1、2009 ASHRAE 793ABSTRACT An interesting approach to reduce building energy consumption is to use green roofs as a part of building enve-lope. However, many building designers ignore this opportu-nity as it is quite difficult to estimate the resulting energy saving. This paper provides results from an
2、 ongoing experi-mental research project that focuses on the thermal perfor-mance of extensive green roofs when buildings are in the cooling mode. The paper discusses the importance of green roofs and reviews previous research studies. In particular, this paper focuses on the role of plants for the h
3、eat flux reduction through the roof structure. The performance of the plant mate-rial was assessed in an environmental chamber by experiments with two samples, one with the plant material, and another one without the plant material. Overall, plants reduced the measured heat flux through the green ro
4、of sample by 40-50% compared to the roof sample without plants. In conclusion, plants have an important role in reducing the heat flux by regu-lating: (1) latent heat flux through better water management and additional water storage in the plant leaves/roots, and (2) sensible heat flux through addit
5、ional shading provided by the plant leaves. Based on these results, future research will focus on thermal modeling of green roof including the role of plants.INTRODUCTIONGreen roofs are an emerging sustainable technology that is becoming more popular in North America (Miller et al., 2005). As a defi
6、nition, green roofs are “specialized roofing systems that support plant growth on rooftops” (Liu et al., 2004). From top to bottom, a typical green roof consists of several layers: (1) vegetation, (2) substrate, (3) filter membrane, and (4) drainage layer. Plants used for extensive green roofs are t
7、ypically drought tolerant, and selected from the group of native or Sedum plants. Substrate is the soil-like layer where plants grow, and it has to be porous, retain mois-ture and nutrients, and support plant growth (Snodgrass et al., 2006). The filter membrane prevents drainage clogging by containi
8、ng the substrate and roots. The drainage layer trans-ports the rainfall water runoff to the roof drainage (Peck, 2002; Snodgrass et al., 2006).There are basically two types of green roofs: extensive and intensive green roofs. Extensive green roofs have lower weight, lower capital cost, minimal maint
9、enance, and a substrate depth between 2 and 6 inches (5 and 15 cm). Inten-sive green roofs have greater weight, higher capital costs, wider planting selection, higher maintenance requirements, and a substrate depth between 8 and 24 inches (20 and 60 cm). However, intensive green roofs are less cost-
10、effective than extensive and required more structural support (Peck et al., 1999; Tanner, 2004). Moreover, extensive green roofs repre-sent about 2/3 of the total green roof square footage installed in North America (Johnston, 2007). Therefore, this research project focuses on summer thermal perform
11、ance of extensive green roofs as a more economically viable solution to be adopted in the building industry. The popularity of green roofs is increasing due to their potential benefits. In general, green roofs have a potential to (Liu et al., 2004):reduce energy demand on space conditioningreduce st
12、orm water runoffimprove air quality, and reduce the urban heat island effect in cities.The Role of Plants in the Reductionof Heat Flux through Green Roofs: Laboratory ExperimentsPaulo Cesar Tabares-Velasco Jelena Srebric, PhDStudent Member ASHRAE Member ASHRAEPaulo Cesar Tabares-Velasco is a graduat
13、e student and Jelena Srebric is an associate professor at the Department of Architectural Engineering, The Pennsylvania State University, University Park, PA.LO-09-076 2009, American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. (www.ashrae.org). Published in ASHRAE Transact
14、ions 2009, vol. 115, part 2. For personal use only. Additional reproduction, distribution, or transmission in either print or digital form is not permitted without ASHRAEs prior written permission.794 ASHRAE TransactionsTherefore, green roofs can help address three of the four top problems facing th
15、e society in the next 50 years: energy, water, and environment (Smalley, 2005).GREEN ROOF ENERGY BALANCEThe main challenge of accurately modeling and measur-ing the thermal performance of green roofs is due to the complex heat and mass transfer processes through the roof by the means of (1) shading,
16、 (2) insulation, (3) evapotranspira-tion, and (4) thermal mass (Liu, 2004). Evapotranspiration represents a combined process of water lost from the soil (evaporation) and plants (transpiration). Transpiration occurs when water from the plant leaf surface goes into the air by diffusion or convection.
