Acoustics and Plenums

With the trend to more lightweight building construction, as well as the development of mixed-use commercial and institutional spaces, the control of noise from heating, ventilating and air-conditioning (HVAC) equipment has become a critical component of building design. A typical fan in a commercial building, for example, might generate 80-90 decibels of low frequency noise.

The noise can be a major annoyance for occupants and it is difficult to mitigate within the building envelope.

The technical committee for sound and vibration (TC-2.6) at the American Society of Heating, Refrigerating, and Air-Conditioning Engineers (ASHRAE) identified the problem with low-frequency noise in the mid-1990s. Their work in this area included evaluating the methodology used by HVAC engineers to predict the performance of a critical element in controlling low-frequency noise — the plenum chamber.

Consequently, this plenum question was addressed through a three-year experimental investigation of the acoustic and aerodynamic characteristics of plenums recently completed at the Vibro-Acoustics Aero-Acoustic Laboratory in Toronto. The goal of this ASHRAE-funded research (ASHRAE RP-1026) was to test a wide range of common HVAC plenums and develop an improved method to predict acoustic performance. As a result, a significant amount of full-scale plenum performance analysis is now available to building engineers.

Plenum as sound attenuator

An HVAC plenum is typically a large sheet metal box connected to a fan and multiple ducts to aid in the distribution of ventilation or conditioned air. In addition to being an effective way to group ducts together aerodynamically, an HVAC plenum is often used to attenuate sound, usually low frequency sound, within a ducted fan system. Acoustically, the plenum has been described as a type of passive sound attenuator that primarily achieves its performance through an abrupt area change. Many HVAC plenums include absorptive wall lining materials such as a standard fibre board or a perforated metal liner backed by fibreglass media to increase the broadband nature of their sound attenuation.

Prior to our research work, the most common plenum prediction tool was the equation found in the ASHRAE HVAC Applications Handbook of 2003. That tool was a result of the limited plenum research work done in 1958. In this early work, a few small plenums were tested, with the largest dimension less than 3 feet. Inlet area, multiple outlet openings, side outlet openings, and the presence of airflow were not considered as variables affecting a plenum’s performance characteristics.

Multiple configurations tested

In our research, 117 unique plenum configurations were tested, with dimensions ranging from 2 x 4 x 3 ft. up to 10 x 10 x 10 ft. This was a large and realistic range of sizes, representing those commonly found in today’s HVAC systems.

To account accurately for the numerous geometric variables, the research project applied a two-stage investigation. The first stage investigated plenums as in-line expansion chambers where the inlet and outlet openings were of the same size and located on opposite walls, each with seven different wall linings typically found in HVAC systems. Following this initial series of tests, the second stage of the investigation considered common geometric variables that had an impact on the inlet/outlet openings, such as opening quantity, size and location (e.g. side vs. end), and offset angles.

Engineers and consultants use either of two methods for calculating sound transmission through ductwork. One method uses the Transmission Loss (TL), or the change in acoustic energy as measured at the inlet and outlet openings of the plenum. The other method, Insertion Loss (IL), takes the sound level difference as measured at a receiving location, with and without the plenum inserted into the duct system.

A large number of the plenum configurations were tested as per industry standards to determine TL and IL respectively.

The difference between the two methods was found to be 3.5 to 5.0 dB at frequencies below the duct opening’s cut-off frequency, and 1.5 to 3.0 dB for frequencies above the cut-off frequency. The cut-off is the point in the frequency spectrum where the sound waves moving down the duct change from low-frequency plane waves to high frequency multi-mode waves. The cut-off frequency is defined by the cross-sectional dimensions of the duct that transmits the sound into the plenum. While our research took both the TL and IL methods into account, we conducted broader investigations using the TL method as it is the more accepted approach.

Simplified surface area coefficient

An analysis between the key low-frequency data and singular geometric variables, such as the wall surface area and volume of the plenum, produced strong correlations among plenums with various wall dimensional configurations. This result suggests that a simplified surface area coefficient may account for the reactive attenuation of low vibration noise in the plenum, analogous to the performance of an automobile’s muffler and independent of the actual plenum wall configuration. For example, a 4 x 5 x 5 ft. plenum (width x height x length) would result in approximately equal low-frequency TL as a 4 x 6 x 4 ft. plenum, in terms of its reactive attenuation of low vibration noise.

For a characterization of the absorptive performance of a plenum, further analysis on wall effects produced interesting results. For a particular plenum configuration a strong trend was shown when we changed the baseline bare metal wall into another more absorptive wall type, independent of the plenum inlet or outlet opening sizes. As a practical method to characterize the absorptive nature of the plenum, a mathematical relationship for “Wall Effect” was derived to be used in combination with the surface area coefficient. These two acoustical phenomena, reactive and passive, were incorporated into the new predictive equation.

New hybrid prediction model

We were able to develop an improved method to predict the acoustic performance of plenums in which the low-frequency reactive and absorptive components of TL were combined in a low-frequency model. This new approach, along with a regression analysis on the high frequency data, resulted in a two-stage, or hybrid, prediction model that appropriately represents the distinct frequency ranges.

The research included an investigation into the acoustical effects due to inlet/outlet geometric variables and corresponding aerodynamic pressure losses. The results are summarized as follows:

* Off-Set Angle (for end-inlet and end-outlet configurations). Significant TL increases occurred for high frequencies and small TL reductions occurred for low frequencies.

* Elbow Effect (change in TL from an end-inlet/end-outlet to an end-inlet/side-outlet configuration). Generally, a broadband increase in TL was realized, except for TL reductions that corresponded to a diminished in-line plenum tuning.

* Multi-Outlet vs. Single Outlet. A negligible change in attenuation (< 1.5 dB) was produced with multiple-outlet plenums as compared to corresponding single outlet in-line or side-out elbow configuration plenums. Therefore once a critical path has been established in a plenum system, the presence of multiple openings may be ignored.

* Outlet/Inlet Area Ratio not equal to 1.0. Below the cut-off frequency, the reduction in TL was found to be proportional to the duct opening area ratio, independent of the plenum size, wall lining type, and frequency.

* Pressure Drop relationships were derived for all configurations in this study, including the in-line or flow-through type plenum that was found to produce a pressure drop at 40% of the inlet duct’s velocity pressure. As anticipated, the pressure drop was found to be independent of w
all type and plenum volume.

This extensive project has derived a new prediction model of transmission loss between 50 and 5,000 hertz for ducted plenums between a 20 cu. ft. to 1,100 cu.ft. size range, and inlet and outlet connection sizes between 12 and 48 inches. The hybrid prediction model is divided into two ranges of frequencies, determined by the cut-off frequency of the duct that transmits the sound into the plenum.

In the low-frequency range (below the cut-off frequency), the amount of internal wall surface area and the corresponding wall absorption characteristics were found to be the key parameters in determining the plenum TL. This is contradictory to the traditional notion that has assumed the primary parameter is the expansion ratio of the duct area to the plenum cross-section area. The existing ASHRAE algorithm, with the addition of two regression coefficients, was found to be a very acceptable prediction tool for frequencies above the cut-off frequency.

With an improved understanding of plenum performance, engineers can help reduce the intrusive low frequency HVAC noise within occupied spaces. Also, by designing appropriately sized plenums, they can realize energy and space usage savings in buildings.

In June 2004, a paper highlighting the practical application of this new TL prediction tool and the pressure drop algorithms derived from this research received an ASHRAE Symposium Paper Award.

Emanuel Mouratidis, P.Eng., is a consulting engineer with Aercoustics Engineering of Toronto.