Air leakage analysis of special ventilation hospital rooms.

INTRODUCTION

Methods to control airborne infectious diseases in hospitals containing afflicted patients are vital to the safety of health-care workers, other patients, and visitors (Rice et al. 2001; Streifel 2003, 2006). Airborne infection isolation (AII) rooms are designed to isolate patients with a suspected airborne infectious disease and rely on special ventilation parameters to contain and filter infectious particles. Protective environment (PE) rooms are designed to protect immune-compromised patients from infectious particles.

Three environmental controls are used for special ventilation (SPV) rooms (AII and PE), including pressure management, room air changes for dilution ventilation, and filtration to remove infectious particles (Saravia et al. 2007). Pressure management refers to the use of differential pressure (patient room compared to adjacent spaces) to promote "clean to dirty" airflow. "Negative" pressure is used for AII rooms and is achieved by adjusting the exhaust/return and supply airflow rates such that the mechanically exhausted airflow is greater than the mechanically supplied air. Positive pressure is required for PE rooms and is achieved by mechanically supplying more air than is exhausted.

In practice, the relative pressure between the patient room and the corridor is most commonly used to indicate acceptable pressure control. In reality, however, the patient room must be pressurized relative to every adjacent space in order to assure that there is not airflow from any of the adjacent spaces to or from the patient room. Poorly sealed rooms require significantly higher flow differentials to establish a substantial differential pressure (Streifel 2005). In addition, greater flow differentials result in higher monetary costs of ventilation for leakier rooms. The requirement to be consistently pressurized over all six sides underscores the need for tightly sealed rooms.

The American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. (ASHRAE) defines air leakage of a building as follows: "The air leakage area of a building is ... the area of an orifice (with an assumed value of [C.sub.D) that would produce the same amount of leakage as the building envelope at the reference pressure" (ASHRAE 2005). The results in this study are presented as the equivalent leakage area, which is defined at a differential pressure of 10 Pa and a discharge coefficient of 0.611. The concept of air leakage gives designers the ability to design pressurized buildings or spaces within buildings (e.g., cleanrooms, SPV hospital rooms) to operate at a specific differential pressure under specific airflow offsets (there is a quantitative difference between supply and exhaust/return airflow rates). However, air leakage is not well defined for SPV rooms, and designers overwhelmingly fail to account for air leakage in their designs (Hewett and Hermans 2006).

While SPV hospital room air leakage has been a topic of interest since the 1980s (Murray et al. 1988), the concept of air leakage is not well defined for the design and construction of these rooms. Organizations including the American Institute of Architects (AIA) and Centers for Disease Control (CDC) attempt to provide guidance on leakage specifications (AIA 2006; CDC 2005). The AIA suggests that SPV rooms should be sealed tightly and the CDC suggests that the leakage area of SPV rooms should be no more than 72 in. (2). While these both seem reasonable initially, the first suggestion complete lacks measurability and the second suggestion fails to specify at what pressure and over what surface area the leakage occurs. These suggestions also fail to include a discharge coefficient assumption, and they do not specify whether the leakage area is ELA ([DELTA][p.sub.r] = 4 Pa, [C.sub.D] = 1), where ELA is the effective leakage area, or EqLA ([DELTA][p.sub.r] = 10 Pa, [C.sub.D] = 0.611). In addition, these organizations fail to describe the means and methods of measuring the leakage area and sealing a leaky room.

Another example that conveys the lack of understanding of leakage area is shown in Guidelines for Environmental Infection Control in Health-Care Facilities from the Healthcare Infection Control Practices Advisory Committee of the CDC (HICPAC 2003). First of all, it suggests that an airflow offset (125 cfm recommended) correlates with the air changes provided (12 ach recommended), which is completely untrue. Also, the only suggestion for achieving the desired pressure differential is that the airflow offset may need to be increased. It is true that the offset will need adjustment, but without a quantitative estimate of the leakage area it is impossible to estimate the required airflow offset. Although airflow offset and pressure differential are related to equivalent leakage area, these suggestions underscore the general physical misunderstanding of leakage area.

In another study, empirical relationships were found between leakage area, airflow offset, and differential pressure (Hayden et al. 1997). This was done by constructing an extremely tight space, creating a controlled orifice of a given size, and determining the average differential pressure over all surfaces of the space. Again, it is true that leakage area, airflow offset, and differential pressure are related; however, this empirical relationship is difficult to apply to all rooms.

The air leakage of the room directly affects the pressure differential that can be established and the airflow offset required. Physically, there is a positive relationship between differential pressure and differential flow. Current standards and suggestions do not account for this and fail to describe an acceptable leakage area, methods to determine leakage area, and methods to locate leakage areas and properly seal them.