17、 Most of the water is lost by transpi-ration through plant stomata, which are adjustable small pores in the leaf that allow the entry of gases needed for photosyn-thesis such as CO2, and the release ofO2and water vapor. Thus, plants can control their transpiration rate by opening and closing their s
18、tomata (Nobel, 1983; Allen et al., 1998; Hillel, 1998).All of the heat transfer processes taking place on a green roof are combined in an energy balance equation as following (Hillel, 1998; Jones, 1992):(1)whereRn= net radiation, equal to solar gain minus infrared heat losses, Btu/hft2(W/m2)ET = eva
19、potranspiration, or latent heat flux, Btu/hft2(W/m2)Qsensible= convective or sensible heat flux, Btu/hft2(W/m2)Qcondution= heat flux trough roof, Btu/hft2(W/m2)Sthermal= thermal storage for substrate, plants, Btu/hft2(W/m2)M = metabolic storage (photosynthesis and respiration), Btu/hft2(W/m2)In Equa
20、tion 1, the metabolic storage is often neglected as its contribution to the total energy budget is around 1% to 2% of the net radiation (Jones, 1992; Gates, 1980). All fluxes are dependent greatly on the capacity of the plants/substrate to evaporate water as latent heat flux uses energy from the env
21、i-ronment to evaporate water, thus cooling down the plants surface and roof temperature. Consequently, the latent heat flux partially controls the heat flux going trough the substrate/roof that eventually converts to building cooling loads. Evapo-transpiration is also a phenomenon that enables green
22、 roofs to decrease the urban heat island effect by lowering the temper-ature of the roofs by evaporation of the rain water.As an example of energy balance, Figure 1 shows the percentages of each heat transfer component divided by the incoming shortwave radiation for a particular experimental setup w
23、ith a green roof. Experiments are described in more details later in the paper, while these results are presented here for the illustration of typical heat fluxes and heat balance on a green roof. Table 1 provides a summary of the total heat flux components for the two different sets of experiments,
24、 one with the plants and another without plants. Because all fluxes were divided by the incoming shortwave radiation having the same units, the results are dimensionless. The sum of all heat fluxes shown in Table 1 was very close to 100% for the experiments without plants. In contrast, the sum of he
25、at fluxes for the experiments with plants was systematically lower than 100% by 10-15% due to the assumption of a horizontal flat plate for the convective heat transfer and simplified infrared radiation model. For proper comparison, the sums of all heat fluxes shown in Table 1 were normalized to 100
26、% for both cases. As shown in the table, the latent heat flux played an important role in the heat transfer process by diverting from the roof about 55% to 80% of the incoming shortwave radiation to the process of evapotranspiration. It is important to mention that these percentages were obtained fo
27、r a wet green roof, while for a dry green roof these percentages are substantially lower. The thermal performance of green roofs has been studied worldwide using three different approaches: (1) field or labo-ratory experimentation, (2) theoretical studies, and (3) a combination of laboratory or fiel
28、d experiments with numerical models. From these three approaches, only the field and labo-ratory experiments have focused on comparing energy fluxes from bare soil surface to planted surfaces. Most of the green roof field studies have focused on heat flux reduction through the roof. Interestingly, a
29、n on-site study found a significant reduction of heat flux from a green roof compared to a bare soil roof (Wong et al., 2003). The study concludes the difference was due to the shading of plants because the heat flux at night was mainly the same for both roofs. Another study came to similar conclusi
30、ons by analyzing an irrigated bare soil roof, and then adding a shading device over the roof (Pearlmutter et al., 2008). Our present study finds RnET QsensibleQconductionSthermalM+ +=Figure 1 Heat fluxes in a green roof for a typical summer day with percentage of heat fluxes normalized with the inco
31、ming shortwave radiation as collected in our experimental setup.ASHRAE Transactions 795that it is not just the shading, but also the evapotranspiration (latent heat flux) that gets alerted by the presence of plants. This finding was possible because we conducted tightly controlled laboratory experim
32、ents with laboratory graded instrumentation.A laboratory study inside a greenhouse compared the latent heat flux of green roof samples with other samples with-out plants (Rezaei, 2005; Berghage et al., 2007). The latent fluxes were measured based on gravimetric method that continuously weigh the sam
33、ples after initial watering. The highest latent heat flux, which corresponds to the highest weight loss, was during the first day of measurements with a value around 111 Btu/h ft2(350 W/m2). This peak latent heat flux coincides with the peak solar radiation, while the soil was the wettest during tha
