1、OR-05-5-4 Modeling Approaches for Indoor Air VOC Emissions from Dry Building Mate ria I s-A Review Fariborz Haghighat, PhD, PEng Member ASHRAE Hongyu Huang, PhD ABSTRACT Physical models have been developed to predict volatile organic compound FOC) emissions from building materials. Accordingly, expe
2、rimental methods have also been established for measuring input parameters to these models. The purpose of this paper is to review the existing material emission models, as well as to analyze the experimental methods. It was found that existing physical models for describing VOC emission from dry bu
3、ilding material can be classified as one-phase models and multi-phase models. Further detailed analysis showed that these models could be converted.fiom one form into another as long as the linear sorption isotherm is used in both approaches. This inter-model conversion has been vali- dated through
4、comparison of the models predictions with the experimental data as well as the inter-model prediction. In addition, it was found that the impact ofmaterialporos- ity on indoor VOC concentration is significant only iftheparti- tion coeficient has the same magnitude as the material porosiy. The result
5、s also indicated that VOC concentration increases as the material porosity increases. Also, when the partition coeficient is more than one order of magnitude greater than the materialporosity, the efect ofmaterialporos- ity diminishes and its impact on the indoor VOC concentration is negligible. INT
6、RODUCTION Many of the materials used in buildings are the source of indoor air pollution, due to their large surface area and their permanent exposure to indoor air. These materials are the source of volatile organic compounds (VOC), which have been associated with certain symptoms of sick building
7、syndrome, multiple chemical sensitivity, and other health Chang-Seo Lee, PhD problems. These sources could be classified as wet materials (e.g., paint, glue, and sealant) and dry materials (e.g., dried wet materials, carpet, gypsum board, wallpaper, and vinyl and wood products). VOC emissions from w
8、et materials are char- acterized by initial high emission rates and fast decay. Surface emission usually dominates the emission process of the wet materials (Haghighat and De Bellis 1998), whereas VOC emissions from dry materials have low VOC emission rate initially and slow VOC decay rates. Hence,
9、the internal di Tiffonnet et al. 2000; Lee et al. 2002, 2003; Huang and Haghighat 2002). A more detailed description of indoor VOC convection and difision transfer as well as VOC distribution can be obtained by using CFD technique (Yang et al. 2001; Murakami et al. 2000). For an intermediate approac
10、h between total mixing and CFD models, a zonal model can be applied to describe the indoor VOC transfer (Huang et al. 2003). Basically, there are two commonly used approaches to describe VOC transfer within the solid material. The first approach assumes the material is homogeneous medium and VOC exi
11、sting within the material is in a material phase (Little et al. 1994,1996; Cox et al. 2000; Yang et al. 2001; Huang and Haghighat 2002; Haghighat and Huang 2003), while the second approach treats the material as a porous medium and VOC existing within the material are in a gas phase and an adsorbed
12、phase (Tiffonnet 2000; Murakami et al. 2000; Lee et al. 2002,2003). The magnitude of the VOC difision coeffi- cient used in the first approach is several orders of magnitude lower than the one used in the second approach. However, Haghighat et al. (2002) found that input parameters estimated using t
13、he first approach to analyze the experimental data could be converted to the input parameters required by the second approach. Therefore, additional research and analysis regard- ing these two types of modeling approach and input parameter would enhance the understanding of the fundamentals of the V
14、OC emissions from dry buildings. This paper first briefly reviews methods used to simulate air and contaminant flow within a room. It will then proceed to review methods used to simulate VOC transport within the material, and finally it will report the results of an inter-model comparison. Furthermo
15、re, a parametric study about the impact of material porosity on VOC concentration is presented in the last part of the paper. VOC TRANSFER IN ROOM AIR VOC transfer in the room air can be through convection and molecular and turbulent diffusion. Different models have been used to describe VOC transpo
16、rt in room air, and the level of modeling complexity depends on the assumption made on the extent of VOC mixing with air. According to the assump- tions made, the models can be classified as well-mixed models, CFD models, and zonal models. Well-Mixed Model In the well-mixed model, a room is treated
17、as a homoge- neous mono-zone and VOC is assumed to be instantaneously and completely mixed with the room air. VOC concentration in the room is approximately represented by a single value, and no information about airflow distribution is required to predict VOC concentration. The transient VOC mass b