MATERIALS AND METHODS

Ten SPV hospital rooms designed and built in the mid-1980s were tested for this study. During construction, special attention was paid to sealing the PE rooms while the AII rooms and intensive care unit (ICU) rooms did not receive that detail (Rice et al. 2001). A summary of the hospital rooms tested is provided in Table 1.

Table 1.Room, Ceiling, and Door Types for SPV Rooms Tested

Room  Room Type  Ceiling Type            Door Type

 A      PE            Solid             Single door
 B      PE            Solid             Single door
 C     ICU          Suspended           Double door
 D     ICU          Suspended      Sliding three-panel door
 E     ICU          Suspended      Sliding three-panel door
 F     All       Perforated panel       Single door
G1     All       Perforated panel       Single door
G2     All       Perforated panel       Single door
H1     All       Perforated panel       Single door
H2     All          Suspended           Single door

* The solid ceiling is a metal pan ceiling that was painted and
completely sealed with caulk. The suspended ceiling is a special
suspended ceiling with cleanable surfaces and gasketed T-bar. The
perforated panel ceiling is a suspended ceiling with perforated metal
panels with radiant heating coils installed above.

This study was completed using duct pressurization equipment and analysis software. A calibrated fan (1) was used to depressurize the room by forcing air into the exhaust duct. The pressure at the flow sensor of the fan was measured using a digital pressure and flow gauge (2) and converted to a flow rate. In addition, the digital pressure gauge was used to monitor the relative pressure of the room to the corridor.

The fan includes four calibrated air inlet settings that were used to vary the airflow. The fan pressure and differential room pressure were logged using data logging software (3) designed to communicate with the digital pressure and flow gauge. By alternating the fan inlet size and the fan speed, several data points were obtained at differential pressures. These data points were manually entered into the data analysis software. (4) This software performs a regression of the measured flow and pressure data to calculate c and n for Equation 1 (see "Theory" section). It then evaluates Equation 1 to determine the flow at the reference pressure (for EqLA, the reference pressure is 10 Pa). Then, Equation 2 is used to calculate the EqLA. Airtightness test results were calculated by the data analysis software using CGSB Standard 149.10-M86, Determination of Airtightness of Building Envelopes by the Fan Depressurisation Method.

The air leakage test method consisted of the following tasks:

1. Determine the surface area of the room.

2. Secure duct adapter to exhaust/return register (ensure that damper is not closed).

3. Seal all remaining supply and exhaust grilles.

4. Connect flexible duct to adapter and fan.

5. Connect pressure gauge to laptop and run tubing to fan housing and corridor.

6. Use TECLOG data logging software to obtain data points to be used in calculating leakage area.

7. Analyze data using TECTITE to determine room leakage.

THEORY

The attainable differential pressure is related to the leakage area and the airflow offset of the room. Empirically it has been found that the airflow offset needed to pressurize most enclosures to various differential pressures follows a power law of the form of Equation 1 (ASHRAE 2005). By performing a linear regression to the natural logs of pairs of Q and [DELTA]p, the coefficients C and n can be determined. Then the flow needed to create any given pressure differential can be determined.

Q = c[([DELTA]p).sup.n] (1)

where

c = flow coefficient, cfm/[(in.[H.sub.2]O).sup.n)]

n = pressure exponent, dimensionless

In order to calculate the leakage area of the room, [A.sub.L], for comparison purposes a reference pressure and a discharge coefficient must be chosen for the type of leak that is being used. For EqLA, the discharge coefficient is assumed to be 0.611 and the reference pressure is 10 Pa. These values ([C.sub.D] and [DELTA][p.sub.r]) along with the required airflow offset (Q) from Equation 1 and the standard air density can be substituted into Equation 2 (ASHRAE 2001) to determine the leakage area.

[A.sub.L] = [CQ.sub.r][square root of ([rho]/2[DELTA][p.sub.r]/[C.sub.D]] (2)

where

[A.sub.L] = equivalent or effective leakage area, in.(2)

[Q.sub.r] = predicted airflow rate at [DELTA][p.sub.r] (from curve fit to pressurization test data), cfm

[rho] = air density, [lb.sub.m]/[ft.sup.3]

[DELTA][p.sub.r] = reference pressure difference, in. [H.sub.2]O

[C.sub.D] = discharge coefficient

C = unit conversion factor = 0.186

RESULTS

The equivalent leakage areas presented in this study were normalized against the general room surface area and reported as in. (2) per 100 [ft.sup.2] of surface area. In order to represent this, the variable NEqLA (normalized EqLA) will be used and is defined as EqLA per 100 [ft.sup.2] of room surface area.

In this study, the equivalent leakage areas of eight SPV hospital rooms were determined. Two rooms were retested after remodeling work was completed. Rooms G2 and H2 represent the second round of testing on rooms G1 and H1. The main component of the remodeling was replacement of the perforated panel ceilings with suspended ceilings (with solid, heavier tiles and special gasket T-bar).

