1、4838 Simulated Influence of Roof Reflectance on the Building Energy Balance in Two Northern Cities Sebastian Freund Daniel J. Dettmers Member ASHRAE ABSTRACT Solar irradiation on the roof surfaces of buildings affects cooling and heating energy demand. In this study the effect of enhanced roof rejec
2、tance is investigated using computer simulations ofa prototype large retail building with four roof insulation levels (R-4 through R-24) driven by hourly weather dataforhrlinneapolis, MN. The dark roofsurface temperatures in the summer reach more than 160F (71 “C); the white roof reaches about 100F
3、(38“C), leading to 6%-17% lower cooling energy consumption. Although the increased heating costs in the winter negate a portion of the cooling savings, the total energy cost savings of the white roof are still positive. Also, a reduction in the maximum hourly compressorpower-respon- sible for the mo
4、nthly billed electric peak demand-of 0.06 to 0.3 WR (0.6 to 3 W/m2) can be consistently attributed to the high albedo roo$ The demand cost savings may be in excess of the energy cost savings. The total annual savings a white roof can attain are in a range between $27/1,000 ft ($0.3/m2) and $10/1,000
5、 ft ($0.1/m2) for R-4 and R-24 insulation. For Denver, CO., which has a warmer climate and is located farther south than Minneapolis, limited sirnulation results indicate total savings twice as high. With higher insulation levels, savings ofa white roof compared to a black roof are lower, but total
6、cost savings of increased roof albedo can be higher than the cost savings of increased insulation. Also, the effects ofsnow cover and the operation of an outside air econ- omizer on the energy savings ofa reflective roof are investi- gated. INTRODUCTION Reflective roof coatings on commercial and res
7、idential buildings have the potential for decreasing cooling energy Douglas T. Reindl, PhD Member ASHRAE consumption and demand by lowering roof temperatures. The energy savings and potential for downsizing cooling equip- ment is due to the lower average roof temperatures obtained with roofing mater
8、ials that have higher solar reflectance. During the heating season a reflective roof will also lower the roof temperature (relative to a non-reflective roof), thereby increasing the heating load and creating a trade-off between cooling energy savings and heating energy penalty. The effect is greates
9、t on low-rise buildings with large flat or south-sloped roofs, typically commercial, institutional, and industrial facil- ities that would otherwise have dark bitumen or sheet metal roofing. Unconditioned attic spaces or high roof insulation values mitigate the effect. The trade-off of cooling savin
10、gs versus heating penalty had not been extensively studied in previously published liter- ature for a northern climate such as Minneapolis. This study was conducted to determine the overall energy consumption in northern climates for heating and cooling when a reflective roof is used. This paper sum
11、marizes the results of analysis using a computer simulation model built in TRNSYS, which is a simulation program specifically designed to model transient heat transfer processes and solar systems. The program includes all necessary routines to effectively calculate solar irradiation on buildings as
12、well as models of building and HVAC components. Simulations are driven by TMY2 weather data for Minneapolis. LITERATU RE REVI E W In previous studies on roof reflectance and building energy usage, authors agreed that roofs with higher reflectance could deliver space conditioning system energy saving
13、s in Sebastian Freund is a graduate student at the Institute for Thermodynamics, Department of Mechanical Engineering, Helmut-Schmidt- University of the Federal Armed Forces, Hamburg, Germany. Daniel J. Dettmers is a research engineer and Douglas T. Reindl is center direc- tor at the HVAC Akbari 199
14、8; Parker, Sherwin et al. 1998). Prior simulation results were supple- mented using experiments to demonstrate the significant roof surface temperature decreases with correspondingly lower cooling energy consumption in Florida (Akridge 1998; Parker, Huang et al. 1998). In warm climates, the savings
15、in air-condi- tioning costs exceed the increased heating costs. For colder climates, it has been considered that the opposite may be the case. Akbari et al. (1998) conclude that highly reflective roofs have a negative impact on the annual building energy costs in Minneapolis. Positive net savings, h
16、owever, may occur in those climates when the HVAC system size is reduced due to a smaller summer peak load, resulting in smaller fans and higher part-load ratio operation for a greater proportion of the year. From Parker, Huang et al. (1 998) it can be concluded that reflective roofs on residential
17、buildings in Minneapolis have a negative impact (if any) on annual energy costs. However, the different methods applied to analyze energy consumption do not yield consistent results (Hildebrandt et al. 1998) and DOE- 2-based simulations may significantly underestimate the effect of roof reflectance
18、on energy usage, including an under- prediction of the savings by as much as twofold (Akbari 1998). In addition, the results may only be applicable to the certain type of analyzed building. For the northern climates, the ques- tion of whether or not reflective roofs offer positive net energy savings
19、 and, if so, for which types of buildings had not been conclusively determined. INFLUENCING PARAMETERS The relation of savings from enhanced roof reflectance in the summer versus losses in the winter is dependent upon the following factors: Roof orientation: Depending on the slope and the time of ye
20、ar, south-sloped roofs (in the northern hemisphere) can receive higher radiation intensity than flat roofs; plus, sloped roofs have less snow coverage. Sunshine: The number of sunshine hours in the winter and the summer affect the maximum heating benefit and cooling savings potential, respectively.
