Impact of moving objects on contaminant concentration distributions in an inpatient ward with displacement ventilation.
Air movement in indoor environments is strongly linked to the transmission and spread of airborne infectious diseases, such as measles, tuberculosis, chickenpox, influenza, smallpox, and SARS (Li et al. 2007). The lack of knowledge of and insufficient data on ventilation requirements in hospitals, schools, and offices make it difficult to understand the spread of airborne infectious diseases (Li et al. 2007; Beggs et al. 2008). A study by Yin et al. (2009) showed that ventilation systems played a very important role in the transmission of exhaled particles from a patient to a caretaker in the same ward. They showed that displacement ventilation can provide much better indoor air quality than overhead mixing ventilation. The ventilation effectiveness of a displacement ventilation system with a reduced ventilation rate of 4 ach can be the same as that of an overhead mixing ventilation system with a ventilation rate of 6 ach. However, there were concerns about whether an object moving, such as a visitor or caretaker walking, changing a sheet on the patient's bed, or a swinging entrance door in an inpatient ward, could destroy the stratified flow created by the displacement ventilation and thus decrease the ventilation effectiveness (Brohus et al. 2006; Bj[empty set]rn and Nielsen 2002; Bj[empty set]rn et al. 1997). Mazumdar (2009) found that a moving passenger in an aircraft cabin could carry a contaminant in his or her wake to positions far from the contaminant source. Thus, it is essential to assess the impact of moving objects on contaminant transmission in inpatient wards with displacement ventilation.
An investigation of the impact of moving objects on airborne contaminant transmission inside an inpatient ward can be made experimentally or through computer simulations. Acquiring experimental data with meaningful temporal and spatial resolution is difficult and time consuming (Poussou 2008; Thatcher et al. 2004). Computer simulations using computational fluid dynamics (CFD) are an efficient alternative and can provide high-resolution data because CFD simulations can handle directly or indirectly the movement of an object in an inpatient ward. The indirect methods model the movement approximately, such as by using a distributed momentum source (Zhai et al. 2002) or a turbulent kinetic energy source (Brohus et al. 2006). These indirect methods need little computing time, but they generate some uncertainties, and the results may not be accurate due to the bold assumptions used. In contrast, the direct methods use moving and dynamic grids to simulate the movement. The accuracy is greatly improved, but unfortunately, the methods are computationally demanding.
To reduce the computing demand, a CFD model with combined dynamic and static mesh scheme can be used (Mazumdar and Chen 2007). The CFD model uses dynamic meshes for regions where the movement takes place and static meshes for the rest of the computational geometry. This CFD model was validated with experimental data for airflow in an airliner cabin to be reliable and, therefore, appropriate for the present study.
The CFD model used a second-order upwind scheme and the SIMPLE algorithm. The renormalization group (RNG) k-[epsilon] model was used to model the turbulent flow inside an inpatient ward. Compared to other turbulence models, the RNG k-[epsilon] model was one of the best in terms of accuracy, computing efficiency, and robustness for modeling indoor environments (Zhang et al. 2007). This study used a commercial CFD program, FLUENT (Ansys 2003). The CFD model was used to calculate the distributions of air velocity, air temperature, gaseous contaminant concentration, air pressure, and turbulence parameters. The contaminant concentration was normalized using
Normalized C = [C.sub.p]-[C.sub.s]/[C.sub.e]-[C.sub.s] (1)
where [C.sub.p] is the contaminant concentration at the point of interest. [C.sub.s] and [C.sub.e] are the steady state contaminant concentrations at the supply inlet and at the exhaust outlet, respectively.
