Displacement Ventilation As a Viable Air Solution for Hospital Patient Rooms

DISPLACEMENT VENTILATION AS A VIABLE AIR SOLUTION FOR HOSPITAL PATIENT ROOMS

Julian Rimmer, P.Eng. Brad Tully, P.Eng. Alf Dyck, P.Eng. Mike Buck Price Industries Price Industries Price Industries University of Minnesota Senior Product Manager, General Manager, Vice President, Engineering Sustainable Technologies Price Mechanical West

Corresponding email: [email protected]

SUMMARY In few environments is the indoor air quality as important as that in hospital patient rooms. The air system must provide protection for the patient, caregiver as well as any visitors that may be in the room. Current codes and standards have been developed with this in mind, often by prescribing an air change rate, thereby ensuring a minimum amount of air movement and purge rate. These have all been put in place on the assumption of a mixing air supply system. Displacement ventilation offers a means of improving the air quality in these spaces, with the indication of lower air change rates. This in turn saves energy and reduces operating costs for the healthcare system. This paper presents the results of a particle study that quantifies the improvement of the air quality in a displacement ventilation system over that of a mixing system in a patient room. The exposure potential of these systems was evaluated along with maximum concentration and decay rate following a sneeze event. The displacement ventilation system showed significant improvement over the mixing ventilation case, even at lower air change rates.

INTRODUCTION In no industry is indoor environmental quality (IEQ) more important that in that of healthcare, an industry also known to be one of the most energy intensive. For years, codes and design best practices have ensured that these facilities provide the best protection against infection for the patient, healthcare providers and visitors. Lately, these design goals have been coupled with an effort to also create an environment that promotes recovery. To this end, patient rooms and waiting rooms, along with most other areas in a hospital, have prescriptive guidelines which outline the ventilation and air change rates in an effort to maintain the IEQ at an acceptable level. The typical solution for patient and waiting rooms today is ceiling based air distribution where filtered air is injected at the ceiling level at high velocity. This air then mixes and dilutes the room air at a rate set by the American Society for Heating, Refrigeration and Air Conditioning Engineers (ASHRAE) [1], the American Institute of Architects (AIA) [2] or by local codes in order to fully change the air in these rooms several times per hour. Recently, there has been growing interest in the use of alternative air distribution strategies in order to maximize IEQ while minimizing energy use in several non-critical areas in hospitals. One such strategy is displacement ventilation. Displacement ventilation has long been demonstrated to provide improved performance in all mechanically related areas of IEQ, namely indoor air quality (IAQ), acoustics and thermal comfort.

Background Displacement ventilation (DV) in practice introduces cool air directly into the occupied zone of a room at low velocity. The temperature of the supply air is slightly lower and therefore slightly more dense than the room air, allowing it to fall to the floor. The velocity of the supply air is low enough to minimize entrainment of, and therefore mixing with, the room air. This low velocity also ensures that there is minimal air movement in the space allowing the formation of thermal plumes. These plumes form around heat sources where the surrounding air is warmed and becomes buoyant, rising and being replaced with fresh air from below. When these heat sources are people, it is their thermal plume that delivers the fresh air directly to the breathing zone, mixing little with the surrounding room air. The resulting room airflow pattern, with cool fresh air initially at the floor, rising and warming in thermal plumes picking up pollutants along the way and collecting at a high level where it is exhausted or returned, is what makes displacement ventilation such an effective air distribution system. With the warm air trapped high in the room and with minimal air movement, the room airflow pattern is essentially one dimensional, from floor to ceiling. This characteristic allows for a reduction in the room design cooling capacity of the supply air [3].

Published 2010 Page 1 of 7

BACKGROUND

Displacement ventilation (DV) in practice introduces cool air directly into the occupied zone of a room at low velocity. The temperature of the supply air is slightly lower and therefore slightly more dense than the room air, allowing it to fall to the floor. The velocity of the supply air is low enough to minimize entrainment of, and therefore mixing with, the room air. This low velocity also ensures that there is minimal air movement in the space allowing the formation of thermal plumes. These plumes form around heat sources where the surrounding air is warmed and becomes buoyant, rising and being replaced with fresh air from below. When these heat sources are people, it is their thermal plume that delivers the fresh air directly to the breathing zone, mixing little with the surrounding room air.

