Study of penetration of outdoor fine particles into nonresidential building with multizone simulation.
Particulate matter (PM), such as dust, dirt, soot, smoke, and liquid droplets found in the air, varies in size. Some particles are large or dark enough to be seen by human eyes. Others are so small that they can only be detected with an electron microscope (EPA 2006). Depending on the diameters, particles can be suspended in the air for a period of time and thus treated as airborne (Liu and Zhai 2006). In most cases, particles less than 10 micrometers (3.94 X [10.sup.-4] inches) in diameter (PM10) are airborne and can remain in the air for a long time. These particles can be inhaled into and accumulate in the respiratory system due to their small size, posing a serious health concern. In particular, particles less than 2.5 micrometers (9.94 X [10.sup.-5] inches) in diameter (PM2.5) are referred to as "fine" particles. Because of their small size (approximately 1/ 30th the average width of a human hair), fine particles can lodge deeply into the lungs. Exposure to fine particles can cause short-term health effects, such as eye, nose, throat, and lung irritation; coughing; sneezing; runny nose; and shortness of breath. Lung function will be affected and worse medical conditions, such as asthma and heart disease, might occur. Scientific studies have linked increase in daily PM2.5 exposure with aggravation of respiratory and cardiovascular disease (as indicated by increased hospital admissions, emergency room visits, absences from school or work, and restricted activity days). Evidence has even shown a significant association between exposure to fine particles and premature death (EPA 2006). Individuals particularly sensitive to fine particle exposure include older adults, people with heart and lung disease, and children.
There are outdoor and indoor sources of fine particles. Outdoors, fine particles primarily come from car, truck, bus, and off-road vehicle exhausts (e.g., construction equipment, snowmobile, locomotive) and other operations that involve the burning of fuel or vegetation. Fine particles may also form from the reaction of gases or droplets in the atmosphere. These chemical reactions can occur miles from the original sources of the emissions. Because fine particles can be carried long distances from their sources, events such as wildfires or volcanic eruptions can raise fine particle concentrations hundreds of miles from the event. P[m.sup.2]5 can also be produced by common indoor activities. Typical indoor sources of fine particles are tobacco smoke, cooking (e.g., frying, sauteing, and broiling), burning candles or oil lamps, and operating fireplaces and fuel-burning space heaters (e.g., kerosene heaters).
Significant health risks associated with exposure to fine particles, together with the fact that a large portion of total personal exposure to ambient particulate matter occurs in indoor environments because people spend 80%-90% of their time indoors, result in renewed interest in indoor fine particle concentrations and their relationship to outdoor particles. Understanding this relationship is especially important because most of the particle exposure studies in the literature occurred indoors, while the measured particle concentrations were obtained from ambient monitoring stations. The inference has been made that the concentration of indoor fine particles that originated outdoors is comparable with the outdoor concentration. facciola (2006) conducted a literature review on researchers' efforts on revealing the relationships. for instance, Colome et al. (1992), Thatcher and Layton (1995), Thornburg et al. (2001), and Chao (2001) studied how ambient PM infiltrates into indoor environments and the roles of building shells and central air-cleaning systems in influencing the process. Spengler et al. (1981), Colome et al. (1992), Clayton et al. (1993), fischer et al. (2000), and Long and Sarnat (2004) measured actual indoor and outdoor concentrations in residential environments and showed that indoor PM concentrations are consistently lower than ambient concentrations without the presence of indoor sources. It has been revealed that indoor concentrations are strongly correlated with outdoor levels of P[m.sup.2]5 (Long and Sarnat 2004) and even more so when the air exchange rate is high (Phillips et al. 1993).
This research investigated the penetration mechanism of outdoor fine particles into large nonresidential buildings. Special building structures, such as atrium, stairwell, and elevator space and complicated HVAC systems for nonresidential buildings, impose extra complexities and challenges to identify the particle indoor/outdoor correlations. To understand and correlate the indoor and outdoor fine PM concentrations for large nonresidential buildings, indoor and outdoor fine particle concentrations have been measured for four nonresidential buildings in Colorado (Facciola 2006). This paper focuses on developing appropriate multizone simulation models and methods that can rapidly and reliably depict overall building ventilation and infiltration characteristics and reveal indoor/outdoor contaminant relationships.
