Whole House Thermal Performance Of Asphalt Shingles Exploiting Special .

ESL-HH-06-07-321000 Hours2000 Hours4000 Hours5000 Hours3000 Hours3.5Total Color Change (ΔE)3.02.52.01.51.00.50.0A (IRR)A (Std)C (IRR)C (Std)E (IRR)E (Std)Roof Materials (Asphalt Shingles)Figure 2. Total color change (ΔE) of asphalt shingles under fluorescent UV light usingthe ASTM G154-04 protocol. (Data from Shepherd Color Company)Table 1. Granules exposed to natural sunlight in south Florida and painted with and withoutIRR coatingsPigmentExposuretime(months)Initial color of asphalt-coated panelL*IRRa*IRRb*IRRColor changeafter exposure(ΔE)Conventional pigmentCarbon black1822.00.4-0.22.4Black iron oxide42.522.92.73.61.6Ferro V-7785826.02.12.60.8Ferro O-1765B23.522.71.50.70.9Cool color pigmentProceedings of the Fifteenth Symposium on Improving Building Systems in Hot and Humid Climates, Orlando, FL, July 24-26, 2006

ESL-HH-06-07-320 Hours1000 Hours4000 Hours5000 Hours2000 Hours3000 Hours35Solar Reflectance (ρ )302520151050A (IRR)A (Std)C (IRR)C (Std)E (IRR)E (Std)Roof Products (Asphalt Shingles)Figure 3. Solar reflectance of asphalt shingles under xenon-arc light using the ASTMG-155 protocol. (Data from 3M Company)1000 Hours4000 Hours2000 Hours5000 Hours3000 Hours3.0Total Color Change (ΔE)2.52.01.51.00.50.0A (IRR)A (Std)C (IRR)C (Std)E (IRR)E (Std)Roof Products (Asphalt Shingles)Figure 4. Total color change (ΔE) of asphalt shingles under xenon-arc light using theASTM G-155 protocol. (Data from 3M Company)Proceedings of the Fifteenth Symposium on Improving Building Systems in Hot and Humid Climates, Orlando, FL, July 24-26, 2006

ESL-HH-06-07-323.0 after 5000 hours of exposure, which can bevisually distinguished but is still considered goodcolor fastness.DEMONSTRATION HOME FIELD TESTINGTwo new homes with identical footprints,layouts, and equipment design were equipped withasphalt shingle roofs with and without IRR pigments.The homes were built in Redding, California, by thefirm Ochoa and Shehan, Inc. The Redding site wasconsidered an excellent location for the field testingof asphalt shingles because high temperatures in thesummer can be as much as 45 C (110 F) and winterstypically have subfreezing temperatures. The twodemonstration homes are on different cul-de-sacs butwithin about 100 yards of each other. Both haveabout the same azimuth orientation with respect tothe sun, which allows direct comparison of thethermal performance of the two roof assemblies.The residences are one-story ranch-style housesbuilt on concrete slab. Each house has about2400 square feet of floor space, and each uses two3½-ton split-system air conditioners for comfortcooling. The attics for each home are ventilated bysoffit and ridge vents. R-19 loose fill insulation wasblown into the attic space, and the indoor air-handlerand air-distribution ducting are located in the attic.All ducts were wrapped in R-5 insulation. Orientedstrand board (OSB) decking facing the attic interioruses a radiant barrier, as required by the buildingcodes for the county of Redding.Elk Corporation donated its WeatherwoodPrestique Cool Color and its conventionalWeatherwood Prestique asphalt shingles for thesefield tests. The IRR shingle has a solar reflectance of0.26 and a thermal emittance of 0.90. Theconventional shingle has solar reflectance of 0.09 andthermal emittance of 0.89. The cool color shinglesare advertised on Elk’s web site 2 as the “first energyefficient cool asphalt shingles offered in a palette ofrich, organic colors.” Elk offers a 40-year limitedwarranty with a limited wind warranty of up to90 mph, and a UL Class “A” fire rating. Theprototype shingles were developed in conjunctionwith 3M to meet the initial 0.25 solar reflectancespecified by ENERGY STAR criteria.Instrumentation and Data AcquisitionInstrumentation was added to the pair ofdemonstration homes to catalogue temperatures andheat flows across the roof and attic assembly ngles prestique ccs.cfmmeasure the relative humidity of the ambient air inthe attic. Sensors were also installed in the livingspace to measure the indoor ambient return andsupply air temperatures and the indoor relativehumidity. Whole house power usage was obtainedfrom the Redding utility with the permission of thehomeowners. The power consumed by thecondensing unit of each air conditioner was measuredusing watt-hour transducers. 3Roof Deck and Attic Floor (Ceiling).Heat flux transducers (HFTs) were embedded inthe roof decks and the attic floor to measure the heatflows crossing the decks and attic floor of eachhouse. The roof decks are made of 5/8-in. OSB.Typical construction uses 15/32-in. OSB; however,the 5/8-in. OSB was selected because it is ofsufficient thickness for embedding a 0.038-m(0.15-in.-) thick HFT in the OSB withoutcompromising the accuracy of the heat flow (Fig. 5).We checked the thermal conductivity of 5/8-in.and 1/4-in.-thick boards because OSB is made fromvarious waste wood products and conductivity mighttherefore vary with thickness. A 0.61-m- (2-ft-)square section of 5/8-in. OSB was placed in a heatflux metering apparatus and calibrated to determinethe thermal conductivity of the OSB. The toptemperatures of the board were set at 7.2, 23.9, 37.8,and 48.9 C (45, 75, 100, and 120 F), which aretypical temperatures observed by Parker, Sonne, andSherwin (2002) for roof decks field tested inFt. Myers, Florida. Results revealed that the thermalconductivity of OSB increased linearly withtemperature. The thinner 1/4-in. OSB board was alsotested and found to have a thermal conductancewithin 0.5% of the measures obtained for the thicker5/8-in. board. The thermal conductivity of the thinnerboard varied linearly and had the same slope as thethicker 5/8-in. board. Tests verified that the thinnerboard could be used as a cover plate to hold the heatflux transducer in place (Fig. 5) and would thereforenot adversely affect the heat flow.Shunting due to the differences in thermalconductance of the HFT and the OSB was alsoaccounted for by calibrating the HFT in a 0.61 0.61 m (2 2 ft) guard of 5/8-in.-thick OSB using theASTM C518 protocol (ASTM 1998). Calibrationshowed a slight but linear drop in sensitivity as thetemperature of the OSB was increased from 4.4 to48.9 C (40 to 120 F). The guard became a portionof a sandwich panel equipped with copper/constantan3The power of the indoor blower was not measured bythe transducers.Proceedings of the Fifteenth Symposium on Improving Building Systems in Hot and Humid Climates, Orlando, FL, July 24-26, 2006

