HVAC Thesis

THE IMPACT OF FILTER LOADING ON RESIDENTIAL HVAC
PERFORMANCE

A Thesis
Presented to
The Academic Faculty

by

Abraham Kruger

In Partial Fulfillment
of the Requirements for the Degree
Masters of Science in Building Construction in the
School of Building Construction

Georgia Institute of Technology
December 2013

Copyright 2013 by Abraham Kruger
THE IMPACT OF FILTER LOADING ON RESIDENTIAL HVAC
PERFORMANCE

Approved by:

Dr. Javier Irizarry, Advisor
Dr. Brent Stephens
Department of Civil, Architectural
and Environmental Engineering
Illinois Institute of Technology

Dr. Daniel Castro-Lacouture

Rick Porter

Date Approved: November 15, 2013

iv
ACKNOWLEDGEMENTS

This thesis would not have been possible without the generous support of my
thesis committee, friends and family, test home volunteers, and numerous industry
professionals. In no particular order, I would like to thank Abbey Kruger, Evonne
Kruger, Isaak Kruger, Matt Laliberte, Carl Seville, Laura Capps, Terrel Broiles, Gary
Nelson, Jeffrey Saules, Ela Orenstein, Georgia Hill, Roxanne Greeson, Kevin Thompson,
Bob Mason, Eyu-Jin Kim, Jacquelyn Strickland, Lorie Wooten, and Amin Esmaeili.
I would also like to acknowledge the generous support of the National Housing
Endowment who provided a Homebuilding Education Leadership Program (H.E.L.P.)
grant to the School of Building Construction at Georgia Tech and through which I was
able to pursue this research as a Residential Construction Industry Applied Research
Fellow.

v
TABLE OF CONTENTS
Page
ACKNOWLEDGEMENTS iv
LIST OF TABLES viii
LIST OF FIGURES ix
LIST OF SYMBOLS AND ABBREVIATIONS xi
SUMMARY xiv
CHAPTER
1 Introduction 1
2 Literature Review 3
HVAC Prevalence 4
Heating and Cooling Systems 5
HVAC Motors 6
Types of Filters 7
Filter Efficiency 8
Coil Fouling 10
Pressures within HVAC Systems 11
Airflow within HVAC Systems 15
Filters Impact on Energy Consumption 18
3 Methodology 22
Research Design 22
Test House Selection 23
Filter Loading Simulation 25
Data Collection 27
vi
Fan Curve Determination Procedure 30
Data Analysis 33
4 Field Results and Data Analysis 35
Test System Descriptions 35
Test System Locations 35
Test System 1 36
Test System 2 37
Test System 3 38
Test System 4 38
Test Systems 5 and 6 39
Analysis 40
Filter Pressure Drop and System Airflow 40
Filter Pressure Drop and Temperature 43
Filter Pressure Drop and Absolute Humidity 47
Airflow and Temperature 49
Airflow and Absolute Humidity 52
Airflow and System Capacity 54
5 Conclusion 60
Filter Pressure Drop and System Airflow 60
Filter Pressure Drop and Temperature 61
Filter Pressure Drop and Absolute Humidity 61
Airflow and Temperature 61
Airflow and Absolute Humidity 62
Airflow and System Capacity 62
Analysis Summary 62
vii
Recommendations 63
APPENDIX A: Field data from 6 test systems 64
REFERENCES 70
viii
LIST OF TABLES
Page
Table 1: Stephens et al. results

20
Table 2: Test system characteristics 24
Table 3: Measurement type and location 31
Table 4: Testing instrumentation 31
Table 5: Filter pressure drop and system airflow 42
Table 6: Filter pressure drop and T 43
Table 7: Z-scores for each T measurement 45
Table 8: Filter pressure drop and absolute humidity differences across the coil 48
Table 9: Filter pressure drop and T across the evaporator coil 50
Table 10: System airflow and W across the evaporator coil 52
Table 11: Relationship of airflow and sensible capacity for all systems 55
Table 12: Relationship of airflow and latent capacity for all systems 56
Table 13: Relationship of airflow and total capacity for all systems 57

