December 15, 2012

Abstract: 

In multifamily buildings, central (typically rooftop) ventilation systems often have poor overall performance, overventilating some portions of the building (resulting in excess energy use), while simultaneously underventilating other portions of the building (resulting in diminished indoor air quality). These issues are often tied to multistory stack effects (warm air rising at cold outdoor conditions), and a lack of compartmentalization (airtightness) between floors and between units. These issues are exacerbated by the presence of multistory shafts (e.g., elevator shafts, stairwells, and ventilation shafts). Central corridor supply and makeup air systems combined with rooftop central exhaust systems are particularly problematic. The recommended solution is to isolate the units from one another and from corridors, shafts, elevators, and stairwells by means of greater airtightness.

Executvie Summary

In multifamily buildings, central (typically rooftop) ventilation systems often have poor overall performance, overventilating some portions of the building (resulting in excess energy use), while simultaneously underventilating other portions of the building (resulting in diminished indoor air quality). These issues are often tied to multistory stack effects (warm air rising at cold outdoor conditions), and a lack of compartmentalization (airtightness) between floors and between units. These issues are exacerbated by the presence of multistory shafts (e.g., elevator shafts, stairwells, and ventilation shafts). Central corridor supply and makeup air systems combined with rooftop central exhaust systems are particularly problematic. The recommended solution is to isolate the units from one another and from corridors, shafts, elevators, and stairwells by means of greater airtightness.

Duct sealing of exhaust shafts has significant energy benefits. Codes require minimum exhaust flows from spaces (kitchens and bathrooms). Leaky exhaust duct shafts pull additional exhaust air out of interstitial spaces (i.e., "stealing" air), which does not help meet the minimum exhaust requirements and results in overventilation.

Building Science Corporation performed a series of field tests at a mid-rise test building with Innova Services Corporation, which is a Philadelphia-based firm that works in the affordable housing industry. The test building was undergoing a major energy audit and retrofit that was completed over the course of 2011 and 2012. Ventilation upgrades were one component of the retrofit, which also included lighting, space heating, domestic hot water, and appliance upgrades.

The retrofit exhaust ventilation system replaced the existing rooftop fans with variable-speed, pressure controlled, electronically commutated motor rooftop units. Apartment unit exhaust registers were changed to localized powered exhaust fans, controlled by wall switch timers, to supply ventilation on an as-needed or timed basis. When the unit exhaust fan is off, some limited baseline ventilation occurs through the fan damper; when a unit exhaust fan is turned on to respond to pollutant loads, the rooftop exhaust rate increases (maintaining negative pressure in the shaft).

The corridor ventilation system was intended to replace a non-operational rooftop makeup air system, which was deactivated because of excess energy costs associated with heating large airflows. The retrofit system switches to floor-by-floor ventilation, tempering ventilation supply air with indoor air to avoid cold air complaints.

Pre-retrofit air leakage was measured in two exhaust duct shafts: leakage was more than double recommended levels (per Zuluaga and Fitzgerald 2010). The fan depressurization leakage measurements were 26% and 13% of the nominal (callout) flow, using the cfm 50 (cubic feet per minute at 50 pascal test pressure) leakage metric. Summed unit exhaust airflows were compared with the rooftop airflow measurements, providing a "calculated leakage" that correlated reasonably well with the nominal leakage.

Airflows from the units and the rooftop fans were compared to the nominal plan callouts. Unit airflows were lower than callouts (78% weighted average), while rooftop airflows were higher than callouts (109% on average). The mismatch between rooftop and unit flows shows the effect of exhaust shaft duct leakage.

Unit air leakage was measured with depressurization testing, showing high air leakage and poor compartmentalization. Much of the leakage appeared to be above the suspended ceiling: air sealing details required for fire rating of the demising walls were never completed.

In the post-retrofit testing, exhaust airflows were measured at the rooftop unit and in the apartment units, with individual unit exhausts turned on and off. Although rooftop measurements had high uncertainty because of wind effects, it appears that the rooftop fan correctly increases its flow when additional unit exhaust fans are turned on. Airflows at the unit exhausts matched expected patterns, including some airflow (bypass) through the unit exhaust fan’s damper with the unit off and higher flow with the unit on. Rooftop exhaust fan airflows were roughly 50% of original plan callout values, but met ASHRAE 62.1 and 62.2 targets (ASHRAE 2010a, 2010b).

Power draw measurements show a substantial improvement in fan efficiency: calculated efficiency based on this measurement was 10–20 cfm/W, compared to the pre-retrofit state of 1.3– 2.1 cfm/W. This is ascribed to improved efficiency of the fan motor and the reduction in airflow (40%–60% of original design), the latter which results in effectively oversized ductwork. Efficiency levels this high should not be expected with a new duct system sized for the measured flows. These metrics do not include the contribution of the unit exhaust fans (at 12 W each).

Similar to the pre-retrofit measurements, calculated duct leakage could be estimated from the difference between rooftop and unit exhaust measurements. The calculated duct leakage (as a percentage of flow from the rooftop fan) is 40%–50% with the unit exhausts off and 15%–25% with the unit fans on. Although the “unit fans off” calculated duct leakage is a very high fraction of the flow (40%–50%), at Exhaust Fan-1, for example, the absolute value (85–100 cfm) is
comparable to or less than the calculated duct leakage in the pre-retrofit system (110 cfm). The roof curb connection was one significant source of duct leakage that was corrected in the retrofit. No duct sealing was implemented beyond rooftop curb work, such as aerosol-based sealant or hand-applied mastic to accessible portions.

Monitoring equipment was installed on two exhaust fans. A correlation was seen between fan speed variations and wind events. There was also a diurnal variation in fan speed on a cycle that matches typical occupancy, with low variations during sleeping hours (10 p.m. to 6 a.m.), and greater variations during the day.