The results of the air leakage tests are shown in Table 2. The CDC and AIA recommend a pressure differential of at least 2.5 Pa (0.01 in. [H.sub.2]O) for these rooms (CDC 2005; AIA 2006). The offset required to pressurize or depressurize the rooms to 2.5 Pa was found using Equation 1 and is seen in Table 2. Figure 1 graphically shows the data presented in Table 2. Figure 2 shows the normalized leakage area of rooms G and H before and after the perforated panel ceilings.

[FIGURE 1 OMITTED]

Table 2. Normalized Leakage Area and Airflow Offset Required for Each
Room

Room  NEqLA (in. (2) per 100 [ft.sup.2] of surface area)  Offest *

  A                            5.6                           61
  B                            3.4                           43
  C                           19.6                          269
  D                           23.6                          235
  E                           23.7                          210
  D                           23.6                          235
  E                           23.7                          210
  F                           21.6                          237
 G1                           22.6                          248
 G2                           10.8                          144
 H1                           31.8                          314
 H2                           14.2                          169

* Calculated airflow offset (cfm) required for 2.5 pa.

From Tables 1 and 2 and Figure 2, the rooms of the study can be compared and evaluated upon the basic room design and the air leakage characteristics. The two rooms that were found to be the most airtight were the two PE rooms. This was expected, as these rooms received extra attention during construction in an effort to make them as airtight as possible. These two rooms were unique to the study as they were the only two tested that had solid ceilings.

[FIGURE 2 OMITTED]

Rooms C, D, and E are ICUs with the ability to convert to negative pressure and were found to have similar air leakage characteristics. These rooms are unique as they were recently remodeled and each had a new suspended ceiling. The suspended ceilings in these three rooms (and in rooms G2 and H2) were not of the type found in office buildings but rather consisted of specially designed materials. The tiles were surface cleanable, unlike standard tiles, and the T-bar was fitted with a gasket that allowed a better seal between the tile and the T-bar. It could be assumed that the gasket would significantly lower the expected leakage area of these rooms; however, it seems to be offset by the type of doors found in these rooms.

The door to room C is an oddly shaped two-panel hinged door consisting of a 17 in. wide section and a 45 in. wide section separated by a 3/8 in. gap. This gap, along with the undercut of the door, undoubtedly account for a significant portion of the air leakage of the room. Rooms D and E each have a single 45 in. door along with a sliding panel door approximately the length of an entire wall of the room. It is expected that this accounts for a large portion of the air leakage of these two rooms.

Rooms F, G1, and H1 are all of the same type and construction. Each is an AII room, and each has a single entrance door and a perforated panel suspended ceiling. Although it was not possible to test the room leakage that was due to the ceiling, it is expected that the ceiling contributed a large portion of the air leakage for these three rooms.

During the course of the study, rooms G and H underwent remodeling. The only portion of the remodeling that changed the physical nature of the room was the replacement of the perforated panel suspended ceiling with the new suspended ceiling described previously. These suspended ceilings were expected to be tighter than regular suspended ceilings found in office buildings and residential construction and significantly tighter than the perforated panel suspended ceilings original to the rooms. Figure 2 shows the air leakage of each room before and after remodeling. As shown, the decrease in leakage area was greater than 50% in each case.

Most of the sources of air leakage in a typical patient room are unintended sources including electrical outlets, multimedia connections, medical gas connections, lighting, the ceiling, plumbing connections, etc. An intended, or controllable, source of air leakage is the corridor door. When possible, the rooms of this study were tested with the door unsealed and again with the door sealed. Subtracting these two values, the door EqLA could be estimated. This was done for rooms A, B, C, F, G2, and H2 and was plotted against the corresponding physical area of the door undercut (see Figure 3). A moderate correlation ([R.sup.2] = 0.757) was found. The EqLA is actually larger than the physical area, and from this data the multiplier to convert from physical area to EqLA is 1.57 (when the trend line is forced to zero).

[FIGURE 3 OMITTED]

DISCUSSION

Research can be performed on the leakage areas of rooms and spaces, but without an understanding of the leakage of room components, designers lack the tools necessary to meet a standard for air leakage. The data from this study suggest that the ceiling and door are the two largest contributors to the leakage area of a room. The door is a controllable source because the undercut can be adjusted and a door sweep can be added. The sides and top of the door can be sealed more tightly by fitting the door frame with gaskets along the joints.

The data obtained in this study followed general expectations. The PE rooms were expected to have the lowest leakage area because of the solid ceilings and the special attention they received during initial construction. The ICU rooms were expected to have moderate leakage areas. All ICU rooms had special suspended ceilings, which likely decreased the leakage area. However, this was likely offset by the leaky doors as described in the Results section. The AII rooms with the perforated panel ceilings were expected to have moderate to high leakage areas. Upon replacement of the perforated panel ceilings with the special suspended ceilings, the leakage areas of the AII rooms decreased markedly as expected.