ESL-HH-06-07-32Figure 5. Heat flux transducer embedded in OSB guard and used to measure heat flow crossing roof decks ofthe demonstration homes with asphalt shingle roofs.thermocouples and the HFT. Once calibrated, thesandwich panels were shipped to the builders andinstalled in the adjacent roof decks of eachdemonstration home. The sandwich panel was madeof the same material as the deck and was also thesame thickness.While the roof products were being installed, thethermocouples attached to the sandwich panels wereepoxy-glued to the roof surface, taped to the topsideof the deck, placed adjacent to the HFT embedded inthe OSB, and taped to the underside of the OSB deckfacing the interior of the attic. The thinner 1/4-in.board was attached to the underside of the deck toprovide access for future maintenance (Fig. 5).A similar procedure was used for setup of theHFT measuring the heat flow crossing the attic floorand entering the conditioned space. Here, however,we taped an HFT in the center of a 0.61 0.61 m(2 2 ft) piece of gypsum board, covered the devicewith R-19 batt insulation, and proceeded to calibratethe transducer. In the field we simply taped the HFTto ceiling drywall and attached a thermocoupleadjacent to the HFT. Later, after insulation wasblown into the attic, we placed a thermocoupleapproximately at the top surface of the blowninsulation.Weather Data.Pyranometers were placed on adjacent slopedroofs of each home for measuring the morning andafternoon solar irradiance. These measures helpedprove that, for instance, the west-facing roofs for thepair of homes had the same intensity of solar flux.The instruments also indicated the daylight hours anddisplayed peak irradiance with time of day. Athermocouple for measuring the outdoor airtemperature was placed underneath the roof soffit,where it was shaded and sheltered from rain. Allother weather data were gleaned from the CaliforniaIrrigation Management Information System (CIMIS);CIMIS provides current weather data fromcomputerized weather stations acquiring hourly,daily, weekly, and/or monthly solar irradiance,ambient air temperature and relative humidity as wellas wind speed, wind direction, and precipitation. 4Power Measurements.Watt-hour transducers (Wattnode Model WNA1P-240-P) measured the electrical energy consumedby each outdoor condenser unit. The transducers withcurrent transducers were installed in the power panelof each home.Data Acquisition System.Campbell Scientific micro-loggers were usedfor remote acquisition and recording of field data.Salient features of the micro loggers are provided inTable 2. The loggers were equipped with 4 MB ofmemory, a 25-channel multiplexer for thermocouples, a rechargeable battery, a 115 Vac-to-24 Vdctransformer, a modem, a modem surge protector, aweatherproof enclosure and associated cables.The mirco-logger was programmed to scan every30 seconds and reduce analog signals to theengineering units. Averages of the reduced data were4Web site: http://www.cimis.water.ca.gov/.Proceedings of the Fifteenth Symposium on Improving Building Systems in Hot and Humid Climates, Orlando, FL, July 24-26, 2006