Multizone network airflow models allow engineers to predict infiltration, ventilation rates, contaminant transfer, and pressure differences in buildings. In multizone models, buildings are represented as networks of connected zones (e.g., rooms) in which uniform air proprieties and contaminant concentrations are assumed. Zones are connected to each other and to the outdoors via paths that represent the building component leakages. A mass balance calculation determines the flow between the zones as well as concentrations of contaminants distributed by this flow. Due to the integral form of the governing equations and the uniform zonal property assumption, multizone models can provide quick solutions with reasonable accuracy.
Although the mathematics behind these models is well understood, it is challenging to determine proper model input parameters, particularly for envelope leakage characteristics. When a building is modeled during the design stage, it is common to specify building leakage and ventilation characteristics according to the design documents. This designbased prediction may significantly deviate from reality because (1) buildings have many leakage paths that are not shown in the design drawings, (2) the leakage associated with identical components may vary significantly with construction, and (3) the designed ventilation capacity may not be fully utilized. This has to be dealt with by analyzing a potential range of the parameters to obtain a bounded solution. for existing buildings, these parameters may be measured to fine-tune the simulation model. Blower door test, tracer gas test, and flow rate measurement are three common techniques that can provide actual building leakage and ventilation conditions. However, both blower door and tracer gas tests are fairly difficult for large buildings with many compartments.
This paper documents the modeling process of a threestory institutional building with fine-particle concentration measurements conducted both indoors and outdoors for a continuous week. The facility was first modeled using standard leakage data and the original design plans. The ventilation flow rates were then refined according to actual measurement during the week. The building leakage coefficients were calibrated based on the measured indoor and outdoor concentration data at night when the HVAC system was shut down. The calibrated model can then be used to predict indoor fine particle dispersions under real weather and outdoor particle conditions. This research used the public domain multizone network airflow model CONTAMW developed by the US National Institute of Standards and Technology (Dols et al. 2000).
The study used a campus building at the University of Colorado (CU) at Boulder to develop proper simulation models and methods. The building is a three-story, 3159 [m.sup.2] (34,000 [ft.sup.2]) integrated teaching and learning (ITL) facility (Figure 1). Completed in 1997, the concrete-walled ITL building is a shared classroom facility for the six departments of CU-Boulder's College of Engineering and Applied Science. This state-of-the-art facility features design studios, an active learning center, and an extensive computer network integrating a wealth of experimental equipment throughout the two large, open laboratory plazas, team workrooms, and manufacturing centers, including a machine shop and three project fabrication facilities.
[FIGURE 1 OMITTED]
The building has a large two-story atrium (laboratory plazas). The floors in the majority of the facility are tiled, while the study rooms and classrooms are carpeted. The air-handling unit serving the building is a variable-air-volume (VAV) system with a design supply volume capacity of 22.8 [m.sup.3]/s (48,300 cfm). The unit runs on a preprogrammed direct digital control (DDC) system, the operating schedule depending on the occupancy schedule. This air-handling unit is unique in that it utilizes a direct evaporative cooling section in addition to the indirect cooling coil. The evaporative cooling is used because the humidity in Colorado is typically low, making this type of cooling more cost-efficient. However, when the humidity in the air increases, the indirect cooling coil is automatically activated. The supply and return fans are, respectively, 40.8 kW (54.7 HP) with a speed of 884 rpm and 22.6 kW (30.3 HP) with a speed of 725 rpm. The efficiency of the supply air filter is around 40%.
The facility has as many as 150 students scheduled to be in the laboratory plazas and classrooms at one time for various classes or projects. The building has two main operating schedules--occupied school year and unoccupied summer months. During the school year, the HVAC system is turned on most weekdays between 6:00 a.m. and 7:00 a.m. and turned off at 6:00 p.m. in the summer and 10:00 p.m. in all other seasons. The weekend system schedule is about 11:00 a.m. to 3:00 p.m. When the building is occupied, the discharge air temperature is set based on the outside air temperature (OAT) according to two schemes. In the first scheme, when the OAT exceeds 55[degrees]F, the discharge air temperature will be 12.8[degrees]C (55[degrees]F). In the second scheme, when the OAT falls below -6.7[degrees]C (20[degrees]F), the discharge temperature will be 18.3[degrees]C (65[degrees]F). During the occupied period, the supply fan speed is controlled to maintain a 248.8 Pa (1.0 in. w.g.) static pressure in the supply duct. The return fan speed is controlled to maintain a 24.9 Pa (0.1 in. w.g.) in the return air plenum. The exhaust air damper position is adjusted to maintain a static pressure of 12.5 Pa (0.05 in. w.g). The outside air dampers are controlled by the DDC to maintain a minimum fresh airflow of 4.1 [m.sup.3]/s (8700 cfm). When the building is unoccupied, the fans are deenergized and the outside and exhaust air dampers are closed, while at the same time the steam coil control valve is opened to the coil. The radiation heating valves are controlled to maintain a temperature of 18.3[degrees]C (65[degrees]F) or above in each zone. This study measured and simulated the building performance in a spring week between April 6 and April 12, 2006.
