- ADVISOR.Parenthetical results indicate percent reductions in fuel consumption and quired to cool the cabin air to 25°C.Eq.(5)was then numerically emission rates relative to the black car. integrated to compute the new cabin air temperature profile Driving Black car cool colored Silver or Hypothetical super- Ta(t). Fig. 13 compares the measured and fitted values of Ta(t) in cycle (p=0.05) car(p=035) white car white car(p=0.80) the fourth cooling cycle to the values of Ta(t) computed after (p=o.60) decreasing the black and silver cars' vent air temperatures by Fuel consumption(L per 100 km) AT=9.3°C and AT=4.9°C, respectively.Also drawn for reference SCO3a 10.95 10.84 10.74 10.67 is the cooldown target temperature(25°C). UDDSb 10.59 10.46 10.36 10.28 AC cooling rates before and after vent temperature reduction HWFET` 6.86 6.80 6.76 6.72 were computed from Eqs.(6)and(7).Fig. 14 shows for the fourth Fuel savings(L per 100 km) cooling cycle in each car the measured AC cooling rate and the AC SC03 NA 0.12(1.1%) 0.21(1.9%) 0.29(2.6%) cooling rate after lowering the vent air temperature to attain a final UDDS NA 0. 8) 0.23 0 0.31(2.9%) HWFET NA 0.00 5 56((0.0.82%) 0.71 0 0(1.5%) 0.14(2.0%) cabin air temperature of 25°C. 5C0 The AC cooling capacity Q(peak AC cooling rate)required to at- emission reduction(g km-1)d CO3 NA 2.7 4.9 6.7 tain Ta=25°C was computed from Eq.(8)for each car and cooling UDDS NA 2.9 5.2 7.1 cycle (Table 6). For example, in the fourth cooling cycle Q was HWFET NA 1.3 2.4 3.3 3.83 kW in the black car, and 3.34 kW in the silver car.The ratio NO emission(g km-1) of Qsiiver to Qblack ranged from 0.83 to 0.87 over the three middle SC03 1.22 1.22 1.22 1.21 cooling cycles. UDDS 0.70 0.69 0.69 0.69 HWFET 0.53 0.53 0.53 0.52 NO„emission reduction(mg km-') 6.3.Fuel savings and emission reductions SCO3 NA 5.4(0.44%) 9.9(0.80%) 13(1.1%) UDDS NA 4.7(0.67%) 8.5(1.2%) 12(1.7%) 6.3.1.Fuel consumption and pollutant emission versus ancillary power HWFET NA 3.3(0.62%) 6.1(1.1%) 8.3(1.6%) load CO emission(g km-1) Table 7 shows values of y obtained by linearly regressing ADVI- SCO3 3.99 3.98 3.96 3.95 SOR simulations of fuel consumption, NO),emission, CO emission UDDS z.69 1.69 z.69 1.68 HWFET 1.69 1.69 1.69 1.68 and HC emission rates to ancillary power load. The variations of , fuel consumption and emissions with ancillary load were highly CO emission reduction(mg km) SC03 NA 17(0.43%) 31(0.79%) 43(1.07%) linear within the simulated range (0-4 kW) and the minimum UDDS NA 8.4(0.41%) 15(0.74%) 21(1.01%) coefficient of determination (R2) was 0.96. Fig. 15 relates reduc- HWFET NA 4.0(0.24%) 73(0.43%) 10(0.59%) tions in fuel consumption and emission to the ancillary loads of HC emission(g km-') the standard (black) and cool (nonblack) cars. Each curve repre- SCO3 1.12 1.11 1.11 1.11 sents a different value for ancillary load of the standard car. For UDDS 0.61 0.61 0.61 0.60 brevity,we present charts only for the SCO3 driving cycle. HWFET 0.46 0.46 0.46 0.46 HC emission reduction(mg km-1) SC03 NA 4.1(0.37%) 7.4(0.67%) 10(0.91%) 6.3.2.Fuel savings and emission reductions versus cool car solar UDDS NA 3.1(0.50%) 5.6(0.92%) 7.6(13%) reflectance HWFET NA 1.4(0.30%) 2.5(0.55%) 3.4(0.75%) Since roof and cabin air soak temperatures peaked in the fourth a SC03 simulates transient driving(average speed=34.8 