VALIDATION OF THE CFD METHOD
In order to verify that the CFD model can also be used in an inpatient ward with displacement ventilation, this investigation first used the measured data of air velocity, air temperature, and contaminant concentration as simulated by a tracer gas (Yin et al. 2009) from a mockup of an inpatient ward as shown in Figure 1 for the validation. Figure 1 shows that the ward was furnished with one bed, a TV set, and a piece of medical equipment. This ward had one patient lying on the bed and one caretaker standing on the right side of the patient. The air was supplied at 4 ach from the diffuser located near the floor on the opposite wall from the patient. The exhausts were located at the ceiling level on the wall near the patient's head and on the wall adjacent to the bathroom. The corresponding ventilation rate was 78 cfm (37 L/s). The thermofluid boundary conditions used for the CFD mockup are shown in Table 1, which were close to the measured data. The heat released from patient, caretaker, equipment, and TV was 106 W, 110 W, 36 W and 24 W respectively. The air velocity and air temperature were measured using omnidirectional anemometers at 56 locations along different heights in Poles 1-8, as shown in Figure 1b. The accuracy of the measurements was 4 fpm (0.02 m/s) for velocity and 0.45[degrees]F (0.25[degrees]C) for air temperature. Yin et al. (2009) used continuous sources of an [SF.sub.6] tracer gas and particles (1 [micro]m and 3 [micro]m in size) to investigate the airborne contaminant dispersion inside the inpatient ward. The corresponding contaminant concentration profiles in the ward looked the same for the tracer gas and particles. Hence, the contaminant was simulated for simplicity by using the tracer gas in this investigation. The [SF.sub.6] concentrations were measured at 30 locations along different heights in Poles TG 1-5 (see Figure 1b). The measurement accuracy of [SF.sub.6] concentration was 0.001 ppm. Experiments were conducted at three different times, and the repeatability was good.
[FIGURE 1 OMITTED]
Table 1. Boundary Conditions for the Inpatient Ward Supply airflow rate 4 ach Supply air temperature 19.5[degrees]C Ceiling surface temperature 22.5[degrees]C Surface temperature of the sidewalls 22.5[degrees]C Surface temperature of the TV 24.8[degrees]C Floor temperature 22[degrees]C Surface temperature of the caretaker 29.1[degrees]C Surface temperature of the patient 31.3[degrees]C Surface temperature of the equipment 25[degrees]C Surface temperature of the visitor (Case 1) 29.1[degrees]C
Figures 2 and 3 compare the profiles of air velocity and air temperature at Poles 1-8 (see Figure 1b), respectively. The height, velocity, and temperature were nondimensionalized with respect to the height of the inpatient ward (H), diffuser face velocity ([u.sub.s]), inlet ([T.sub.s]) and main exhaust ([T.sub.e]) temperatures. The repeatability of the experimental data is evident from the plots. The computed air velocity and temperature profiles were in good agreement with the measured data. The temperature stratification is clear in the results.
[FIGURE 2 OMITTED]
[FIGURE 3 OMITTED]
Figure 4 further compares the measured and computed profiles of the normalized [SF.sub.6] concentrations at Poles TG 1-5 (see Figure 1b). The computed concentration at Poles 4 and 5 was higher than at the other poles. Our past experience found that it is very difficult to obtain a good agreement on contaminant concentrations calculated and measured, but discrepancies are acceptable for such a study. Again, the displacement ventilation shows very clear contaminant stratification.
[FIGURE 4 OMITTED]
Though the CFD validation for the inpatient ward was performed for a steady-state case, it could be used for cases with moving objects. For example, our group previously validated the CFD model in an airliner cabin under steady-state conditions (Zhang et al. 2009). The model performed similarly for waterflow in a small-scale experimental model airliner cabin with transient conditions (Mazumdar 2009). Thus, the CFD program FLUENT with the RNG k-[epsilon] model was used to further study the effect of moving objects on transient contaminant distributions in the inpatient ward.
To study the impact of moving objects on contaminant transmission and ventilation system performance in an inpatient ward, four different scenarios of moving objects were used, as shown in Figure 5:
Case 1: Visitor moving along the foot of the patient bed from left to right. The visitor moved at a speed of 200 fpm (1 m/s) for 3 s and then stopped. The thermal boundary conditions used for the moving visitor were the same as those for the caretaker (refer to Table 1).
[FIGURE 5 OMITTED]
Case 2: A caretaker moving backwards and forwards along the bedside. The caretaker started his or her movement near the center of the bed at a speed of 200 fpm (1 m/s) towards the head of the patient for one second, then moved in the opposite direction for two seconds, and finally came to his/her initial position by moving for another second.
Case 3: Changing a sheet by moving it up at 200 fpm (1 m/s) for 0.75 seconds and then back down to the original position at the same speed over the patient.