34、t day. The latent heat flux for the sample with no plants was about half the value compared to the sample with plants (Berghage et al., 2007). For the sample with plants, Figure 2 shows latent heat fluxes during the first four days of measurements for summer conditions in Pennsylvania (Rezaei, 2005)
35、. It is important to notice that the latent heat flux decrease substantially from day to day as the available water content decreased. LABORATORY EXPERIMENTSControlled laboratory experiments are proposed as a solu-tion to understand non-steady state heat transfer phenomena through a green roof. Afte
36、r gaining insight from previous on-site and laboratory studies, our research team designed and built a new experimental apparatus called “Cold Plate” to test the thermal performance of green roofs. The design of the Cold Plate was inspired by ASTM standards C177 (ASTM 1997a) and C1363 (ASTM 1997b),
37、and later modified based on experiments with non-homogeneous samples (Tabares-Velasco et al., 2007). A detailed description of the experimen-tal apparatus is available in the literature (Tabares-Velasco et al., 2007). Most importantly, tests performed with the Cold Plate are conducted under tightly
38、controlled conditions inside a state-of-the-art environmental chamber. This chamber contains data acquisition systems that measure and/or control the energy consumption, air quality, and thermal comfort of different heating, ventilating and air conditioning systems (HVAC). Figure 3 shows locations o
39、f several data acquisition sensors installed in green roof samples to measure heat fluxes, irradiance, volumetric water content (VWC) in substrate, air and substrate temperatures, and air speed. Both latent and conductive heat fluxes are measured by two independent approaches to add redundancy and c
40、heck the accuracy of both measurement methods. Incoming shortwave radiation was calculated using measurements from a second-ary class pyranometer. Absorbed shortwave radiation was calculated from the measured incoming radiation and variable albedo for the specific wavelength of the lamps obtained fr
41、om the literature for different soils and plants (Escadafal, 1990; La et al., 2008; Gates, 1980). Convective heat flux was calculated assuming turbulent natural convection on a flat horizontal plane because the Raleigh number was in the order of 107. Infrared heat transfer was calculated assuming tw
42、o large parallel planes, and used temperature readings at the top layer of substrate and plant leaves, as well as temperature readings for the lamp surface. Lamp temperature was measured by a thermistor located near the light bulbs. Another set of therm-istors located under the plant leaves measured
43、 the surface temperature of the leaves. Finally, a set of thermistors located under a thin layer of substrate measured the top substrate temperature. Substrate and plants emissivity was set to 0.95 (Pielke, 2005; Gates, 1980; Nobel, 1983). Lamp emissivity was set to 0.90 (Incropera et al., 2002).Tab
44、le 1. Percentage of Heat Fluxes Relative to Incoming to Shortwave Radiation Obtained inCurrent Experiments, Day TwoNo Plants PlantsIncoming Shortwave Radiation 100% 100%Infrared Radiation/Incoming Shortwave Radiation 3% 24%Reflected Shortwave Radiation/Incoming Shortwave Radiation 5% 11%Latent Heat
45、Flux/ Incoming Shortwave Radiation 56% 82%Sensible Heat Flux/ Incoming Shortwave Radiation 20% 16%Conductive Heat Flux/Incoming Shortwave Radiation 22% 15%Figure 2 Green roof latent heat fluxes for early summer conditions in Pennsylvania (Rezaei 2005; Berghage et al. 2007).796 ASHRAE TransactionsExp
46、erimental ProcedureTwo green roof samples were tested under similar envi-ronmental conditions. Both samples had inner dimensions of 48x42x3 in3(122x107x9 cm3). The samples contained 3 in. (9 cm) deep green roof substrate, which consists mainly of expanded clay typically used in the green roof indust
47、ry. As shown in Figure 4a and 4b, the first sample used bare substrate, without any plants, while the second sample was completely covered with Sedum Spurium. Sedums are succulent plants, and have the ability to limit their water loss due to transpiration (VanWoert et al., 2005). The sample without
48、plants was tested first, and the sample with plants was tested one week after the first set of experi-ments. Reflective side panels were added to the Cold Plate for the sample without plants to improve the uniformity of the irradiance. Both samples were watered 48 hours and 24 hours before testing t
49、o allow for proper saturation of the substrate and successive drainage of excess water. The sample without plants was tested for 6 days to observe the drying process of substrate to reach very low water content. The sample with plants was tested for 3 days, and further drying would cause damage to plants, which would eventually start to behave as a bare substrate sample. For each day the lamps were on approx-imately 14 hours, and were switched off for another 10 hours to allow transition between “day” and “night” environmental conditions. The chamber environmental controls w