The resulting room airflow pattern, with cool fresh air initially at the floor, rising and warming in thermal plumes picking up pollutants along the way and collecting at a high level where it is exhausted or returned, is what makes displacement ventilation such an effective air distribution system. With the warm air trapped high in the room and with minimal air movement, the room airflow pattern is essentially one dimensional, from floor to ceiling. This characteristic allows for a reduction in the room design cooling capacity of the supply air [3].

METHODS

This study intended to determine whether a displacement system would provide an adequate level of IEQ for a patient room, particularly at lower air change rates. To this end, both thermal comfort and particle distribution studies were carried out in order to evaluate the relative performance between the DV system and a mixing (MR) system from METHODSan occupant perception standpoint.. For this study, the MR airflow rate was set to match that required by ASHRAE [1] to six air changes per hour (ACH). The DV airflow rate This studywas intended set toto determine four ACH whether which a displacement is equivalent system would to six provide ACH an inadequate the occupied level of IEQ forzone a patient for patientroom, particularly at lower air changerooms rates. with To thisa 3m end, ceiling both thermal height comfort. The and particle distribution study wasstudies carried were carried out out on in the order standard to evaluate pa thetient relative performanceroom between shown the DV in system Figure and 1 a. mixing These (MR) rooms system frominclude an occupant a patient, perception located standpoint.. on the For bed, this study, a caregiver the MR airflow rate was set to match that required by ASHRAE [1] to six air changes per hour (ACH). The DV airflow rate was set to four ACH which is equivalent and visitor in the vicinity of the bed (all shown in pink), a television mounted to the foot to six ACH in the occupied zone for patient rooms with a 3m ceiling height. The particle study was carried out on the standard patient room shown in Figurewall 1.and These a window rooms include to simulatea patient lying the on perimeterthe bed, a caregiver loads and for visitor various in the vicinityclimate of thes. bed (all shown in pink), a television mounted to the foot wall and a window to simulate the perimeter loads for various climates. Two displacement diffusers were installed in the corners of the room (shown in blue) but Two displacementonly one diffusers of these were installed was active in the corners for the of thetest. room The (shown return, in blue), red, but onlywas one located of these atwas the active ceiling for the level,test. The return, red, was locatedabove at the the ceilingdoor. level, above the door.

a) b) Figure 1 Plan (a) and isometricFigure 1- Plan (a)(b) and viewsisometric (b)of views the of patient the patient roomroom under under study study

Particle Distribution The patient room was set up in the mockup room at an air distribution lab located in Winnipeg, Canada. A procedure for evaluating the effect of displacement ventilation on the exposure of occupants to particles was developed in conjunction with two international engineering firms. It was determined the standardized particle study, developed by the team, would be conducted to evaluate the indoor air quality of the room with a mixing system and a DV system. The particles were generated and injected into the room with a common nebulizer and the values were measured with a particle counter. The location of both instruments was varied from test to test, according to Table 1, for the test conditions outlined in Table 2.

Table 1. Particle Tests Configuration Location of Nebulizer Location of Particle Counter 1 Patient Bed Caregiver 2 Caregiver Patient Bed

Table 2. Details of test patient rooms Room MR DV Air Distribution System Mixing DV Room Air Temperature 23.3°C (74°F) 25°C (77°F) Supply Air Volume (ACH) 137L/s (6) 85L/s (4) Supply Air Temperature 18.3°C (65°F) 18.3°C (65°F) Diffuser Location Ceiling Low Level, Across from Bed Room Occupant Load 250 W 250 W Exhaust Location Room and Toilet Room and Toilet Exhaust Temperature 24.4°C (76°F) 26.7°C (80°F)