Case 4: Swinging the entrance door by opening it inward and then immediately closing it. The duration of the opening and closing processes was two seconds each at constant speed.
As the CFD simulations used a combination of dynamic and static meshes, only 2.1%, 1.5%, 1.1%, and 1.0% of the meshes inside the inpatient ward were dynamic for Cases 1, 2, 3, and 4, respectively. The maximum mesh size in the ward was 2 in. (0.05 m). The total number of meshes used for the ward was approximately 1.2 million. User-defined functions were implemented in FLUENT to describe and track the object's movements. Each case was computed using eight processors (1.8 GHz AMD 64) and 4GB of memory. The smallest time step used during the computations was 0.01 s. The residual for mass (sum of the absolute residuals in each cell/the total mass inflow) and the residual for energy (sum of the absolute residuals in each cell/the total heat gains) was less than 0.1% and 1%, respectively, for the computations. The normalized residual for the contaminant transport equation was below 1.0 x [10.sub.-5]. The initial conditions inside the ward were steady state with a continuous passive release of the tracer gas from the mouth of the patient. This implies that the tracer gas had zero velocity. Gupta et al. (2009, 2010) found that the momentum of exhalation was important for coughing but not significant for breathing and talking. Thus, the passive release could be acceptable. Each case was computed for two minutes of real time in order to obtain the transient contaminant distributions due to the body movement. The actual computing time took about two weeks for each case.
Contaminant Distributions in the Inpatient Ward with a Moving Object
Figure 6 shows the normalized distributions of the contaminant concentration across the plane of a visitor moving track in Case 1 at t = 0, 1, 2, 3, 60, and 120 s, respectively. The contaminant distribution at t = 0 s was that for the steady state with a continuous contaminant release from the mouth of the patient. The contaminant concentration was low in the lower part of the ward, which is typical for displacement ventilation. The moving visitor could bring the contaminant in the wake. The contaminant concentration distribution at t = 120 s was not the same as at t = 0 s, since the initial and final conditions were different due to the visitor relocating. For this cross section, the contaminant concentration did not change much between t = 60 s and t = 120 s.
[FIGURE 6 OMITTED]
Figure 7 shows the normalized distributions of the contaminant concentration in the ward with the forward and backward motion of the caretaker (Case 2). As with Case 1, the wake of the caretaker induced the contaminant generated from the patient bed during his/her movement. Once the caretaker stopped, the contaminant concentration returned to steady-state level in approximately 90 s.
[FIGURE 7 OMITTED]
Figure 8 depicts the contaminant concentration distributions through a plane of the sheet movement above the patient in Case 3. When the sheet was moved up, it also induced the contaminant in the wake. Once the sheet was moved down over the patient, the contaminant induced was pushed out through both sides. The contaminant concentration returned to the initial level of the steady state in approximately 50 s.
[FIGURE 8 OMITTED]
Figure 9 illustrates the contaminant concentration distributions in the inpatient ward with the swinging movement of the entrance door in Case 4. The concentration distributions were at the height of breathing level of a standing person. The concentration distributions changed with the processes of the door opening and closing. But as soon as the door was closed, the concentration quickly returned to that of the steady state within 10 to 15 s. Compared to the other cases, the movement of the swinging door made the least impact on the contaminant distributions in the ward, as the door was far away from the contaminant source.
[FIGURE 9 OMITTED]
Table 2 summarizes the time needed to eliminate the effect of various object movements in the ward with displacement ventilation on the contaminant concentration distributions. Due to the relocation of the visitor, the thermofluid boundary conditions were different between the initial state and the final state. For Cases 2, 3, and 4, the time needed to return to the original state was between 10 to 90 s. Such a short change was unlikely to cause a major change in the exposure risk of the visitor or caretaker to the contaminant concentration.