The experimental study continuously measured the indoor and outdoor fine particle (PM_0.7) concentrations by using an aerosol mass spectrometer (AMS) and an ultra high sensitivity aerosol spectrometer (UHSAS) (Facciola 2006). The size of the fine particles (PM_0.7) varies from 55 to 700 nm (2.16 X [10.sup.-6] to 2.75 X [10.sup.-5] in.; 0.055 to 0.7 [micro]m), with an average of about 0.4 [micro]m (1.57 X [10.sup.-5] in.). The indoor sampling points were located at the main level (second floor) laboratory plaza (Figure 2). The plaza consists of 15-20 computer workstations and is open to the high ceiling approximately 12.2 m (40 ft) above. The floor is tiled, and there are several sealed windows in the plaza. The fabrication facilities and machine shop are not located on this level. One indoor sampling line was hung by the yellow support beam seen in Figure 2a, near a corner of the laboratory space. The other indoor line was routed to the stairwell between the second and third floors of the building, near the study rooms and open to the main laboratory plaza (Figure 2b). The average of the readings from the two indoor lines gives the fine particle concentration in the atrium. The outdoor measurement point was the HVAC intake plenum on the roof, sampling air just before it entered the air-handling unit. All sampling equipment was kept in the mechanical room. The experiment also recorded the total supply, outside, return, and exhaust airflow rates for the building with the built-in sensors. During the testing week, the average outdoor temperature was 8.9[degrees]C (48[degrees]F) with a significant swing from 23.5[degrees]C to -2.1[degrees]C (74.3[degrees]F to 28.2[degrees]F). The maximum wind speed reached about 13.9 m/s (31.1 mph), while the average wind speed was about 5 m/s (11.2 mph).
[FIGURE 2 OMITTED]
FIRST STAGE OF SIMULATION--DESIGN DOCUMENT MODEL
The modeling study first simplified the building with the original building and system design plans. The research modeled the special building structures (e.g., atrium, stairwell, and elevator shaft) with particular flow path models in CONTAMW, among which the atrium was treated as a well-mixed zone. An HVAC system was incorporated into the model, for which the supply and return airflow rates for each diffuser and grille were specified according to the design conditions. A fixed outside air ratio of 15% was applied in the model. The HVAC systems use the pleated media type of filters. Since the average diameter of measured particles is around 0.4 [micro]m (1.57 X [10.sup.-5] in.), the efficiency of the filters was set at 40% in the simulation model as suggested by Annis (1991).
Two types of leakage factors were considered to represent the total leakage characteristics of the building in the model. They are easily identified leakage sites (e.g., doors and windows) and distributed envelope leakages (e.g., exterior and interior wall constructions and other details). The 2001 ASHRAE Handbook--Fundamentals does not provide leakage data for windows and doors of nonresidential buildings. The door and window leakage data of this model were taken from the ASHRAE residential database (ASHRAE 2001). In a study of eight US office buildings, Persily and Grot (1986) found the air leakage rate ranging from 1080 to 5220 [cm.sup.3]/(s*[m.sup.2]) (0.213 to 1.028 cfm/ft) at a pressure difference of 74.6 Pa (0.3 in. of water) (ASHRAE 2001). These correspond to the distributed envelope leakages of 0.00347 to 0.0139 [dm.sup.2]/[m.sup.2] (0.005 to 0.02 [in.sup..2]/[ft.sup.2]) at a reference condition of 4 Pa (0.016 in. of water) and discharge coefficient [C.sub.D] = 1 (ASHRAE 2001). As a result, an averaged distributed envelope leakage of 0.00854 [dm.sup.2]/[m.sup.2] (0.0123 [in.sup..2]/[ft.sup.2]) was introduced to the first simulation model. Interior partitions, ceilings, and floors were modeled with twice the leakage for the exterior walls (Musser et al. 2001). A deposition rate sink was placed in each zone to represent the overall zone deposition effect of PM_0.7 particles. The deposition rate was determined as 8.33 [s.sup.-1] according to particle deposition research on residential buildings by Howard-Reed et al. (2003). Figure 3 shows the layout of zones, leakage paths, HVAC diffusers, and special architectural features for a portion of the third floor of the building.