Published 2010 Page 2 of 7 For both the MR and DV layouts, the supply air was filtered with a HEPA 99.97% efficient filter to ensure that it contained a minimal count of particles. The nebulizer was selected to provide a range of particle sizes in line with sizes of common pathogens. Table 3. Characteristics of the compressed air nebulizer Output Concentra- Droplet Size Distribution Operating Pressure Flow Rate (1/min) tion MMD (μm) GSD Nebulizer 10 11 16 4.2 1.8

Where MMD is the mass mean diameter and GSD is the geometric standard deviation of the droplet size distribution. At the start of each test, fan filter units equipped with HEPA filters were used to clear the room of particles. These were run for 15 minutes (5 air changes). Once shut off, the room was allowed to stabilize for 5 minutes. The nebulizer pump was then run for 30 seconds. The particle counter then monitors the local concentration of particles as they decay over time. The counter was allowed to run for 45 to 60 minutes after the pump was shut off to give a good description of the concentration decay.

RESULTS For the analysis, the data was corrected to a background concentration of 0 particles as each run had a slightly different, though low, background concentration when the tests were initiated. To accomplish this, the average background concentration was taken for each run during the five minute stabilization period. This average value was then subtracted from the measured particle count during the test. The magnitude of this correction varied from 0.4% of the test average count for the 0.5μm particles to 1.5% of the test average for the 2μm particles. Figure 2 shows the data for the study where the particle source was at the patient location and the particle counter at the caregiver’s location. Time 0 is at the moment when the injection of particles start, with the duration of the injection shown in grey. subtracted from the measured particle count during the test. The magnitude of this Incorrecti the figure,on the variedMR cases from(shown in0.4% orange of and the red) indicatetest average a more gradual count increase for andthe decrease 0.5μm in particleparticles counts tothan 1.5% the two of displacement cases (shown in blue). This is not unexpected because lateral air movement in a mixing system can be significant, causing thethe injected test particlesaverage to move for in the all three 2μm dimensions. particles. In contrast, Figure displacement 2 shows ventilation the systemsdata for have the a bulk study air motion where that is largelythe one dimensional,particle sourcefrom floor wasto ceiling. at theThis wouldpatient cause locationthe particle counterand thelocated particle directly above counter the source at tothe have caregiver’s a near immediate response tolocation. the injection ofTime particles, 0 asis isat noted the from moment Figure 2. Tablewhen 4 and the Table injection 5 show the decayof particl of the measuredes start, particle with count the through duration the first tenof minutes the injection of the test. shown in grey.

Figure 2. ConcentrationFigure 2 - Concentration ofof 2 and2 5and micron 5 particles micron vs. time particles for (a) particle source vs. attime the patient for and (a) counter particle at the caregiver source and at the patient and counter at the caregiver(b) particle source and at (b)the caregiver particle and counter source at the patient at the caregiver and counter at Resultsthe patient for the mixing and DV systems are given in Table 4 for the case of the particle source located at the patient and particles measured at the caregiver. As indicated in the table, the measured particle count in the DV system decays in number of particles 93-94% in the first 10In minutes, the figure, whereas the mixing MR system cases, caused shown a decay in of 11-80%.orange What and is notablered, indicate is that the quicker a more decay gradual is realized evenincrease at a lower and air change rage, 4ACH for the DV system vs. 6ACH for the mixing system. decrease in particle counts than the two displacement cases, shown in blue. This is not Results for the mixing and displacement ventilation systems are given in Table 5 for the case of the particle source located at the caregiver andunexpected particles measured because at the patient. lateral In this air case, movement the decay for thein DVa mixingsystem is 66-85% system in the can first be 10 significant,minutes, compared causing to 36- 87% for thethe mixing injecte system.d Again,particles these results to move are for ain lower all air three change dimensions. rate for the displacement In contrast, system. displacement ventilation systems have a bulk air motion that is largely one dimensional, from floor to ceiling. This would cause the particle counter located directly above the source to have a near Published 2010 Page 3 of 7 immediate response to the injection of particles, as is noted from Figure 2. Table 4 and Table 5 show the decay of the measured particle count through the first ten minutes of the test.