ESL-HH-06-07-32Table 2. Salient features of Campbell scientific data loggersItemItem descriptionCSI part No.1CR-23x data logger with 4-MB memoryCR10X-2M2Array-based operating system for CR-23X-4m98013Thermocouple reference thermistor for CR23X wiring panelCR10XTCR412-V power supply with charging regulator and rechargeable batteryPS100518-V, 1.2-A wall charger, 6 ft9591616-channel (4-wire) or 32-channel (2-wire) relay multiplexerAM16/32725-channel solid-state multiplexer for thermocouplesAM25T8Data cable, two peripheral connector cables for data logger, 2 ftSC1298-channel switch closure moduleSDM-SW8A10Telephone modemCOM21011Phone modem surge protector636212Enclosure 16/18, weather-resistant15873written electronically to an open file every 15 min.The averages were calculated over the 15-mininterval and are not running averages; they are resetafter each 15-min interval. The electronic format iscomma-delimited for direct access by spreadsheetprograms. Data files consist of one full week of datacontaining 672 rows of averaged measurementsrepresenting the instrument measurements writtenevery 15 min over the 168 hours of a week. Themicro-logger automatically closes the existing fileand opens a new data file every Friday at midnightfor recording the next week of data. A dedicateddesktop computer calls the micro-logger and acquiresthe previous week of data over a modem connected toa dedicated phone line.Demonstration ResultsThe attics and roof decks of each demonstrationhome were instrumented to document the immediateeffects of IRR pigments on the deck and attic airtemperatures and the heat flows crossing the roofdeck. Our intent was to demonstrate typical summerperformance of IRR shingles and to collect datauseful for formulating and validating computer codescapable of calculating the heat transfer occurringwithin the attic. Once validated, these atticsimulations can be coupled to whole-house buildingmodels for simulating and predicting local, state, andnational energy savings afforded by roofs with IRRshingles. The following sections describe ourfindings about temperature and heat flow results forthe roofs and attics of the demonstration home andcooling energy savings.Roof Temperature and Heat Flow.July and August ambient air temperatures inRedding often exceed 45 C (110 F). Redding fielddata for August 2005 show that the conventionalshingles on the demonstration homes experiencedpeak temperatures of about 73 C (163 F) ascompared with peak temperatures of about 70 C(158 F) for the IRR shingles (Fig. 6). Similar dropswere observed in the attic air temperature (Fig. 6).The reduction in peak temperatures for the coolshingles has potential benefits for improving thechemical and flexural properties of the shingles.Terrenzio et al. (2002) and Shiao et al. (ASTM 2003)showed that heating of asphalt shingles promotes thevaporization and diffusion of oils from the asphalt,with the subsequent migration of oxygen into theasphalt. Terrenzio noted that as aging progresses, thestiffness of the asphalt increases. Therefore, IRRpigments may help increase the service life ofshingles, although there is no definitive data thatcorrelate service life to stiffness of the shingle.The 3 C (5 F) drop in surface temperature inturn resulted in a reduction in the heat flow that waspenetrating the west-facing roof for the pair ofhomes. A sustained and consistently lower heat fluxoccurs when IRR pigments are applied to the shingles(Fig. 7). The IRR shingles had about a 30% lowerheat flux than the conventionally pigmented shingleover the daylight hours during a week in August2005. The reduced heat flow lessens the burden onthe air-conditioning system and should lead to areduction in electrical energy consumed for comfortcooling.Proceedings of the Fifteenth Symposium on Improving Building Systems in Hot and Humid Climates, Orlando, FL, July 24-26, 2006