[FIGURE 3 OMITTED]
The study used the first model to predict the indoor airflow and fine particle transport patterns with the real outdoor weather and fine particle conditions during the experimental week. The simulation assumed that the outdoor fine particle is the only source of the indoor fine particle, which is quite reasonable for this educational building, which does not have typical indoor particle sources such as tobacco smoke, cooking devices, fireplaces, and fuel-burning space heaters. Figure 4 shows the simulated and measured results in which a significant discrepancy is observed. This leads to the necessity of calibrating the model.
[FIGURE 4 OMITTED]
SECOND STAGE OF SIMULATION--SUPPLY AIR REFINEMENT WITH MEASUREMENT
Two primary parameter uncertainties for the current model are ventilation supply conditions and building leakage characteristics. Building ventilation and leakage are exactly the penetration paths of outdoor particles into the building. When the building is air conditioned, ventilation systems deliver fresh air from outside to inside, playing a dominant role in carrying fine particle indoors. The VAV ventilation systems vary supply airflow rates dynamically with room conditions and operation schedules, not always at full design capacity. It is thus necessary to refine the supply air volume flow rates and fresh air percentages with the actual conditions. In practice, it is much easier to record the total supply, return, and fresh airflow rates rather than measuring the supply and return conditions of individual vents in each zone. Figure 5 shows the variation of total airflow rates in seven days measured by this study.
[FIGURE 5 OMITTED]
The actual total supply flow rates were then compared against the design values to generate an actual total supply air schedule (ratio), as shown in Figure 6. The actual supply air schedule was used for every diffuser in all zones to produce the actual supply airflow rates. This implies a linearity assumption for air distribution behavior in duct networks. Similarly, an actual fresh air fraction schedule can be calculated by dividing the total outside airflow rate by the total supply airflow rate of the same time, as shown in Figure 7. These schedules were input to the simulation model to account for the real-time supply air conditions for each zone when the HVAC system was on during the day. The return air schedule was assumed to be the same as the supply air schedule. Besides the zones served by the main HVAC system, several special "service" areas located at the southeast corner of each floor of the building are ventilated by local exhaust fans. These areas were simulated in the model as zones served by another simple air-handling system with only exhaust air. The total direct-exhausted airflow rates are very small and there are no schedules associated with them for simplification.
[FIGURE 6 OMITTED]
The simulated results shown in Figure 4 reveal an apparent improvement in modeling the daytime concentration variations when the mechanical system was on. However, significant flat results are observed during the nights when the system was off, which indicates that an inappropriate leakage or infiltration property might be assigned to the building by using the standard data.
THIRD STAGE OF SIMULATION--BUILDING LEAKAGE CALIBRATION WITH MEASUREMENT
When the mechanical system is turned off during the nights, infiltration through building leakage areas becomes the primary approach for outdoor fine particles penetrating into the building. To find proper building leakage characteristics requires extensives field expertments, such as fan pressurization/depressurization tests and tracer gas tests. However, it is usually difficult and challenging to apply these test technologies for a large nonresidential building with many compartments and special design features. By contrast, it is relatively easy to measure point contaminant concentrations indoors and outdoors. This study investigated the feasibility and the process of using these point concentration measurements to adjust building leakage conditions so as to match simulation with experiment.