km h-1,21.6 mph). cycle(Table 5),AC capacity requirements Q(black car)and QH(sil- b UDDS simulates urban driving(average speed=31.5 km h-'=19.6 mph). ver car)were based on values computed for the fourth cooling cy- ` HWFET simulates highway driving(average speed=77.7 km h-1=48.3 mph). cle.The following analysis assumes AC capacities QL=3.83 kW and ° CO2 emission reduction calculated at the rate of 2321 g CO2 per L gasoline[30]. QH=3.34 kW and shell solar reflectances pi_=0.05 and pH=0.58, for a capacity reduction of 92.5 W per 0.1 increase in shell solar economy by 0.10 km L-1 [0.24 mpg](1.1%).It would also decrease reflectance. CO2 emissions by 2.7 g km-1(1.1%),NO),emissions by 5.4 mg km-1 Hendricks [33]obtained a maximum COP of 1.6 when optimiz- (0.44%),CO emissions by 17 mg km-1(0.43%),and HC emissions by ing the COP of a mechanically driven compressor for the SCO3 cy- 4.1 mg km-1 (0.37%). Selecting a typical white or silver shell cle.Here we select a COP of 2 to conservatively estimate reduction (p=0.60) instead of a black shell would lower fuel consumption in ancillary power load,which is inversely proportional to COP. by 0.21 L per 100 km(1.9%), raising fuel economy by 0.19 km L-1 Table 8, Fig. 16 and Fig. 17 present fuel savings and emissions [0.44 mpg] (2.0%). It would also decrease CO2 emissions by reductions attained when a cool (solar reflective) car shell is 4.9 g km-1(1.9%),NO),emissions by 9.9 mg km-1(0.80%),CO emis- substituted for a standard(black)car shell(p=0.05).Dashed ver- sions by 31 mg km-1 (0.79%), and HC emissions by 7.4 mg km-1 tical lines in Figs. 16 and 17 mark the shell solar reflectances of a (0.67%).A hypothetical super-white car shell(p=0.80)could save typical cool colored car (p=0.35), a typical white or silver car 0.29 L per 100 km(2.6%),increasing fuel economy by 0.25 km L-1 (p=0.60), and a hypothetical super-white car (p=0.80)4. Fig. 16 [0.59 mpg] (2.7%) and decreasing CO2, NO„, CO and HC emissions shows fractional fuel savings and emission reductions and Fig. 17 by 6.7 g km-1 (2.6%), 13 mg km-1 (1.1%), 43 mg km-1 (1.1%), and shows absolute fuel savings and emission reductions. 10 mg km-1 (0.91%),respectively. Results from our model with y values from the SCO3 drive cycle As discussed previously, emissions in standardized driving cy- indicate selecting a typical cool colored shell(p=0.35)would re- des are typically lower than those in real-world(off-cycle)driving. duce fuel consumption by 0.12 L per 100 km(1.1%),increasing fuel Hence, our simulation results may underestimate emission reductions. n Many white metal roofing products have initial solar reflectances in the range of We can compare fuel and emission reductions of urban versus 0.7-0.8[34].We present the super-white shell(p=0.80)as a limiting case. highway driving by observing results for the UDDS and HWFET R.Levinson et al./Applied Energy 88(2011)4343-4357 4355 (a) 5.0 (b) 3.0 4.5 - -UDDS t5 cf0i -UDDS m `m m U U , U a m a`) a) 0 2.5 - -0 a� °3 • 4.0 SCO3 0 3 c -HWFET o > 0_:= a)3.5 - O w o o w m 3 c -HWFET " 2.0 - -SCO3 L 3.0 °0 3 o lc N U 3 TL) 2.5 - 0 1.5 - 2.0 - Z , °2 1.5 - 0 1.0 - u 1.0 - 0.5 - 0.5 - 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Solar reflectance of cool car Solar reflectance of cool car (c) 3.0 - (d) 3.0 - SCO3 c0i -UDDS L 2.5 a3 m 0 2.5 - `- m °3 0 -UDDS o > Q-� -SCO3 > °L .N cnn3 0 °o 'w rani TD 3 2.0 - -HWFET 0 w .0 2.0 - -HWFET o TD t a o lc a) c°,1 3 d o 3 p 1.5 - cj 1.5 - U = Ta To 0 1.0 - 0 1.0 - U �J R u 0.5 - Lt 0.5 - i 0.0 0.0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Solar reflectance of cool car Solar reflectance of cool car Fig.16. Fractional reductions in rates of(a)fuel consumption,(b)NOx emission,(c)CO emission and(d)HC emission versus solar reflectance of the cool car shell,assuming a vehicle power rating of 84 kW.Reference values of solar reflectance for a typical cool colored car,a typical white or silver car,and a hypothetical super-white car are shown as dashed vertical lines. driving cycles (Fig. 16). Relative to the SCO3 drive cycle, fuel say- and emission reductions attainable by using a cool shell to reduce ings are larger for the UDDS cycle(urban driving)and smaller for ancillary load. the HWFET cycle (highway driving). Further, relative to the SCO3 The air conditioners in the experimental vehicles were in most cycle, emissions reductions for NOR, CO, and HC are smaller for trials too small to lower cabin air temperature to 25°C within both the UDDS cycle and HWFET cycle. Emissions reductions are 30 min.We estimate that if the vehicle ACs were resized to meet larger for UDDS than HWFET for NOR, CO, and HC. This may be this target,the AC cooling capacity would be 3.83 kW for the car due to the fact that emissions are more sensitive to transients with low solar reflectance and 3.34 kW for the car with high solar (e.g.,simulated vehicle accelerations)in driving cycles [31]. reflectance(Table 6). Assuming that potential reductions in AC capacity and engine ancillary load scale linearly with increase in shell solar reflectance, 7.Summary ADVISOR simulations of the SCO3 driving cycle indicate that substi- tuting a typical cool-colored shell (p=0.35) for a black shell In this study we estimated the decrease in soak temperature, (p=0.05) would reduce fuel consumption by 0.12 L per 100 km potential reduction in AC capacity,and potential fuel savings and (1.1%),increasing fuel economy by 0.10 km L-1 [0.24 mpg] (1.1%). emission reductions attainable through the use of solar reflective It would also decrease CO2 emissions by 2.7 g km-1 (1.1%), NOR car shells.First,we experimentally characterized component tem- emissions by 5.4 mg km-1 (0.44%), CO emissions by 17 mg km-1 peratures and cooling demands in a pair of otherwise identical (0.43%),and HC emissions by 4.1 mg km-1(0.37%).Selecting a typ- dark and light colored vehicles,the former with low solar reflec- ical white or silver shell (p=0.60) instead of a black shell would tance (p=0.05) and the latter with high solar reflectance lower fuel consumption by 0.21 L per 100 km (1.9%), raising fuel (p=0.58). Second, we developed a thermal model that predicted economy by 0.19 km L-1 [0.44 mpg] (2.0%).It would also decrease the AC capacity required to cool each vehicle to a comfortable final CO2 emissions by 4.9 g km-1(1.9%),NOR emissions by 9.9 mg km-1 temperature of 25°C within 30 min.Third,we used the ADVISOR (0.80%),CO emissions by 31 mg km-1(0.79%),and HC emissions by vehicle simulation tool to estimate the fuel consumption and pol- 7.4 mg km-1(0.67%).A hypothetical super-white car shell(p=0.80) lutant emissions of each vehicle in various standard drive cycles could save 0.29 L per 100 km (2.6%), increasing fuel economy by (SC03,UDDS, and HWFET).Finally,we calculated the fuel savings 0.25 km L-1 [0.59 mpg] (2.7%) and decreasing CO2, NOR, CO and 4356 R Levinson et al./Applied Energy 88(2011)4343-4357 (a) 0.50 (b) 25 - E 0.45 - -UDDS 0 CO CO E --SCO3 U v 0 o 0.40 N °' w. - 20 o m -SC03 > a) 0..