21、When the roof is snow covered, the effect of the roof reflectance is negated. Consequently, the penalty of a reflective roof in the winter may be smaller for locations with a long snow season. Roof construction and insulation: The higher the R- value of the roof, the lower the transmission gains; th
22、erefore, there is less influence of the roof temperature on the building energy usage. A roof construction with high thermal capacity, such as thick concrete, will delay the solar heating effect and diminish the effect of roof reflectance. Internal gains and length of cooling season: During the cool
23、ing season, any additional heat gain (or reduc- tion of heat loss) due to increased roof temperature increases the cooling load. If the cooling season is very long, either due to warm climate or to high internal gains (commercial and industrial buildings), a reflective roof will tend to provide high
24、er savings. MODEL SUMMARY The building modeled is a large, single-story (i.e., “big- box”) retail store with a height of 25 ft, a length of 3 16 ft, and a gross area of 100,000 ft2. The four external walls have an R- value of 11 h-ft2-F/Btu. A total of 5% of the wall area is composed of single-pane
25、windows with internal shading of The internal gains of the building originate from sources such as lights and other electric equipment, as well as humans, and are scheduled in accordance with the building occupancy. A lighting intensity of 2.5 W/ft2 is used for the simulation. The electric equipment
26、 in the building other than lights is assumed to operate 24 hours a day at a rate of 0.3 W/ft2. The number of persons in the facility varies hourly and is different for weekdays and weekends. An occupancy schedule provided by a big box retailer was used in the simulation. The maximum number of peopl
27、e in the facility during the open hours is 400; a person density of 4/1,000 fi2 is assumed. The heat generation per person for the activity level expected of a customer in a store is assumed to be 3 15 Btu/h sensible and 325 Btu/h latent, based on IS0 Standard 7730 (IS0 1994). The building modeled i
28、n this simulation has a large inter- nal heat gain. Applying the results ofthis simulation to a build- ing without a significant internal load, such as a residential building, may not yield similar results to those presented here. This analysis compares the modeled hourly energy consump- tion for he
29、ating and cooling. The parameters varied are the R- value of the insulation installed in the built-up roof (from R- O to R-24), the roof albedo (black = 6% vs. white = 65%), and the presence of snow cover on the roof. The snow cover model is based on hourly weather data and also accounts for addi- t
30、ional melting due to the heat flux through the roof. It is important to emphasize our definition of an “R-24” roof as a sheet metal roof with an EPDM membrane and suffi- cient levels ofan insulation material that adds an R-value of24 h-ft2-F/Btu (nominal value). The actual overall heat transfer resi
31、stance of the roof includes the variable inside and outside convective and radiative heat transfer coefficients; conse- quently, the overall R-value used is updated during each time step in the simulation based on prevailing outdoor and indoor conditions. 80%. HVAC System The mechanical systems serv
32、ing the simulated facility are typical constant-volume packaged rooftop units. The heating stipend of the building is constant at 72F without humidifi- cation, while the cooling stipend is 76F with a maximum rela- tive humidity of 60%. The supply air rate in the building is 1.5 cfm/ft2, a typical va
33、lue for large retail facilities to meet cooling peak demand. The outside airflow rate is 0.2 cfm/ft2, a value in accordance with ASHRAE Standard 62-2001 for retail 172 ASHRAE Transactions: Research stores, malls, and arcades (ASHRAE 2001). During unoccu- pied hours, the outside air ventilation is se
34、t to a minimum value of 2.5% of the supply air, or 0.0375 cfm/ft2. The fan power is estimated based on the assumption of a total static pressure drop of 2 in. H,O, a total supply flow rate of 150,000 cfm with a fan efficiency of 0.5 (motor 0.85, fan 0.6). These assumptions lead to a total fan power
35、of 70 kW, which adds to the internal gains. Heat is provided by gas-fired units in the rooftop units. The heating units are assumed to have a constant efficiency of 0.8, based on the lower heating value. The rooftop units assumed in this study operate with a full-load cooling EER of 9.0 (COP = 2.64)