ESL-HH-06-07-32Figure 6. Surface (solid lines) and attic air (dashed lines) temperatures for the conventional andIRR asphalt shingle roofs: week of data in August 2005.Figure 7. Heat flows penetrating the roof deck for the demonstration houses with conventionaland IRR asphalt shingle roofs.Proceedings of the Fifteenth Symposium on Improving Building Systems in Hot and Humid Climates, Orlando, FL, July 24-26, 2006

ESL-HH-06-07-32Cooling Energy Savings.The principal focus of the field demonstrationswas on collecting attic and roof data to prove thethermal benefits of the cool pigment technology. Butthe experiments also included measurement of poweruse and of air-conditioning supply and return airtemperatures in an effort to estimate annual savingsindependently of the effects of occupancy habits or,put differently, corrected for the effects of occupants.Redding Electric provided the 2005 summerrevenue meter readings for the pair of demonstrationhouses (Fig. 8). The summer 2005 electrical use forthe house with IRR shingles was 67% higher thanthat of the house with conventional shingles.Discussions with the homeowner of the house withconventional shingles revealed, however, that thishouse was used as a second residence and was notconsistently occupied during the summer nor was thethermostat maintained at 22 C (72 F). In contrast, thehouse with a IRR shingle roof was occupiedthroughout the summer.To correct for these differences in occupancyhabits, power measurements were reduced to dailyelectrical energy consumptions. The electrical usageof the air conditioners was plotted against the dailyaverage outside air-to-indoor air temperaturedifference. Regression analysis of the reduced datahelped to remove the effects of different thermostatsettings and occupancy habits.When the outdoor air and return air temperaturesare about the same, the regression lines for the pair ofhouses nearly intersect (Fig. 9). This simply meansthat no cooling is required by the houses, and there isno air conditioner usage. However, as thetemperature difference from the outdoor air to thehome’s return air increases, the cooling savings alsoincrease for the pair of shingle demonstrations(Fig. 9). At an outdoor ambient air-to-indoor airtemperature difference of 10 C (18 F) [about 32.2 C(90 F) outdoor air temperature], the house with IRRFigure 8. Monthly energy use (revenue meter readings) for demonstration houses in Redding, California.Proceedings of the Fifteenth Symposium on Improving Building Systems in Hot and Humid Climates, Orlando, FL, July 24-26, 2006

ESL-HH-06-07-32Figure 9. Air-conditioning cooling energy usage and savings, measured during the daylight hours (Maythrough July 2005) for demonstration houses with conventional and IRR asphalt shingle roofs. Indoor blowerpower is not included.asphalt shingles uses 6.3 kWh per day lesselectricity 5 than the house with conventionalshingles. The cool pigments are therefore providingsavings of roughly 0.90 kWh per day per ton ofcooling capacity. The whole house electricalconsumption for the home with IRR shingles (whichwas more fully utilized than the home withconventional shingles) was obtained from ReddingElectric’s revenue meters and averaged over thesummer months (June through October 2005) atabout 92 kWh per day. Therefore, whole housepower drops about 7% due to the use of IRRmaterials in asphalt shingles.It is interesting to note that higher outdoor airtemperatures yield greater energy savings for the airconditioners operating in the houses with IRR roofs.The trend is slight yet could be important in terms ofthe time-dependent valuation of energy, which placesa premium cost on energy consumed during thehottest portion of the day.The new asphalt shingles showcased in Reddinghave resulted in the demonstrable reduction of heatbuildup in the roof deck through the use of IRR5Indoor blower energy is not included in the 6.3 kWhsavings.pigments. John McCaskill, product brand managerfor Elk Corporation, reports that the additional costfor the finished product of approximately 25 persquare foot makes this product a reasonablealternative to radiant barriers for the reduction ofcooling load. The combination of the costeffectiveness of IRR asphalt shingles with thepotential reduction of urban heat island effectsprovides a mainstream solution for the residentialreroofing market as well as new construction. Elknow offers four “cool” shingle colors across twoproduct lines.CONCLUSIONSIn accelerated weathering tests, xenon-arc andfluorescent light exposures were conducted to judgethe effectiveness of IRR pigments in asphalt shingleroof products. Test results showed that new coolpigmented asphalt shingle roofs maintain their solarreflectance as well as their standard productioncounterparts. Total color change is reduced or is atleast similar to that of conventional shingles,meaning that the new IRR products have

Xenon-arc testing of the IRR asphalt shingles showed slight increases in solar reflectance through 3000 hours of exposure (Fig. 3). For example, solar reflectance increased from 0.27 to 0.29 before leveling at about 0.28 for the code A shingle with IRR pigments. The standard shingles also showed slight increases in solar reflectance as exposure