The retrofit corridor supply ventilation system was installed and tested; basic function is as per design, with outdoor supply air being tempered or diluted with interior air (at ratios of 1:2 to 1:3.5) for occupant comfort. The system, however, had relatively high static pressures and relatively low fan efficiencies (0.9 cfm/W for net outside air).

Some limited conclusions can be drawn from the collected data. If the reduction in ventilation flow is applied across all fans, heating energy savings are estimated at roughly 4,000 therms/yr, which can be compared with wintertime heating use estimated at 16,000 therms/yr. The electrical savings resulting from upgraded fans (and reduced ventilation rates) are also significant, changing from roughly 4,500 kWh/month to 290 kWh/month in fan use energy, including the
estimated contribution of unit exhaust fans.

1 Inroduction

In multifamily buildings, central (typically rooftop) ventilation systems often have poor overall performance, overventilating some portions of the building (resulting in excess energy use), while simultaneously underventilating other portions of the building (resulting in diminished indoor air quality). At this point, there are some tested and recommended solutions for solving these issues in both new construction and retrofit situations.

The recommended retrofit solutions, however, have ramifications in terms of installed cost and simultaneous access to multiple units to execute the retrofit. Alternate solutions, involving variable-speed rooftop exhaust fans and individual unit powered exhaust fans, can be considered. This research project examined the performance of existing multifamily ventilation central systems and explored alternative solutions in retrofit situations. Engineering calculations and field-testing of components and systems were used to complete this work.

Building Science Corporation (BSC) performed this research with Innova Services Corporation (“Innova”), a Philadelphia-based firm that works in various sectors of affordable housing, including construction project management, general contracting, and building retrofit services.

This research on multifamily ventilation systems was conducted at a building retrofit project that was recently completed. The test building was the James J. Wilson Mercy-Douglass Residences,1 which is senior housing constructed in the mid-1980s (see Figure 1).

Figure 1: Mercy-Douglass Residences overview (L) and site conditions (R)

2 Multifamily Building Ventilation Background

2.1 Stack Effect, Ventilation, and Compartmentalization

The dominant forces causing air movement in buildings are wind, temperature differenceinduced stack effects (otherwise known as natural buoyancy), and mechanical pressurization and depressurization (Straube and Burnett 2005; Hutcheon and Handegord 1995).

Taller buildings are often dominated by stack effect. Wilson and Tamura (1968) and Hutcheon and Handegord (1995) discuss the fundamental physics. The simplified consequences of stack effect are shown in Figure 2. In cold weather, outdoor air infiltration occurs on lower floors and interior air exfiltration occurs on upper floors. Research has shown that in cold climates in the winter, stack effect dominates over wind effects (Feustel and Diamond 1996; Palmiter et al.
1995; Francisco and Palmiter 1994).

If there is interior leakage between floors (common in the stock of multifamily buildings), upper floors are effectively “ventilated” with air from lower floors (i.e., replacement air comes from other units), as shown in Figure 2. This results in odor and pollutant transfer, compromised smoke control and fire safety, highly varying rates of air change between floors, difficulties in maintaining even temperature set points (especially in buildings without zoned controls or thermostats), and excess energy use.

Figure 2: Stack effect in multifamily buildings (simplified)

Stack effect problems are exacerbated by the presence of multistory shafts, such as elevator shafts, stairwells, and ventilation shafts. These shafts have stack-driven pressure differences across their walls, resulting in an additional potential air transfer path (Figure 3). Note that if the shaft is a multistory mechanical duct, the stack effect pressures will be superimposed on the mechanically induced pressures on a seasonal basis, resulting in uneven distribution of
ventilation flows. Upper floor units are typically overventilated in wintertime because of combined stack and mechanical pressures.

Figure 3: Stack effect in multifamily buildings (simplified), showing shaft effects

The solution proposed by Lstiburek (2005) and others is to isolate the units from one another and from the corridors, shafts, elevators, and stairwells, by means of greater airtightness or compartmentalization (Figure 4). This limits stack effects largely to the floor-to-ceiling height difference of a given unit, as opposed to stack acting over the height of the building. Elevators should be located in vestibules, lobbies, or other airlocks to isolate them from corridors.

Figure 4: Idealized compartmentalized multifamily building, showing stack per floor

Figure 5 shows the air barrier compartmentalization concept. Each unit is isolated from adjacent units and from the exterior by an air barrier system with a maximum air leakage rate of 2.0 L/(s∙m2) at 75 Pa, which is equivalent to 0.30 cfm at 50 Pa/ft2 of enclosure (cfm 50/ft2). The interunit separation must also meet the specific fire resistance rating requirement for the given separation.

Compartmentalization also reduces overall air leakage through the building enclosure by limiting stack effect pressures across the exterior enclosure to those associated with one compartmentalized unit’s height. This is shown in concept by comparing Figure 3 with Figure 4.

This compartmentalization principle can be applied to ventilation systems as well. Ideally, ventilation air would be supplied and exhausted through the exterior wall (as shown in Figure 5, right), not across interior pressure boundaries, which compromises compartmentalization.

Individual unit ventilation systems have a further benefit in that they can be controlled on a unit-by-unit basis, either by the occupant or by building management. A central ventilation system, in contrast, typically provides a constant exhaust rate for all units at all times, resulting in overventilation in some units and underventilation in others, assuming diversity of pollutant loads. For instance, a temporarily unoccupied unit would be overventilated if operated identically to an occupied unit. Of course, overventilation has an associated energy penalty.

For reference, the calculated stack effect pressure over a 30-ft and 40-ft height is graphed against exterior temperature . . .

Download complete report here.

Footnotes:

  1. 4511 Walnut Street, Philadelphia, PA, 19139