36、. The EER includes the supply fan electric power as well as the degradation of the cooling capacity due to the heat added by the fan. The EER for rooftop units is established in accordance with ART Standard 340/360 (AR1 2000). Under part-load conditions, when the cooling demand of the building is si
37、gnificantly lower than the capaciy of the rooftop units, there are three options to meet the reduced demand: the compressors can be unloaded to a certain level, the compressors can be cycled on and off, or the total load can be split between the rooftop units servicing the same thermal zone so that
38、only a few units operate under high load while others remain off. The first two options, unloading and cycling, compromise the compressor and refrigeration system efficiency. This, however, is partially compensated for by the general coincidence of lower outside air temperatures, leading to better h
39、eat rejection capaciy and lower condensing pres- sure. The calculation of exact part-load efficiency requires detailed compressor and system modeling beyond the scope of this investigation, but this will not have any significant impact on the results. However, actual energy savings of a highly refle
40、ctive roof could be somewhat lower than expected due to longer operation with lower part-load ratio when reducing cooling loads while leaving the equipment and design capacity the same. The costs are evaluated based on the 2004 general service rate for commercial customers of the local utility in Mi
41、nneap- olis. The electric rate is $0.03 lkWh with demand charges of $6.61/kW (October to May) and $9.26kW (June to Septem- ber); the natural gas rate is $0.481/therm. Heat Transfer The convective heat transfer coefficient on the outside surface of the roof is dependent on the wind speed. Limited res
42、earch has been conducted in the area of building envelope convective heat transfer, and various existing correlations yield conflicting results (Beckman and Duffie 1991; ASHRAE 1997). The approach made here is the combination of a Nusselt number correlation for turbulent flow over a hori- zontal fla
43、t plate (Incropera and DeWitt 2002) and a constant term for natural convection at low wind speeds (Beckman and Duffie 1991). The Nusselt number is given as 0.8 0.33 Nu = Re Pr , while the free convection constant is estimated as 5 W/m2K (0.88 Btu/h-ft2-F). A curve fit of these equations using aver-
44、age air properties yields a convection coefficient of h, = max5,2.5 + 1.19vl (2) with the convection coefficient h, (W/m2K) and the wind speed v (ds). The inside convective heat transfer coefficient on the roof surface is usually not of great significance since the thermal resistance of the wall is
45、often very high compared to the added resistance of the convection coefficient. For simulations, the coefficient is often set to a default value of 0.5 Btu/h-ft2-F (3 W/m2K). However, since in this project a detailed analysis of the roof is desired, the inside convection coefficient is not constant
46、but calculated based on airflow conditions and temperatures. In an experimental investigation of heat transfer coeffi- cients on the walls of a typical office room with ceiling venti- lation, a correlation was developed that relates the convection coefficient solely to the ventilation rate (Fisher a
47、nd Pedersen 1997). In that equation, h is the convection coefficient on the ceiling in W/m2K and ACH is the air change rate per hour based on the supply airflow rate and the room volume: h = 0.49ACH0“ (3) However, these results are assumed to have limited rele- vancy in a larger facility due to the
48、special configuration of the room and diffuser and due to the negligence of free convec- tion, the effect of different room and ceiling temperatures. Assuming undisturbed free convection dominates the heat transfer from the ceiling to the room, the coefficients will depend on the temperature differe
49、nce of the ceiling surface and bulk room air. Furthermore, the coefficients will be differ- ent for a cool ceiling (winter) and a warm ceiling (summer). The buoyancy forces occurring with the density differences of colder and warmer air enhance convection when heavier, cooled air is sinking and warm air is rising toward the cooler ceiling (winter). When the ceiling is warmer than the room air (summer), the buoyancy suppresses convection, resulting in stratification disturbed only by the induced draft from supply air diffusers pending under the ceiling and the return air intakes. The