ix
LIST OF FIGURES
Page
Figure 1: U.S. energy use by sector 3
Figure 2: Air Conditioning equipment growth in the US 5
Figure 3: System and fan curves for medium-, high-, and low-pressure-drop filters 14
Figure 4: The Alnor EBT Balometer 16
Figure 5: TrueFlow Air Handler Flow Meter 17
Figure 6: The research process. 22
Figure 7: Percent of new homes containing central air conditioners in the south 23
Figure 8: An example of a partially taped filter (TrueFlow Taped 1) 26
Figure 11: Vertical HVAC system with one central filter. 28
Figure 12: Horizontal HVAC system with one central filter. 29
Figure 13: Vertical HVAC system with two central filters. 30
Figure 14: The location of the 4 test homes in the Atlanta metro-region. 36
Figure 15: Test system 1 37
Figure 18: Test systems 5 and 6 40
Figure 19: Relationship of induced filter pressure drop and system airflow 41
Figure 20: System airflow versus filter pressure drop. 42
Figure 21: Relative T across the coil versus filter pressure drop for each test system 46
Figure 22: Relationship of filter pressure drop and T across all systems. 47
Figure 23: Relationship of filter pressure drop and W for each test systems 48
x
Figure 24: Relationship of filter pressure drop and W for all systems. 49
Figure 25: Relationship of airflow and T for each system. 50
Figure 26: Relationship of airflow and T for all systems. 51
Figure 27: The relationship of airflow and absolute humidity for each system 53
Figure 28: Relationship of airflow and absolute humidity for all systems. 54
Figure 29: Relationship of airflow and sensible capacity for all systems 55
Figure 30: Relationship of airflow and latent capacity for all systems. 56
Figure 31: Relationship of airflow and total capacity for all systems. 57

xi
LIST OF SYMBOLS AND ABBREVIATIONS

AC Air conditioner
AC Alternating current
ACCA Air Conditioning Contractors of America
ACDX Direct expansion AC
AHS American Housing Survey
AHU Air handling unit
ANSI American National Standards Institute
ASHRAE American Society of Heating, Refrigerating and Air-Conditioning
Engineers
Btu/hr British thermal units per hour
C Celsius
C Specific heat of air, kJ/kgK
C
total
Coefficient of proportionality for the entire system
C
return
Coefficient of proportionality for the return duct
C
filter
Coefficient of proportionality for the filter
C
CC
Coefficient of proportionality for the cooling coil
C
HC
Coefficient of proportionality for the heating coil
C
S1
Coefficient of proportionality for the first supply duct branch
C
S2
Coefficient of proportionality for the second supply duct branch
C
S3
Coefficient of proportionality for the third supply duct branch
CFM Cubic feet per minute, ft
3
/min
COP Coefficient of performance
DC Direct current
DOE Department of Energy
ECM Electronically Commutated Motor also known as a Brushless DC Motor
xii
EER Energy Efficiency Ratio
EIA Energy Information Administration
EST Eastern Standard Time
F Fahrenheit
g Gram
fan
Efficiency of the fan
h
fg
Latent heat of vaporization for water, assumed constant, (970 Btu/lb)
HUD Housing and Urban Development
HVAC Heating, Ventilation and Air Conditioning
IAQ Indoor air quality
ICC International Code Council
IRC International Residential Code
IWC Inches of water column
k Unit of measurement equal to 1,000
K constant based on measured airflow and pressure
K Kelvin
kWh kilowatt-hours
motor
Efficiency of the fan motor
m
3
/s cubic meters per second
MERV Minimum efficiency reporting value
Q
reference
Volumetric flow rate of air with measuring device installed, m
3
/hr (cfm)
P System pressure, Pa (IWC)
Air density, kg/m
3

Pa Pascal
PSC Permanent split capacitor
Q System airflow rate, m
3
/s (cfm)
Q
fan
Volumetric flow rate of air flowing across the fan, m
3
/s (cfm)
xiii
s Standard deviation
VAV Variable air volume
VOC Volatile organic compound
W Watts
W Humidity ratio
W
fan
Power draw of fan, W
i
Sample value
Sample mean
P
reference
Supply plenum pressure with measuring device installed, Pa (IWC)
P
operating
Operating supply plenum pressure, Pa (IWC)
T Temperature difference across the cooling coil, K (F)
W Humidity ratio difference across the cooling coil, kg/kg