Table 2. Time Taken to Eliminate the Impact of Various Object Movements in the Inpatient Ward in Contaminant Concentration Case Time (s) 1 Not applicable 2 90-100 3 50-60 4 10-15
Contaminant Concentration at Breathing Levels in the Inpatient Ward
A quantitative comparison of the effect of various movement cases on contaminant transmission is presented in Figures 10-13. Case 1 was also computed with an overhead air supply diffuser that supplied air at 6 ach. This is a conventional mixing ventilation system currently used in many wards. To compare the ventilation performance of the displacement ventilation cases with 4 ach, this paper shows the averaged contaminant concentrations at the breathing level of sitting and standing positions--i.e., at the breathing levels of a sitting person (3.7 ft [1.1 m] above the floor) and a standing person (5.3 ft or [1.6 m] above the floor).
[FIGURE 10 OMITTED]
[FIGURE 11 OMITTED]
[FIGURE 12 OMITTED]
[FIGURE 13 OMITTED]
Figure 10 shows the variation of averaged concentration at the breathing level for all cases. A normalized concentration was used for comparison. When the normalized concentration is 1.0, it is perfect mixing for an overhead ventilation system at 6 ach. If the normalized concentration for the displacement ventilation cases at 4 ach is the same as that for the overhead ventilation case at 6 ach, the air quality becomes the same with the same strength of the contaminant source. With the overhead ventilation, the movement of the visitor did not cause a notable swing on the averaged contaminant concentration at the two breathing levels. This is because the air velocity was generally high and the contaminant concentration was uniform. Thus, the wake of the moving visitor did not make an impact on the contaminant concentration distribution in the ward. However, the displacement ventilation system was more sensitive to the moving objects. The variation of the averaged contaminant concentration was within 25% for all cases studied. At the breathing level in a sitting position, the averaged contaminant concentration with displacement ventilation was very low. However, at the breathing level in a standing position, the displacement ventilation system with the moving objects could lead to a higher or a lower contaminant concentration in the ward. Since the changes only lasted for less than 90 s, a slight increase or decrease in the contaminant concentration level likely would not have a notable impact on the risk level of the visitor or caretaker. This conclusion can be further confirmed by using the accumulated contaminant inhaled by the visitor or caretaker during this transitional period for the four cases.
In order to further evaluate the risk level, Figure 11 depicts the breathed contaminant concentration at the sitting and standing levels of the caretaker and the visitor for the movement cases. The study was mainly for standing visitors and the caretaker, but the results for the sitting level were also provided for reference. Again, the variation of the contaminant concentration breathed by the visitor and caretaker due to their body movements was the least for the overhead ventilation case with 6 ach. The concentration near the breathing level of the visitor and the caretaker was generally lower with the displacement ventilation system. In Case 2 with displacement ventilation, the caretaker moved and his/her wake carried the contaminant, which caused a very significant change in the concentration. A similar phenomenon happened near the moving visitor in Case 1 but was not as significant as in Case 1, due to the visitor not being as close to the contaminant source as was the caretaker. At most times, the displacement ventilation system provided a better air quality than the mixing ventilation system. The very large swing in the concentration level was generally short (about 30 s), and the averaged value did not change much, so it was less likely to cause a higher risk for the visitor or caretaker in the displacement system compared to in the mixing system.
Further quantitative analysis was performed at the pole positions TG 1-5 (see Figure 1b). Except for TG 4, all the other positions were around the contaminant source. Figure 12 illustrates the normalized contaminant concentration at the breathing levels at those positions. The contaminant concentration at the breathing level of a sitting position in the ward with the displacement ventilation at 4 ach was much lower than that with the perfect mixing ventilation system at 6 ach (normalized C = 1). However, at the breathing level of a standing person, the displacement ventilation system led to a poorer air quality than that of the perfect mixing ventilation. The main reason was that the locations were closed to the contaminant source. This was also confirmed by the experimental data for steady-state cases (Yin et al. 2009). For TG 4, it was located in a dead zone, so the concentration was especially high. If the exhaust location were moved to a wall next to the TG 4 location, the concentration could become lower. Moreover, the contaminant concentration was unstable, especially at Pole TG 5 in Case 1, where visitor movement was simulated. A spike in contaminant concentration was observed for Case 3 with the changing of the patient's sheet. This happened as the entrapped contaminant below the sheet in Case 3 was pushed out through the sides as the sheet came down. The contaminant pushed out was convected upwards due to the effect of the ventilation system, which caused an abrupt increase in the contaminant concentration. As observed before, the movement of the door did not make much impact on the contaminant concentration at the pole locations.