Measurement results for the second night of the week were chosen to calibrate the building leakage properties. Through simultaneously adjusting flow resistances of all the flow paths by multiplying a certain factor, a reasonable match is achieved between the simulation and experiment (Figure 8). The flow path characteristics then remained fixed for predicting the rest of the days and nights. The calibration revealed that the exterior envelope of the building was about six times leakier than the initial values assumed according to the 2001 ASHRAE Handbook--Fundamentals. As a consequence, the distributed exterior envelope leakage of the model was modified from 0.00854 to 0.0525[dm.sup.2]/[m.sup.2] (0.0123 to 0.0757 [in.sup..2]/[ft.sup.2]) to match the experimental results. Persily (1999) examined the air leakage values by fan pressurization testing for 139 commercial buildings (office, school, or retail facilities) in the US, Canada, and UK and indicated an average distributed envelope leakage of 0.062 [in.sup..2]/[ft.sup.2]. This number is close to the value (0.0525 [dm.sup.2]/[m.sup.2] [0.0757 [in.sup..2]/[ft.sup.2]]) identified by this research. In the simulation, the interior leakage was concurrently adjusted to maintain the ratio of 2:1 between interior and exterior envelope leakage as suggested by Musser et al. (2001).
[FIGURE 7 OMITTED]
[FIGURE 8 OMITTED]
Figure 9 shows the improvement of the calibrated simulation. The second-stage model predicts fairly flat curves at night, while the third-stage model reflects the measured indoor particle concentration swing trend. The figure depicts the validation of the new model by comparing measured data with simulation results for all the testing days. Although calibrated solely by one night of experimental data, the simulation model provides overall good results for all the other nights and the days, which validates the final model as well as the calibration approach.
[FIGURE 9 OMITTED]
The calibrated model can then be used to predict the fine particle indoor distribution patterns in the whole building and during arbitrary time periods as long as the outdoor weather and particle conditions are available. It is almost impossible to conduct such thorough field tests for different spaces and different time periods because of the cost and time restrictions. The simulation results are valuable for identifying potential correlations between indoor and outdoor contaminant concentrations for different types of nonresidential buildings (e.g., school and office buildings). As an example, Figure 10 shows the fine particle concentration variation during the testing week for an office room at the northeast corner of the second floor predicted by the three models. Obvious result disparities between the three models are observed, justifying the importance of the model calibration and validation.
[FIGURE 10 OMITTED]
RESULTS AND DISCUSSION
This study reveals the effectiveness of using a multizone model to simulate contaminant dispersions in a complex nonresidential building. Figure 9 shows a good match between the simulation and measurement results for most of the days. However, discrepancies do exist for some time periods of the week. As illustrated in Figure 9, the left and middle vertical lines, labeled P1 and P2, respectively, indicate the starting points of two major discrepancies. The right vertical line P3 shows another odd point where a sudden indoor concentration increase was observed while outdoor fine particle concentration was decreasing. Since this engineering teaching laboratory facility is almost free of various typical indoor fine particle sources, penetration of outdoor fine particles via building ventilation and infiltration is the primary cause for indoor particle fluctuation. P1 occurred at 7:00 a.m. on a Saturday when the HVAC was only on from 10:30 a.m. to 3:00 p.m. The measurement data show that the indoor fine particle concentration slowly increased for the whole night but came to a sudden steep drop after 7:00 a.m. This implies a sudden appearance of an indoor fine particle sink, which is almost impossible for such an early Saturday morning with no occupant and no active ventilation. P2 occurred at 3:00 p.m. on a Sunday when the HVAC system was already shut off. According to the mass balance principle and the facts that infiltration-induced airflow rates are far less than those driven by mechanical fans and the volume of the atrium is huge, the airflow pattern of the atrium would be relatively small inflows from certain neighboring zones and corresponding small outflows to other adjacent zones. Since all the zones in the building have already established certain fine particle concentration levels after one day's operation of the HVAC system, it is anticipated that a slow varying of the atrium fine particle concentration would occur after the system is off, gradually following the outdoor concentration fluctuation due to the infiltration and cross-flows between zones. The phenomenon was predicted by the simulation but was not observed in the experiment, which actually found a sudden drop of the atrium fine particle concentration at P2 and a steep rise at P3 during the night. Hence, there is a chance that measurement errors may have happened during that period or some unknown activities or weather conditions occurred. In fact, the experimental devices were very sensitive to environmental conditions.
To quantify the calibration performance, the normalized mean square error (NMSE) was employed to estimate the magnitude of the difference between the predicted ([C.sub.p]) and measured ([C.sub.m]) concentrations. NMSE is defined as
NMSE = [[bar.[([C.sub.p] - [C.sub.m]).sup.2]]/[[bar.[C.sub.p]] * [bar.[C.sub.m]]]], (1)
where [bar.C.sup.p] and [bar.C.sup.m] stand for the average of predicted and measured concentrations. NMSE equals zero when the predicted and measured concentrations are the same. The smaller the NMSE is, the closer the predicted and measured concentrations are.