xiv
SUMMARY
Residential and commercial buildings account for approximately 41% of total
energy use in the US. Within households, approximately 50% of total energy use is
associated with space heating and cooling. Heating, ventilation, and air-conditioning
(HVAC) systems are designed to provide occupant comfort by meeting heating and
cooling loads safely and efficiently. Air cleaning devices, particularly particle air filters,
are important components of HVAC systems that prevent damage to HVAC equipment
and improve indoor air quality (IAQ) by reducing airborne particle concentrations.
HVAC filters, however, can also have significant impacts on the performance of HVAC
equipment in both residential and commercial buildings.
Coil fouling, or the deposition of airborne particles on the evaporator coil inside
an HVAC systems air handling unit (AHU), will increase system pressure drop and
reduce heat transfer effectiveness, which decreases airflow and air conditioner (AC)
performance. Although filters can increase AC performance by decreasing coil fouling,
filters may also have energy implications, particularly if they are higher pressure drop
filters or if they become dirty (or loaded) over time. In large commercial HVAC
systems with variable speed blowers, energy implications are simple: fans will simply
draw more power to overcome a greater pressure drop to deliver the same amount of
required cooling. In smaller residential systems without sophisticated airflow controls,
recent research has shown that as filters become loaded, pressure drop across the filter
increases and airflow is restricted. Cooling systems should therefore run longer as airflow
is reduced to provide adequate cooling at the reduced capacity, although little quantitative
information exists on the magnitude of the impacts of filter pressure drop on airflow
rates, cooling capacities, and system runtimes in real residential systems. Complicating
the issue is that while most homes currently have inefficient blowers without flow
controls (i.e., permanent split capacitor, or PSC, motors), new AHU products on the
xv
market utilize more sophisticated fans with flow controls (i.e., those with electrically
commutated motors, or ECM, blowers, also called brushless permanent magnet, or BPM,
blowers).
Therefore, the following report reviews recent research on HVAC filters for
central forced-air air-conditioning units and electric heat pumps and presents the findings
of an in-situ evaluation of AC performance under simulated filter loading conditions. The
study hypothesized that it was possible to develop a methodology for simulating filter
loading in-situ that would allow for the observation of the impact of filter loading on AC
performance in-situ and provide a greater understanding of when a filter is dirty and
thus inform when it should be replaced.
Six central AC systems in the Atlanta metro-region were evaluated in this work.
Several conditions of filter loading were artificially induced in the test systems and filter
pressure drops, airflow rates, temperature and humidity differences across the coil were
all measured during pseudo-steady-state cooling operation, which allowed for developing
relationships between simulated filter loading, airflow, and sensible and latent capacity.
Filter loading was simulated by installing an Energy Conservatory TrueFlow plate
airflow metering device and partially taping off the face at three different increments.
This resulted in measurements at 5 discrete static pressure conditions: no filter, TrueFlow
measurement, TrueFlow Taped #1, TrueFlow Taped #2, and TrueFlow Taped #3,
increasing in simulated filter pressure drop at each increment. These in-situ
measurements revealed that as filter pressure drop increased, airflow rates generally
decreased, particularly for the known PSC blowers, as is expected from the literature.
Two of the test systems were apparently ECM blowers as they responded to increased
filter pressure drop by nearly maintaining airflow rates until reaching a maximum
pressure and rapidly decreasing in flow, which is consistent with other ECM data.
Therefore the data herein are considered generally representative of both types of
systems, even with a limited data set.
xvi
With moderate certainty, it was found that as induced filter pressure drop
increased, the difference in temperature across the evaporator coil (T) also increased in
these systems. These data support what other laboratory and field studies have shown: as
airflow rates are reduced in the presence of larger filter pressure drops, sensible capacity
will not decrease linearly with flow because the temperature difference across the coil
increases slightly and supply air is delivered at a lower temperature. This data also shows
that there is a stronger relationship between coil T and airflow as opposed to T and
induced filter pressure drop, as filter pressure drop did not impact airflow rates uniformly
in the systems.
There was no observed correlation between absolute humidity differences across
the evaporator coil (W) and either filter pressure drop or system airflow rates. In other
words, as airflow decreased so did sensible, latent, and total capacity, although these
relationships were not linear. Because reductions in cooling capacity can be linked to
increased system runtimes, this research can be used to inform decisions about maximum
filter loading values that should inform filter replacement schedules. Once a maximum
acceptable reduction in sensible capacity is established, this data can be used to identify
the airflow and filter pressure drop thresholds, which can impact future decisions about
filter replacement timing.
This relatively limited pilot study provides valuable proof of concept for an
approach to simulating in-situ filter loading and characterizing associated capacity
impacts. Taping the face of the TrueFlow plate was shown to work consistently well to
simulate filter loading. In the future, this study should be expanded to a greater number of
central air conditioning units and continuous measurements should be recorded using
data loggers, as one challenge was determining exactly when the AC systems reached
steady state. Greater certainty may be possible by recording continuous measurements
and waiting longer between each simulated filter condition.