It might be subjective to emphasize the concentration levels at Poles TG 1-5, since they were mainly around the contaminant source. Figure 13 shows the fraction of the space where the normalized contaminant concentrations at the breathing levels in the ward with displacement ventilation at 4 ach were less than 1, or those with perfect mixing ventilation at 6 ach. The figure shows that in most areas of the ward, the air quality was better than in the perfect mixing condition, although the perfect mixing condition had a much higher ventilation rate.
It should be noted that the study presented in this paper was conducted with limited conditions. For example, this investigation did not study different postures of the patient and caretaker; did not vary the walking speed or the speed of the changing sheet; did not simulate the actual exhalation with a variable momentum; and did not simulate other exhaling activities, such as talking, sneezing, and coughing. Thus, the subject deserves further study.
One should not extend the conclusions obtained from this study to other conditions that were not studied. For example, Qian et al. (2006) studied a two-patient award with different postures and found that the contaminant could be locked in the breathing zone of a person lying down. Since this study was only for one person lying down, the phenomenon was not produced.
This study investigated the effects of moving objects on contaminant concentration levels in a single inpatient ward with displacement ventilation. The investigation was conducted by assuming a constant contaminant generated from the mouth of a patient lying on the bed and by using a validated CFD program. The study assumed a 4 ach ventilation rate for the displacement ventilation. To compare with a conventional design, this investigation also used a case of overhead mixing ventilation with a 6 ach ventilation rate.
This study considered four different moving objects in the ward for up to four seconds of movement: a visitor walking near the foot of the bed, a caretaker walking alongside the bed, the changing of the sheet on the patient's bed, and the swinging of the door. This study found the following to be true:
1. Moving objects can carry a contaminant in their wake. The movement can cause a swing in the contaminant concentration at the breathing level of sitting and standing positions for 10 to 90 s.
2. The variation of the averaged contaminant concentration due to the moving objects was within 25% for all cases studied. The closer the location of the moving object to the contaminant source, the larger was the change in the contaminant concentration. The variation at the breathing level was lower in the sitting position than in the standing position. Since the variation only lasted for less than 90 s and the averaged contaminant concentration did not change much, the variation would not likely change the risk level in the ward.
3. At most times, the displacement ventilation system with 4 ach ventilation rate provided a better air quality than the mixing ventilation system with 6 ach. The walking visitor and caretaker could experience a large swing in the concentration level due to the body movement in the ward. However, the swing was generally short (about 30 s), and the averaged contaminant concentration did not change much, so it was not likely to change the risk level for the visitor or the caretaker.
4. The contaminant concentration in the ward with displacement ventilation was not uniform. The closer to the contaminant source, the higher the contaminant concentration. Hence at locations close to the contaminant source, the risk of transmission is higher with the displacement ventilation system compared to the mixing ventilation system.
Sagnik Mazumdar, PhD Student Member ASHRAE
Yanggao Yin, PhD Student Member ASHRAE
Arash Guity Associate Member ASHRAE
Paul Marmion Member ASHRAE
Bob Gulick Member ASHRAE
Qingyan Chen, PhD Fellow ASHRAE
Sagnik Mazumdar was a doctoral student at Purdue University, West Lafayette, IN, when this study was performed and is currently a post-doctoral fellow at the University of Medicine and Dentistry of New Jersey, Newark, NJ. Yonggao Yin is a lecturer at Southeast University, Nanjing, China. Arash Guity is a mechanical engineer and Bob Gulick is a senior principal at Mazzetti Nash Lipsey Burch, San Francisco, CA. Paul Marmion is a senior principal at Stantec, Vancouver, BC, Canada. Qingyan Chen is a Changjiang professor at Tianjin University, Tianjin, China, and a professor at Purdue University.
Received December 16, 2009; accepted June 16, 2010
Ansys. 2003. FLUENT user's guide, version 6.2. Ann Arbor, MI: Ansys, Inc.
Beggs, C.B., K.G. Kerr, C.J. Noakes, E.J. Hathway, and P.A. Sleigh. 2008. The ventilation of multiple-bed hospital wards: Review and analysis. American Journal of Infection Control 36(4):250-59.