- .a c`� o n co 3 E _-UDDS o N =3 x0.35 -HWFET ° a) o =- HWFET ' O 0.30 15 - Z 0 1E 3 ca 0.25 - o = iv v+ 0.20 - O 10 = 73 Z w 0.15 - N a) Z 0.10 - / y 5 = A 0.05 Q - Q 0.00 , , i , , , , I , , , I , , , 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Solar reflectance of cool car Solar reflectance of cool car (c) 70 (d) 20 - E 60 -SCO3 iTs 0 iTs " 18 - -SCO3 0 0 c0 a a) E -o a) a) -UDDS o .N y 3 ca 16 -UDDS o y w 3 50 0 m 14 c -HWFET w -HWFET 0 c o 40 0 3 .° 12 _ v 3 3 = 10 - d 30 2 8 O U C.) 2 - W 20 a) 6 - 3 4 3 0 a 10 2 - Q Q _ ' 0 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Solar reflectance of cool car Solar reflectance of cool car Fig.17.Absolute reductions in rates of(a)fuel consumption,(b)NO emission,(c)CO emission and(d)HC emission versus solar reflectance of the cool car shell.Reference values of solar reflectance for a typical cool colored car,a typical white or silver car,and a hypothetical super-white car are shown as dashed vertical lines. HC emissions by 6.7 g km-1(2.6%),13 mg km-1(1.1%),43 mg km-1 Transportation and Air Quality, US Environmental Protection Agency. EPA- t 0 420-D-09-003,September 2009. (1.1%), and 10 mg km (0.91%), respectively. These results may [3] Bevilacqua OM.Effect of air conditioning on regulated emissions for in-use underestimate emission reductions in real-world driving. vehicles.Clean Air Vehicle Technology Center,Oakland,CA.Phase I final Report Prepared for Coordinating Research Council,Inc.Atlanta,GA,CRC Project E-37; Acknowledgments October 1999. [4] Rugh JP, Hendricks TJ, Koram K. Effect of solar reflective glazing on Ford Explorer climate control, fuel economy, and emissions. Paper presented at This work was supported by the California Energy Commission International Body Engineering Conference&Exposition,October,Detroit,MI, through its Public Interest Energy Research Program. It was also USA;2001.doi:10.4271/2001-01-3077. supported by the Assistant Secretary for Energy Efficiency and [S]Turley D, Hopkins D, Goudey H. Reducing vehicle ancillary loads using advanced thermal insulation and window technologies. Paper presented at Renewable Energy,Office of Building Technology,State,and Com- SAE World Congress,Detroit,MI,USA;2003.doi:10.4271/2003-01-1076. munity Programs,of the US Department of Energy under Contract [6] Hoke PB,Greiner C.Vehicle paint radiation properties and affect on vehicle No.DE-ACO2-05CH11231.We wish to thank the California Depart- soak temperature, climate control system load, and fuel economy. Paper ment of General Services for use of their vehicles and facility,with presented at SAE World Congress,April,Detroit,MI,USA;2005.doi:10.4271/ ty lt 2005-01-1880. special appreciation to Kimberly Harbison for her assistance;John [7] Rugh J,Farrington R.Vehicle ancillary load reduction project close-out report: Rugh of the National Renewable Energy Laboratory, for technical an overview of the task and a compilation of the research results.Golden,CO: National Renewable Energy Laboratory (NREL), NREL/TP-540-42454; 2008. advice; former California Energy Commissioner Arthur Rosenfeld, <http://www.nrel.gov/vehicl esandfuel s/ancillary_loads/pdfs/42454.pdf>. for his support; and Philip Misemer of the California Energy Com- [8] Han T, Chen KH. Assessment of various environmental thermal loads on mission,for guiding our project. passenger compartment soak and cool-down analyses.Paper presented at SAE World Congress&Exhibition,April,Detroit,MI,USA;2009.doi:10.4271/2009- 01-1148. References [9] Sullivan R, Selkowitz S. Effects of glazing and ventilation options on automobile air conditioner size and performance. Paper presented at SAE [1] Daly S.Automotive air-conditioning and climate control systems.Butterworth- International Congress & Exposition. February, Detroit, MI, USA; 1990. Heinemann,Lincare House,Jordan Hill,Oxford 0X2 8DP,30 Corporate Drive, doi:10.4271/900219. Suite 400 Burlington,MA 01803;2006. [10] Akabane T,Ikeda S,Kikuchi K,Tamura Y,Sakano R,Bessler W,et al.Evaluation [2] EPA.Draft regulatory impact analysis:proposed rulemaking to establish light- of an electrically driven automotive air conditioning system using a scroll duty vehicle greenhouse gas emission standards and corporate average fuel hermetic compressor with a brushless DC motor.SAE Technical Paper 890308; economy standards. Assessment and Standards Division, Office of 1989.doi:10.4271/890308. R.Levinson et al./Applied Energy 88(2011)4343-4357 4357 [11] Heydari A, Jani S. Entropy-minimized optimization of an automotive air [22] Hendricks TJ.Optimization of vehicle air conditioning systems using transient conditioning and HVAC system. Paper presented at SAE World Congress, air conditioning performance analysis. Paper presented at Vehicle Thermal March,Detroit,MI,USA;2001.doi:10.4271/2001-01-0592. Management Systems Conference&Exposition,May,Nashville,TN,USA;2001. [12] Huang D,Wallis M,Oker E,Lepper S.Design of vehicle air conditioning systems <http://www.nrel.gov/vehiclesandfuels/ancillary_loads/pdfs/ using heat load analysis.Paper presented at SAE World Congress&Exhibition, 2001_01_1734.pdf>. April,Detroit,MI,USA;2007.doi:10.4271/2007-01-1196. [23] Wipke K,Cuddy M,Bharathan D,Burch S,Johnson V,Markel A,et al.ADVISOR [13]Junior CS,Strupp NC,Lemke NC,Koehler J.Modeling a thermoelectric HVAC 2.0:a second-generation advanced vehicle simulatory for systems analysis.14 system for automobiles.J Electron Mater 2009;38(7):1093-7. pp.NREL Report No.TP-540-25928,Golden,CO:NREL;1999. [14] Henry RR,Koo J,Richter C.Model development,simulation and validation,of [24] Wipke K,Cuddy M,Burch S.ADVISOR 2.1:a user-friendly advanced powertrain powertrain cooling system for a truck application.Paper presented at Vehicle simulation using a combined backward/forward approach.14 pp.NREL Report Thermal Management Systems Conference&Exhibition,Nashville,TN,USA; No.JA-540 26839,Golden,CO:NREL;1999. 