18、alance for the entire room is expressed as (Huang and Haghighat 2002) dC As = NC, - NCgUs + LR(t) , dt where C, is the VOC gas phase concentration in the room air (pg/m3), C, is the room supply air VOC concentration (pg/m3), N is the air exchange rate (s-l), R(t) is the material VOC emis- sion rate
19、(pg/m2s), and L is the material loading factor (m2/m3). The well-mixed model is mostly applied for a small, mechanically ventilated room in which VOC emitted from building materials is effectively mixed with the supply air. It does not provide the detailed VOC distribution within a room. CFD Model I
20、n the computational fluid dynamics (CFD) approach, no assumption is made about the mixing of VOC and room air. In this approach the room is divided into a large number of small control volumes or elements. VOC transfer by air convection, molecular diffusion, and turbulent difision is explicitly mode
21、led for each control volume or element. Thus, informa- tion about airflow distribution is needed, and this information is obtained by solving equations expressing the conservation of momentum alone or the conservation of momentum equa- tions coupled with conservation of energy. The VOC mass conserva
22、tion at each small control volume or element is described by Murakami et al. (2000) as where uj is the velocity (ds), D, is the VOC molecular difi- sion coefficient in the air (m2/s), vt is the turbulence viscosity (m2/s), and Sc, is the turbulent Schmitt number. CFD models can provide detailed know
23、ledge of airflow, temperature, and contaminant distributions within a room. However, in most applications there is no need for such detailed information since CFD simulations are too compli- cated and time consuming to be used as a daily design tool. Moreover, the accuracy of the simulation results
24、depends on the users experience and skills in numerical simulations. Zonal Model Zonal models are intermediate models between CFD models and well-mixed models. In zonal models, a room is divided into a much smaller number of cells than for CFD. In each cell it is assumed that VOC is totally mixed. T
25、he heat and mass exchange between cells is through cell interfaces, and the mass and energy conservation principles are applied to each cell. The commonly used zonal models apply the empirical power law to calculate the airflow across cell interfaces (Wurtz et al. 1999; Haghighat et al. 2001). The V
26、OC mass transfer between cells is modeled mainly through convection and molecular difision, and the VOC mass conservation for each cell is expressed by an ordinary differential transport relation as follows (Huang and Haghighat 2003): 636 ASHRAE Transactions: Symposia from its material phase to its
27、gas phase. It is assumed that the vidCgasji dt - “voc,jfS (3) material phase concentration is always in equilibrium with the gas phase concentration. At atmospheric pressure, for low j= 1 where Cgas,i is the VOC gas phase concentration in cell i (pg/m3), mvoc,ij is the VOC mass flow across the cell
28、i and cellj interface (pds), S is the VOC source term (pgs), V, is the volume of cell i (m3), and n is the number of interfaces of cell i. The advantage of this approach lies in its relative straight- forwardness for the user to define the problem, and the formu- lated algebraic equations are relati
29、vely small and far easier to solve than the conventional partial differential equations asso- ciated with CFD methods. Therefore, compared to well-mixed models, zonal models can provide users with a global over- view of airflow, temperature, and contaminant distributions within a room. Zonal models
30、have advantages over CFD models in their simple use, time saving, and satisfactory preci- sion Characteristics (Haghighat et al. 2001). VOC TRANSFER WITHIN MATERIALS VOC transfer within materials is mainly through molec- ular difision, Knudsen diffusion, and molecular adsorption/ desorption. The phy
31、sical emission models for dry materials can be classified as one-phase models and multi-phase models. This classification is based on the assumption about VOC phase existing in dry materials. One-Phase Models Mass Transfer inside the Materials. In the one-phase models, from a macroscopic point of vi
32、ew the building mate- rial is treated as a single homogeneous medium. Since it is very hard to define the actual physical status of the VOC within the material, the VOC within the material is said to be in a material phase in order to distinguish it from its gas phase in the room air, as shown in Fi
33、gure 1. This material phase actu- ally is a lumped sum of phases of various kinds of VOC phases within the material. VOC mass transfer within the material is through diffusion. The dry material is assumed to have a homogeneous diffusivity. One-phase models use Ficks second law to describe the diffus