Bjorn, E., M. Mattsson, M. Sandberg, and P.V. Nielsen. 1997. Displacement ventilation--Effects of movement and exhalation. Proceedings of Healthy Buildings, 5th International Conference on Healthy Buildings, Washington, D.C. pp. 163-68.
Bjorn, E., and P.V. Nielsen. 2002. Dispersal of exhaled air and personal exposure in displacement ventilated rooms. Indoor Air 12:147-64.
Brohus, H., K.D. Balling, and D. Jeppesen. 2006. Influence of movements on contaminant transport in an operating room. Indoor Air 16:356-72.
Gupta, J.K., C.-H. Lin, and Q. Chen. 2009. Flow dynamics and characterization of a cough. Indoor Air 19:517-25.
Gupta, J.K., C.-H. Lin, and Q. Chen. 2010. Characterizing exhaled airflow from breathing and talking. Indoor Air 20:31-9.
Li, Y., G.M. Leung, J.W. Tang, X. Yang, C.Y.H. Chao, J.Z. Lin, J.W. Lu, P.V. Nielsen, J. Niu, H. Qian, A.C. Sleigh, H.J.J. Su, J. Sundell, T.W. Wong, and P.L. Yuen. 2007. Role of ventilation in airborne transmission of infectious agents in the built environment--A multidisciplinary systematic review. Indoor Air 17(1):2-18.
Mazumdar, S., and Q. Chen. 2007. Impact of moving bodies on airflow and contaminant transport inside aircraft cabins. Proceedings of the 10th International Conference on Air Distribution in Rooms, Roomvent 2007, Helsinki, Finland, pp. 165.
Mazumdar, S. 2009. Transmission of airborne contaminants in airliner cabins. PhD Thesis, Purdue University, West Lafayette, IN.
Poussou, S.B. 2008. Experimental investigation of airborne contaminant transport by a human wake moving in a ventilated aircraft cabin. PhD Thesis, Purdue University, West Lafayette, IN.
Qian, H., Y. Li, P.V. Nielsen, C.E. Hyldgaard, T.W. Wong, A.T.Y. Chwang. 2006. Dispersion of exhaled droplet nuclei in a two-bed hospital ward with three different ventilation systems. Indoor Air 16:111-28.
Saravia, S.A., P.C. Raynor, and A.J. Streifel. 2007. A performance assessment of airborne infection isolation rooms. American Journal of Infection Control 35(5):324-31.
Thatcher, T.L., D.J. Wilson, E.E. Wood, M.J. Craig, and R.G. Sextro. 2004. Pollutant dispersion in a large indoor space: Part 1--Scaled experiments using a water-filled model with occupants and furniture. Indoor Air 14(4):258-71.
Yin, Y., W. Xu, J.K. Gupta, A. Guity, P. Marmion, A. Manning, R.W. Gulick, X. Zhang, and Q. Chen. 2009. Experimental study on displacement and mixing ventilation systems for a patient ward. HVAC&R Research 15(6):1175-91.
Zhai, Z., Q. Chen, P.W. Scanlon. 2002. Design of ventilation system for an indoor auto racing complex. ASHRAE Transactions 108(1):989-98.
Zhang, Z., W. Zhang, Z. Zhai, and Q. Chen. 2007. Evaluation of various turbulence models in predicting airflow and turbulence in enclosed environments by CFD--Part 2: Comparison with experimental data from literature. HVAC&R Research 13(6): 871-86.
Zhang, Z., X. Chen, S. Mazumdar, T. Zhang, and Q. Chen. 2009. Experimental and numerical investigation of airflow and contaminant transport in an airliner cabin mockup. Building and Environment 44(1):85-94.
|Printer friendly Cite/link Email Feedback|
|Author:||Mazumdar, Sagnik; Yin, Yanggao; Guity, Arash; Marmion, Paul; Gulick, Bob; Chen, Qingyan|
|Publication:||HVAC & R Research|
|Date:||Sep 1, 2010|
|Previous Article:||Editorial: expanding BIM to meet the grand challenges in buildings--what is needed?|
|Next Article:||Design of model-based optimizing control scheme for an air-conditioning system.|