2001.doi:10.4271/2001-01-1731. [25] Samuel S, Austin L, Morrey D. Automotive test drive cycles for emission [15] Taxis-Reischl B,Morgenstern S,Mersch T,Brotz F.Progress in the optimized measurement and real-world emission levels-a review.Proc Inst Mech Eng application of simulation tools in vehicle air conditioning.Paper presented at 2002;216:555-64. Vehicle Thermal Management Systems Conference & Exhibition, May, [26] Konopacki S,Akbari H,Pomerantz M,Gabersek S,Gartland L.Cooling energy Nashville,TN,USA;2001.doi:10.4271/2001-01-1699. savings potential of light-colored roofs for residential and commercial [16] Wolfahrt J,Baier W,Wiesler B,Raulot A,Rugh J,Bharathan D.Aspects of cabin buildings in 11 US metropolitan areas. Technical Report LBNL-39433, fluid dynamics, heat transfer, and thermal comfort in vehicle thermal Lawrence Berkeley National Laboratory,Berkeley,CA;1997. management simulations.Paper presented at Vehicle Thermal Management [27] Levinson R,Akbari H,Berdahl P.Measuring solar reflectance-Part I:Defining a Systems Conference & Exhibition, May, Toronto, ON, Canada; 2005. metric that accurately predicts solar heat gain. Solar Energy doi:10.4271/2005-01-2000. 2010;84:717-1744. [17] Zhang H,Dai L,Xu G,Li Y,Chen W,Tao W.Studies of air-flow and temperature [28] Levinson R,Akbari H,Berdahl P.Measuring solar reflectance-Part II:Review of fields inside a passenger compartment for improving thermal comfort and practical methods.Solar Energy 2010;84:1745-59. saving energy.Part I:Test/numerical model and validation.Appl Therm Eng [29] EPA.Federal test procedure revisions:driving cycles.Office of Transportation 2009;29(7):2022-7. and Air Quality,US Environmental Protection Agency;2007.Online Resource. [18] Zhang H,Dai L,Xu G,Li Y,Chen W,Tao W.Studies of air-flow and temperature <http://www.epa.gov/oms/sftp.htm#cycles>.[accessed December,2010]. fields inside a passenger compartment for improving thermal comfort and [30] EPA.Emission facts:average carbon dioxide emissions resulting from gasoline saving energy. Part II: Simulation results and discussion. Appl Therm Eng and diesel fuel. Office of Transportation and Air Quality,US Environmental 2009;29(7):2028-36. Protection Agency.EPA-420-F-05-001,February;2005. [19] Cullimore BA, Hendricks TJ. Design and transient simulation of vehicle air [31] EPA. www.fueleconomy.gov. US Environmental Protection Agency and US conditioning systems. Paper presented at Vehicle Thermal Management Department of Energy;2011.<http://fueleconomy.gov>. Systems Conference & Exposition, May, Nashville, TN, USA; 2001. [32] EPA.Fuel economy labeling of motor vehicle revisions to improve calculation doi:10.4271/2001-01-1692. of fuel economy estimates. Final technical support document. US [20] Bharathan D, Chaney L, Farrington RB, Lustbader J, Keyser M, Rugh J. An Environmental Protection Agency. Office of Transportation and Air Quality. overview of vehicle test and analysis from NREL's A/C fuel use reduction EPA420-R-06-017;2006. research. Paper presented at Vehicle Thermal Management Systems [33] Hendricks TJ.Vehicle transient air conditioning analysis:model development Conference&Exhibition(VTMS-8), May 20-24, Nottingham,UK: NREL/CP- &system optimization investigations.Golden,Colorado,USA: NREL/TP-540- 540-41155; 2007. <http://www.nrel.gov/vehic lesandfuels/ancillary_loads/ 30715;2001.<http://www.nrel.gov/docs/fy0losti/30715.pdf>. pdfs/41155.pdf>. [34] CRRC. Cool Roof Rating Council Rated Products Directory; 2011. <http:// [21] Rugh J. Integrated numerical modeling process for evaluating automobile coolroofs.org>. climate control systems.Paper presented at Future Car Congress,June,Crystal City,VA, USA; 2002. <http://www.nrel.gov/vehic lesandfuels/ancillary_loads/ pdfs/evaluati ng_cli mate.pdf>.