CONCLUSION

Currently, there is not a standardized method for determining leakage areas of rooms. In this study, the leakage areas of 8 SPV rooms were determined through a time-consuming process that isn't practical for projects of larger scope. There are currently other methods in development to determine room leakage areas in a speedier and less disruptive manner.

In addition, once the leakage area is determined, there is not a standard method for locating and sealing leakage areas. Aside from installing a solid ceiling during initial construction, a commonly used method is the application of a bead of caulk along any cracks or connections between two surfaces. Although this will likely help reduce leakage, it is difficult to note progress. A practical method used to locate leaks involves temporary large airflow offsets and a smoke pencil. In this method, a temporary large airflow offset can be created by covering the exhaust and return grilles with paper. This offset will over-pressurize the space and force air out through the leakage sites. If the offset isn't large enough, the door can be sealed with tape, which will force more air through the unknown leakage sites (rather than around the door). We have found this method useful and straightforward for carpenters.

From the data obtained in this study, some recommendations for further study and suggestions for leakage area standards can be made. A commonly used recommendation for airtight energy-efficient homes is a normalized EqLA of 2.5 in. (2) per 100 [ft.sup.2] of surface area (Lstiburek 2004). In this study, the most airtight rooms were found to have leakage areas on this order, but the poorly sealed rooms were found to fall far short. Based on the results, we propose a normalized equivalent leakage area standard for AII and PE rooms of 2.5 in. (2) per 100 [ft.sup.2] of surface area for unintended room leakage areas (not including door leakage). These rooms are high-risk areas, and if this recommendation is applicable and achievable for residential construction, a high-risk area should certainly be capable of achieving such a standard. It is suggested that additional study in this area be completed to further identify this value as an attainable standard as well as develop more efficient methods to determine room leakage areas for projects of larger scale.

REFERENCES

AIA. 2006. Guidelines for Design and Construction of Hospitals and Healthcare Facilities. Washington, DC: American Institute of Architects.

ASHRAE. 2005. 2005 ASHRAE Handbook-Fundamentals, I-P Ed, Chapter 26, "Ventilation and Infiltration." Atlanta: American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc.

CDC. 2005. Guidelines for Preventing the Transmission of Mycobacterium Tuberculosis in Health-Care Settings. MMWR 2005, 54(No. RR-17): 63-65. Washington, DC: Centers for Disease Control and Prevention.

HICPAC. 2003. Guidelines for Environmental Infection Control in Health-Care Facilities. Chicago: American Society for Healthcare Engineering Publications.

Hayden, C.S., M. Fischbach, and O.E. Johnston. 1997. A model for calculating air leakage in negative pressure isolation areas. National Institute of Occupational Safety and Health Report ECTR 212-05c. Cincinnati, OH: Department of Health and Human Services.

Hewett, M.J., and R.D. Hermans. 2006. Strategies to Reduce the Spread of Airborne Infections in Hospitals: Survey of Design Practice. NIST GCR 05-883. Gaithersburg, MD: National Institute of Standards and Technology.

Lstiburek, J. 2004. Builder's Guide to Cold Climates. West-ford, MA: Building Science Press Inc.

Murray, W.A., A.J. Streifel, T.J. O'Dea, and F.S. Rhame. 1988. Ventilation for protection of immune compromised patients. ASHRAE Transactions 94:1185-91.

Rice, N., A. Streifel, and D. Vesley. 2001. An evaluation of hospital special-ventilation-room pressures. Infection Control and Hospital Epidemiology 22:19-23

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.

Streifel, A.J. 2006. Pressure management in health-care facilities. HPAC Engineering 78(8):34-40.

Streifel, A.J. 2003. Airborne infectious disease: Best practices for ventilation management. HPAC Engineering 75:97-104.

CGSB. 1986. CGSB Standard 149.10-M86, Determination of Airtightness of Building Envelopes by the Fan Depressurisation Method. Canadian General Standards Board.

Gary Nelson

Member ASHRAE

Andrew Geeslin is a medical student at the University of Minnesota-Twin Cities, Minneapolis, MN. Andrew Streifel is an environmental-health specialist in the Department of Environmental Health and Safety, University of Minnesota, Minneapolis, MN. Gary Nelson is president of The Energy Conservatory, Inc., Minneapolis, MN.

(1.) Minneapolis Duct Blaster[R]. Donated by The Energy conservatory, Minneapolis, Minnesota, for this project.

(2.) DG-700 digital pressure and flow gauge. Donated by The Energy Conservatory, Minneapolis, Minnesota, for this project.

(3.) TECLOG, Version 1.2. Donated by The Energy Conservatory, Minneapolis, Minnesota, for this project.

(4.) TECTITE, Version 3.1.