Table 4. Particle count decay, source located at the patient and measured at the caregiver 0.3μm 0.5μm 1μm 2μm 5μm 10μm Room ACH Particles Particles Particles Particles Particles Particles Peak Count 74832 15093 2366 643 18 2 Count after 15852 3037 622 229 16 0 MR 6 10 min Difference 58980 12056 1744 414 2 2 % Reduction 79% 80% 74% 64% 11% 99% Peak Count 132147 30720 4619 1139 22 3 Count after 9801 1827 269 68 2 0 10 min DV 4 Count 122346 28893 4350 1071 20 3 Reduction % Reduction 93% 94% 94% 94% 93% 100%

Results for the mixing and DV systems are given in Table 4 for the case of the particle source located at the patient and particles measured at the caregiver. As indicated in the table, the measured particle count in the DV system decays in number of particles 93-94% in the first 10 minutes, whereas the mixing system caused a decay of 11-80%. What is notable is that the quicker decay is realized even at a lower air change rage, 4ACH for the DV system vs. 6ACH for the mixing system.

Results for the mixing and displacement ventilation systems are given in Table 5 for the case of the particle source located at the caregiver and particles measured at the patient. In this case, the decay for the DV system is 66-85% in the first 10 minutes, compared to 36- 87% for the mixing system. Again, these results are for a lower air change rate for the displacement system Table 4. Particle count decay, source located at the patient and measured at the caregiver 0.3μm 0.5μm 10μm Room ACH Table 5. Particle count decay, source located1μm Particles at the 2μmcaregiver Particles and5μm measured Particles at the patient Particles Particles Particles 0.3μm 0.5μm 1μm 2μm 5μm 10μm Room ACH Peak Count 74832 15093Particles Particles2366 Particles643 Particles18 Particles2 Particles Count after 10 min 15852 3037 622 229 16 0 MR 6 Peak Count 67109 13090 951 268 10 0 Difference Count58980 after 12056 1744 414 2 2 10695 1710 332 103 7 0 MR% Reduction6 1079% min 80% 74% 64% 11% 99% Peak Count Difference132147 3072056414 461911380 1139619 16522 4 3 0 Count after 10 min % Reduction9801 182784% 26987% 65%68 62%2 36% 0 100% DV 4 Count Reduction Peak122346 Count 2889315677 43503026 16531071 45920 11 3 3 % Reduction Count93% after 94% 94% 94% 93% 100% 2997 538 252 83 4 0 10 min DV 4 Table 5. Particle count decay, source locatedCount at the caregiver and measured at the patient 12680 2488 1401 376 7 3 Reduction0.3μm 0.5μm 10μm Room ACH 1μm Particles 2μm Particles 5μm Particles %Particles Reduction Particles81% 82% 85% 82% 66%Particles 100% Peak Count 67109 13090 951 268 10 0 CountIt is interestingafter 10 min to note10695 that though1710 the values332 for particle103 count in Table7 5 shows0 a slightly MR 6 lowerDifference reduction as 56414a percentage11380 of the peak619 value than 165that in Table 44, the peak 0itself is approximately% Reduction an order84% of magnitude87% lower.65% It is unlikely62% that this36% is solely caused100% by a fluctuationPeak Count in particle15677 generation3026 between tests1653 but is more459 likely contributabl11 e to3 the Countmanner after at 10 which min DV2997 systems move538 the air 252in the room. 83 4 0 DV 4 Count Reduction 12680 2488 1401 376 7 3 From the tables, the maximum number of particles observed between the tests for the % Reduction 81% 82% 85% 82% 66% 00% mixing varies less than 10%, suggesting that the tests and particle distribution are repeatable. As previously discussed, the air pattern generated with the DV system is It is interesting to notegenerally that though theone values-dimensional, for particle count from in Table floor 5 shows to ceiling. a slightly lowerThis reduction would as account a percentage for of the the peakvariance value than that in Table 4, the peak itself is approximately an order of magnitude lower. It is unlikely that this is solely caused by a fluctuation in particle generation betweenfound tests, in thebut ismagnitude more likely attributable of the particleto the manner spikes in which for DV the systems DV movecase the with air in the the lowerroom. value being where the source is downstream (above) from the measurement location. By contrast, the From the tables, the maximumcharacteristics number of ofparticles mixing observed systems between also the tests explain for the themixing very varies little less thandifference 10%, suggesting observed that the in tests the and particle distribution are repeatable. As previously discussed, the air pattern generated with the DV system is generally one-dimensional, peaks because the air pattern tends to distribute the particles throughout the room. from floor to ceiling. This would account for the variance found in the magnitude of the particle spikes for the DV case with the lower value being where the source is downstream (above) from the measurement location. By contrast, the characteristics of mixing systems also explain the very little differenceWhen the observed decay in thedata peaks is becausenormalized the air patternfor each tends particle to distribute size the particlesto the magnitudethroughout the ofroom. the respective count peaks, or When the decay data is normalized for each particle size to the magnitude of the respective count peaks, or