34、ive mass transfer within the material (Little et al. 1994, 1996; Cox et al. 2000; Yang et al. 2001; Huang and Haghighat 2002; Haghighat and Huang 2003). For one-dimensional VOC difision, it is dem aLc, - - Dm- at ay2 (4) where C, is the VOC material phase concentration in the material phase (pg/m3),
35、 D, is the VOC material phase difi- sion coeficient (m2/s), y is the coordinate in which diffusion takes place (m), and tis the time (s). Boundary Condition. Due to the large surface area within the material, VOC can be adsorbed or accumulated on the surface of the material; therefore, the VOC mater
36、ial phase concentration is always higher than the indoor VOC gas phase concentration. At the material/air interface, VOC changes VOC concentration and isothermal conditions, the equilib- rium relationship between VOC concentration in the gas phase and VOC concentration in the material phase can be d
37、escribed by a linear isotherm (Axley 1991): = KmCgus + (5) Cm ly = b- ly=b is the VOC material phase concentration at is the VOC gas phase the materiaf ; therefore, for a new building material, it is assumed that the VOC concentration is uniform inside the material. O = c, Cm LO (7) where Cm is the
38、initial VOC material phase concentration (pg/m3). Multi-Phase Models Mass Transfer inside the Materials. Multi-phase models describe VOC transfer in a porous building material, which consists of voids (pores) and solid parts. VOC in a pore Figure 1 VOC within material and room air: ASHRAE Transactio
39、ns: Symposia 637 is in a gas phase and/or adsorbed phase on the pore surfaces, as shown in Figure 1. VOC transfer within the material is mainly by gas-phase diffusion (i.e., molecular and/or Knudsen diffusion) through the pores, and the adsorbed-phase diffu- sion, or diffusion through the solid part
40、s, is assumed to be negligible (Tiffonnet et al. 2000; Murakami et al. 2000; Lee et al. 2002,2003). Lee (2003) further developed the multi-phase approach, including surface difision (adsorbed-phase diffu- sion) as well as gas-phase diffusion in the pores. The multi- phase models also use Ficks secon
41、d law to describe VOC gas- phase difision within the materials. The following equation is usually used to describe the one-dimensional problem considering only the gas-phase diffusion and sorption: where Cgus is the VOC gas-phase concentration in the material (pg of gas phase VOC/m3 of air), Ca, is
42、the VOC adsorbed- phase concentration in the material (pg of adsorbed phase VOC/m3 of material), DFUs is the effective VOC gas-phase diffusion coefficient within the material (m2/s), and E is the porosity of the material (m3 of air/m3 of material). In the multi-phase approach, it is assumed that the
43、 gas- phase VOC is always in equilibrium with the adsorbed-phase VOC, since the time scale of adsorptioddesorption is much faster than the time scale of gas-phase diffusion. At indoor low VOC concentration and isotherm conditions, the linear adsorption isotherm is usually adopted to describe the rel
44、a- tionship between the gas-phase concentration and the adsorbed-phase concentration, which is given as (9) where K is VOC adsorbed/gas-phase partition coefficient. Substituting Equation 9 into Equation 8 gives Nonlinear sorption isotherms such as Freundlich, Langmuir isotherms, can also be implemen
45、ted in the multi-phase approach (Tiffonnet et al. 2000; Murakami et al. 2000; Lee 2003). Boundary Condition. Since VOC in the room air is in a gas phase, VOC gas-phase concentration at the material/air interface is continuous and given by The VOC mass transfer rate at this boundary is also continuou
46、s: Initial Condition. For a new building material, the initial VOC gas-phase concentration within the material is The previously proposed physical models for dry building materials are summarized according to their approaches of internal diffusion and surface emission in Table 1. These emis- sion mo
47、dels can be generally applied for sink behavior, except some models with zero concentration in the air such as the analytical models by Yang et al. (1998) and Huang and Haghighat (2002). CORRELATIONS BETWEEN ONE-PHASE MODELS AND MULTI-PHASE MODELS One of the critical parts for successful model predi
48、ction lies on the availability and correctness of the model input parameters. One-phase models use three material property parameters to express the VOC material-phase mass transfer process within the material: VOC material-phase difision coefficient, D,; VOC material-phase/gas-phase partition coeff
49、icient, K,; and initial VOC material-phase concentration, Ci. But multi-phase models utilize four material property parameters to describe the VOC gas-phase mass transfer mechanism within the material: material porosity, E; VOC gas- phase diffusion coefficient, Dgas; VOC gas-phase/adsorbed- phase partition coefficient, K; and the initial VOC gas-phase concentration, C, . The following sections discuss the parameter relationship between the one-phase model and the multi-phase model. Material Phase vs. Gas Phase In the one-phase model, it is assumed that material is homogeneous a