Excluding the problematic data set (from 7:00 a.m. Saturday to 1:00 p.m. Sunday), the NMSE values for the three simulation models are calculated as presented in Table 1. It is clear that the third-stage model provides the best prediction.
Table 1. Calculated NMSE Values for Three Simulation Models Model First-Stage Model Second-Stage Model Third-Stage Model NMSE 4.901 0.594 0.079
The study shows that properly defining fan-driven flows are crucial to the simulation in which the mechanical system is used, while reasonable building leakage characteristics are important for simulating buildings with passive ventilation and infiltration. Relatively simple concentration measurements can be valuable for calibrating simulation models and are especially appropriate for complex buildings where both fan pressurization tests and tracer gas tests are difficult to implement. Most of today's new buildings conduct real-time monitoring of supply and return airflow rates, which can be directly used to refine simulation models.
This study used a multizone modeling program to simulate the penetration of outdoor fine particles into indoor spaces of a large nonresidential building. In comparison to the measurement, the simulation predicted fairly good results for both mechanically ventilated daytimes and infiltration-dominated nighttimes. In the simulation, proper specification of fan-driven ventilation airflow rates and building leakage characteristics is crucial. The research indicates that simulation with design and/or standard data may not provide good results because buildings may deviate from the design (or standard) conditions for diverse reasons. Individual calibration with experimental data is necessary to obtain a reasonably accurate prediction.
The study found that proper specification of the fandriven airflow rates is most important for achieving good simulation results during the day when the mechanical system is on, while properly defining the exterior envelope leakage is more critical for the night when the system is off. This study presents a relatively straight measure to conduct a multizone airflow and contaminant simulation for large multicompartment buildings, in which both blower door and tracer gas tests are difficult to perform. To characterize the primary airflow and contaminant spread patterns, only two sets of measurement are needed: the inside and outside contaminant concentrations and the total airflow rates, which can be used to calibrate and validate the actual building leakage characteristics and supply conditions.
The authors would like to thank the American Society of Heating, Refrigerating and Air-Conditioning Engineers (ASHRAE) for supporting and funding this project under contract 1281-RP.
Annis, P.J. 1991. Fine Particle Pollution--Residential Air Quality. North Central Regional Publication.
ASHRAE. 2001. 2001 ASHRAE Handbook--Fundamentals. Atlanta: American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc.
Chao, C.Y.H. 2001. Comparison between indoor and outdoor air contaminant levels in residential buildings from passive sampler study. Building And Environment 36:999- 1007.
Clayton, C.A., R.L. Perritt, E.D. Pellizzari, K.W. Thomas, R.W. Whitmore, L.A. Wallace, H. Ozkaynak, and J.D. Spengler. 1993. Particle total exposure assessment methodology (Pteam) study--Distributions of aerosol and elemental concentrations in personal, indoor, and outdoor air samples in a Southern California community. Journal of Exposure Analysis and Environmental Epidemiology 3:227-50.
Colome, S.D., N.Y. Kado, P. Jaques, and M. Kleinman. 1992. Indoor outdoor air-pollution relations--Particulate matter less than 10 mu in aerodynamic diameter (Pm-10) in homes of asthmatics. Atmospheric Environment, Part A--General Topics 26:2173-78.
Dols, S., G. Walton, and K. Denton. 2000. CONTAMW User Manual, NISTIR 6476. Gaithersburg, MD: National Institute of Standards and Technology.
EPA. 2006. PM 2.5 designation, Frequent questions. US Environmental Protection Agency. www.epa.gov/pm designations/faq.htm#2.
Facciola, N. 2006. Identification, classification, and correlation of ultrafine indoor airborne particulate matter with outdoor values. Master's thesis. University of Colorado at Boulder.
Fischer, P.H., G. Hoek, H. van Reeuwijk, D.J. Briggs, E. Lebret, J.H. van Wijnen, S. Kingham, and P.E. Elliott. 2000. Traffic-related differences in outdoor and indoor concentrations of particles and volatile organic compounds in Amsterdam. Atmospheric Environment 34: 3713-22.
Howard-Reed, C., L. Wallace, and S.J. Emmerich. 2003. Deposition rates of fine and coarse particles in residential buildings: Literature review and measurements in an occupied townhouse, NISTIR 7086. Gaithersburg, MD: National Institute of Standards and Technology.