Counti Counti, N  , (1) Counti, Max a trend is clearly visible. Figure 3 shows the 0.3 - 2μm normalized decay curves which are nearly identical for the displacement case, and are largely similar for athe trend mixing is case. clearly The curves visible. for the Fi5 andgure 10μm 3 showswere omitted the due0.3 to - their2μm relatively normalized low magnitudes decay and curves high noise which are levels. Fundamentally,nearly we would identical expect to see for very the good displacement agreement between case, these and curves are becauselargely the similar range of particlefor the sizes mixing which arecase. The presented are also thosecurves that exhibit for the aerosol 5 and behaviour 10μm and were do not omittedsettle. The factdue that to they their remain relatively in suspension low means magnitudes that they can and high be used to track the motionnoise of levels. air in the roomFundamentally, and that the time we required would for the expect particles to to seebe noticed very (peak)good by agreement the caregiver shouldbetween be these indicative of the effectivenesscurves of because the air distribution the range system of at particle delivering sizesfresh air which to the occupants. are presented are also those that exhibit Figure 3 shows that theaerosol time to peakbehaviour for the DV and and mixingdo not systems settle. are approximatelyThe fact that two theyand five remain minutes, in respectively. suspension The response means time that indicates that the DV thesystemy can is 2.5 be times used more to effective track atthe delivering motion air toof the air occupants in the roomthan the and mixing that systems. the timeThis value required is not outside for the of the range of ventilationparticles effectiveness to be (VE) noticed values typically (peak) associated by the withcaregiver DV systems. should The data be is indicativeonly presented of for the the caseeffectiveness where of the source is located theat the air patient distribution because in VE system studies thisat deliveringis akin to the step-up fresh procedureair to the where occupants. the label is added to the airstream upstream of the measurement location.

Published 2010 Page 4 of 7

Figure 3. Normalized particle counts for the DV and mixing systems for the case with patient source, measured at the caregiver.

Figure 3 shows that the time to peak for the DV and mixing systems are approximately two and five minutes, respectively. The response time indicates that the DV system is 2.5 times more effective at delivering air to the occupants than the mixing systems. This value is not outside of the range of ventilation effectiveness (VE) values typically associated with DV systems. The data is only presented for the case where the source is located at the patient because in VE studies, this is akin to the step-up procedure where the label is added to the airstream upstream of the measurement location.

The stability of the displacement ventilation curves is an indication of the low turbulence experienced with a DV system. The count decay curves for all of the particles that display

aerosolFigure behaviour 3. Normalized are nearly particle identical, counts whereasfor the DV they and are mixing not for systems the mixing for the case, case with with the Figure 3 - Normalized particle counts for the DV and mixing systems for the case with morepatient massive source, particles measured decaying at the atcaregiverpatient a slower source,. measured rate. at the caregiver.

The stabilityExposure of the displacement Potential ventilation curves is an indication of the low turbulence experienced with a DV system. The count decay curves forFi allgure of the 3 particlesshows that that display the timeaerosol to behaviour peak for are nearlythe DV identical, and whereasmixing they systems are not for are the approximately mixing case, with the more In infection control it is the exposure potential, or number of particles in a period of time, massive twoparticles and decaying five minutes, at a slower respectively.rate. The response time indicates that the DV system is 2.5 thattimes influences more effective whether at cross delivering-infection air tocan the take occupants place. Tthanhe area the mixingunder the systems. curve s Thisfrom ExposureFigurevalue Potential 2 is defines not outside this exposure of the range in particle of ventilation-minutes. effectiveness It is expected (VE) that values the particletypically count vs. In infectiontimeassociated controlcurve it isis withthesmooth exposure DV andsystems. potential, therefore or numberThe Simpsons’s data of particles is only in Ruleapresented period was of time, used for that the to influences caseevaluate where whether the the areacross-infection source under is can take place.the Thelocated curvearea unders at in the Figure curves patient from4. becauseWhen Figure 2 evaluateddefines in VE this studies, exposurewith athis inparticle particle-minutes. is akin count, to the It p,is step expectedthough-up thatprocedure time the andparticle nwhere count= 15 vs. the time curve is smoothintervalslabel and therefore is of added 2 minutesSimpson’s to the Rule eachairstream was, the used exposure upstreamto evaluate the potentialof area the under measurement can the curvesbe determined in Figurelocation. 4. When from evaluated: with a particle count, p, though time and n = 15 intervals of 2 minutes each, the exposure potential can be determined from: The stability of the 1displacement ventilation curves is an indication of the low turbulence EX  p  4p  2p  4p  2p  ...  4p  p  (3) experienced with a DV3 0system.1 The2 count 3decay 4curves for29 all of30 the particles that display Using thisaerosol method, behaviourthe exposure potentialare nearly in particle-minutes identical, whereasfor the two systemsthey are is shown not for in Table the 6mixing and Table case, 7. with the Usingmore this massive method, particles the exposure decaying potential at a slower in particle rate. -minutes for the two systems is shown in Table 6 and Table 7. Exposure Potential In infection control it is the exposure potential, or number of particles in a period of time, that influences whether cross-infection can take place. The area under the curves from Figure 2 defines this exposure in particle-minutes. It is expected that the particle count vs. time curve is smooth and therefore Simpsons’s Rule was used to evaluate the area under the curves in Figure 4. When evaluated with a particle count, p, though time and n = 15 intervals of 2 minutes each, the exposure potential can be determined from:

1 EX  p  4p  2p  4p  2p  ...  4p  p  (3) 3 0 1 2 3 4 29 30

Using this method, the exposure potential in particle-minutes for the two systems is shown in Table 6 and Table 7.

Figure 4. Area under the concentration curve of the 2 micron particles vs. time with Figure 4 - Area under the concentration curve of the 2 micron particles vs. time with displacement ventilation displacement ventilation

Table 6. Exposure potential to caregiver, source at patient Published 2010 Page 5 of 7 0.3μm 0.5μm 1μm 2μm 5μm ACH Particles Particles Particles Particles Particles Mixing 6 537398 107047 19554 6663 309 DV 4 548964 116345 17710 4530 105 Difference -11566 -9298 1844 2133 204 % Difference -2% -9% 9% 32% 66%

As shown in Table 6, the difference between the exposure potential for the DV case at four ACH and the mixing case at six ACH, respectively, is minor for the 0.3, 0.5 and 1.0µm particle sizes. The difference is more pronounced for the larger particle sizes with the DV system showing significantly less exposure to the caregiver than the mixing system, even with the higher maximum particle count observed in Table 4.