Liu, X., and Z. Zhai. 2006. Identification of appropriate CFD models for simulating aerosol particle and droplet indoor transport. Submitted to Indoor and Built Environment (under review).
Long, C.M., and J.A. Sarnat. 2004. Indoor-outdoor relationships and infiltration behavior of elemental components of outdoor Pm 2.5 for Boston-area homes. Aerosol Science And Technology 38:91-104.
Musser, A., O. Schwabe, and S. Nabinger. 2001. Validation and calibration of a multizone network airflow model with experimental data. SIM Conference, Session 6, June 13-14.
Persily, A. 1999. Myths about building envelopes. ASHRAE Journal 41(3):39-47.
Persily, A.K., and RA. Grot. 1986. Pressurization testing of federal buildings. Measured Air Leakage of Buildings, ASTM STP 904, H.R. Trechsel and P.L. Lagus, eds., pp. 184-200. American Society for Testing and Materials. Phillips, J.L., R. Field, M. Goldstone, G.L. Reynolds, J.N. Lester, and R. Perry. 1993. Relationships between indoor and outdoor air-quality in 4 naturally ventilated offices in the United Kingdom. Atmospheric Environment, Part A--General Topics 27:1743-53.
Spengler, J.D., D.W. Dockery, W.A. Turner, J.M. Wolfson, and B.G. Ferris. 1981. Long-term measurements of respirable sulfates and particles inside and outside homes. Atmospheric Environment 15:23-30.
Thatcher, T.L., and D.W. Layton. 1995. Deposition, resuspension, and penetration of particles within a residence. Atmospheric Environment 29:1487-97.
Thornburg, J., D.S. Ensor, C.E. Rodes, P.A. Lawless, L.E. Sparks, and R.B. Mosley. 2001. Penetration of particles into buildings and associated physical factors. Part I: Model development and computer simulations. Aerosol Science And Technology 34:284-96.
Joe Huang, Lawrence Berkeley National Laboratory, Berkeley, CA: Calibration is being done on airflow rates, but comparisons are being done on concentrations. What are the chances that fine-particle concentrations are decoupled from air movement and flow rates under certain regimes?
Zhiqiang Zhai: As mentioned in the first paragraph of the paper, particles less than ten micrometers in diameter (PM10) are airborne and move well with airflow. Fine particles fall into this category. Hence, it is reasonable to assume fine particle concentration depends on air movement and airflow rates. In addition to air movement, the study did consider some other effects that may influence the concentration, such as deposition and filtration. Also, as Professor Ezzat Khalifa points out, fine particle resuspension due to human activities might result in the concentration disparity between the modeling and the measurement.
H. Ezzat Khalifa, Professor, Syracuse University, Syracuse, NY: (1) Did you consider the filtering effect of the walls on infiltration air? This could be done in CONTAM by introducing a size-dependent filter efficiency for the infiltration air.
(2) Human activities result in PM resuspension. This may explain part of the disparity between the model and the measurement at night when there is no resuspension due to human activity.
Zhai: (1) The paper does not consider the filtering effect of the walls on infiltration air. Since infiltration mainly happens at night when the HVAC system is off, taking into account this effect might help explain why the model predicts larger concentrations than the measurement at night.
(2) Good comment. Lack of fine particle resuspension from human activity might be another reason for the disparity between the modeling and the measurement at night. By incorporating these two types of particle effects, an even better match between the model prediction and the field measurement could be anticipated.
Zhiqiang Zhai, PhD
Nick A. Facciola
Shelly L. Miller, PhD
Xiang Liu is a student and Zhiqiang Zhai is assistant professor in the Department of Civil, Environmental and Architectural Engineering and Nick A. Facciola is a student and Shelly L. Miller is associate professor in the Department of Mechanical Engineering, University of Colorado, Boulder, CO.
|Printer friendly Cite/link Email Feedback|
|Author:||Liu, Xiang; Zhai, Zhiqiang; Facciola, Nick A.; Miller, Shelly L.|
|Article Type:||Technical report|
|Date:||Jul 1, 2007|
|Previous Article:||Application of neural networks trained with multizone models for fast detection of contaminant source position in buildings.|
|Next Article:||Exergy analysis of psychrometric processes for HVAC&R applications.|