Table 7. Exposure potential to patient, source at caregiver 0.3μm 0.5μm 1μm 2μm 5μm ACH Particles Particles Particles Particles Particles Mixing 6 384873 65904 11776 3612 164 DV 4 92092 16866 9174 2803 118 Difference 292781 49039 2602 809 46 % Difference 76% 74% 22% 22% 28%

Table 7 presents the data for the exposure potential to the patient with the source at the caregiver. In this case, there is a large reduction in the exposure of the patient to the smaller particles. Again, due to the fact that these exhibit aerosol behaviour and the single-pass air pattern generated by the DV system, one would expect that the particles whose movement is dominated by the air transport would show this characteristic. This trend continues to the more massive particle sizes which show a significant reduction in the exposure of the patient to the particles injected at the caregiver location.

Table 6. Exposure potential to caregiver, source at patient ACH 0.3μm Particles 0.5μm Particles 1μm Particles 2μm Particles 5μm Particles 6 537398 107047 19554 6663 309 Mixing DV 4 548964 116345 17710 4530 105 Difference % -11566 -9298 1844 2133 204 Difference -2% -9% 9% 32% 66%

As shown in Table 6, the difference between the exposure potential for the DV case at four ACH and the mixing case at six ACH, respectively, is minor for the 0.3, 0.5 and 1.0μm particle sizes. The difference is more pronounced for the larger particle sizes with the DV system showing significantly less exposure to the caregiver than the mixing system, even with the higher maximum particle count observed in Table 4.

Table 7. Exposure potential to patient, source at caregiver ACH 0.3μm Particles 0.5μm Particles 1μm Particles 2μm Particles 5μm Particles 6 384873 65904 11776 3612 164 Mixing DV 4 92092 16866 9174 2803 118 Difference % 292781 49039 2602 809 46 Difference 76% 74% 22% 22% 28%

Conclusions This study has demonstrated that the air quality as it relates to measured counts of particles exchanged between room occupants is equivalent or better with a DV system when compared to a mixing system. The fact that these results are achieved at 1/3 less airflow is an indication that the DV system improves the effectiveness of the air delivery system’s ability to protect the occupants. The data presented in this study provide some experimental support for the DV system as an effective solution for patient room supply. That the DV system demonstrated this level of performance at a lower air change rate suggests that the healthcare provider can use DV technology to provide superior air quality while reducing the amount of energy required to operate the system. If the DV system is implemented in similar environments within the facility, there is also the possibility of larger scale reductions in air volume and the other mechanical advantages that stem from this such as reduced ductwork and smaller mechanical equipment.

DISCUSSION It is expected that the displacement ventilation system is an appropriate choice for patient rooms due to the improvement in indoor air quality and opportunity for air volume and energy savings. It is also believed that this result should be applicable to more non-critical spaces in a healthcare facility including waiting rooms, exam rooms, MRI rooms, cafeterias, Atria, respite rooms, family rooms as well as nurse station and corridors. This result highlights a significant opportunity to save energy in hospitals by adopting a more efficient DV system for a significant percentage of the building footprint.

ACKNOWLEDGEMENT The authors would like to thank Andy Streifel from the University of Minnesota, Bob Gulick from Mazzetti and Associates as well as Paul Marmion from Stantec Engineering for their assistance.

REFERENCES

1. ASHRAE. 2008. ANSI/ASHRAE/ASHE Standard 170-2008, Ventilation of Health Care Facilities, Atlanta: American Society of Heating, Refrigerating, and Airconditioning Engineers, Inc.

2. AIA. 2006. Guidelines for Design and Construction of Hospital and Health Care Facilities, Dallas: Facilities Guidelines Institute

Published 2010 Page 6 of 7 3. ASHRAE. 1999. RP-949 -- Performance Evaluation and Development of Design Guidelines for Displacement Ventilation , Atlanta: American Society of Heating, Refrigerating, and Airconditioning Engineers, Inc.

4. Hinds. 1999. Aerosol Technology Properties, Behavior, and Measurement of Airborne Particles: William C. Hinds. Wiley, New York (1999). ISBN 0-471-19410-7

Published 2010 Page 7 of 7