1
CHAPTER 1
INTRODUCTION
Buildings account for approximately 41% of total energy use in the US (US
Energy Information Administration 2011). Within households, approximately 50% of
total energy use is for providing heating and cooling. Heating, ventilation, and air-
conditioning (HVAC) systems are designed to provide occupant comfort safely and
efficiently. HVAC systems control the indoor environment by heating, cooling,
recirculating, and filtering the air and managing humidity. When not properly designed,
installed, and/or maintained, these systems can increase energy consumption while
compromising indoor air quality (IAQ).
Air cleaning devices, particularly particle air filters, are important components of
HVAC systems that prevent damage to HVAC equipment and improve indoor air quality
(IAQ) by reducing airborne particle concentrations. HVAC filters, however, can also
have significant impacts on the performance of HVAC equipment in both residential and
commercial buildings. Coil fouling, or the deposition of airborne particles on the
evaporator coil inside an HVAC systems air handling unit (AHU), will increase system
pressure drop and reduce heat transfer effectiveness, which decreases airflow and air
conditioner (AC) performance. Although filters can increase AC performance by
decreasing coil fouling, filters may also have energy implications, particularly if they are
higher pressure drop filters or if they become dirty (or loaded) over time. In large
commercial HVAC systems with variable speed blowers, energy implications are simple:
fans will simply draw more power to overcome a greater pressure drop to deliver the
same amount of required cooling. In smaller residential systems without sophisticated
airflow controls, recent research has shown that as filters become loaded, pressure drop
across the filter increases and airflow is restricted. Cooling systems should therefore run
longer as airflow is reduced to provide adequate cooling at the reduced capacity, although
2
little quantitative information exists on the magnitude of the impacts of filter pressure
drop on airflow rates, cooling capacities, and system runtimes in real residential systems.
Complicating the issue is that while most homes currently have inefficient blowers
without flow controls (i.e., permanent split capacitor, or PSC, motors), new AHU
products on the market utilize more sophisticated fans with flow controls (i.e., those with
electrically commutated motors, or ECM, blowers, also called brushless permanent
magnet, or BPM, blowers).
To date, studies on the impact of filters on residential HVAC system performance
have used: (i) laboratory test systems (Siegel et al.

2002, Palani et al 1992); (ii) computer
simulation models (Nassif 2012); (iii) small samples of 2-16 systems in-situ (Chimunk
and Sellers 2000, Stephens et al 2012, Rodriguez et al 1996); and (iv) some combination
of these three methods (Yang et al 2004).

Research Objectives
The study hypothesized that it was possible to develop a methodology for
simulating filter loading in-situ that would allow for the observation of the impact of
filter loading on AC performance in-situ and provide a greater understanding of when a
filter is dirty and thus inform when it should be replaced. In-situ data can then be used
to evaluate and update computer simulation models as well as increase the model of
knowledge of actual HVAC performance in residences.

3
CHAPTER 2
LITERATURE REVIEW
The U.S. Energy Information Administration (EIA) publishes the Annual Energy
Review. The report includes data on: total energy production, consumption, and trade for
petroleum, natural gas, coal, electricity, nuclear energy, renewable energy, and
international energy; and financial, and environmental indicators. According to the most
recent edition, buildings account for approximately 41% of total energy use in the US as
shown in Figure 1 (US EIA 2011).

Figure 1. U.S. energy use by sector (US EIA 2011)

In 2005, US households consumed 0.88 quadrillion Btus and spent $25.26 billion
dollars on electricity for air conditioning (US EIA 2011)
1
. In that same year, US

1
The 2005 data is what is available in the most recent Annual Energy Review from the DOE/EIA.
4
households consumed 0.28 quadrillion Btus and spent $7.42 billion dollars on electricity
for heating purposes. Even a slight improvement in heating and cooling system
performance would result in significant energy consumption reductions across the
residential building stock.

HVAC Prevalence
Most homes require space conditioning, defined as heating, cooling, or both,
depending on the local climate. The International Residential Code (IRC, the building
code) requires that dwellings in cold climates contain heating equipment that can
maintain indoor temperature at a minimum of 68F (ICC 2012). There is no code
requirement for cooling equipment in any US climate, although central air-conditioning is
become ubiquitous in warmer climates.
The American Housing Survey for the United States (AHS) is sponsored by the
U.S. Department of Housing and Urban Development (HUD) and conducted by the U.S.
Census Bureau. National data is collected every 2 years and provides the most
comprehensive national housing survey in the United States. The data covers a range of
housing types, including single family, manufactured housing, and multifamily housing;
and housing and resident characteristics, such as family composition, income, housing
quality, neighborhood quality, housing costs, equipment, and fuel type. The most recent
data from 2009 was published in 2011 (U.S. Census Bureau 2009).
Based on the most recent AHS, there are a total of 130,112,000 housing units,
ranging from single family to multifamily and owner occupied to rental in the U.S.
Approximately 1.5% of these dwellings are seasonal properties. The median dwelling
was constructed in 1974 and contains 1,700 square feet. Approximately 63% of all
American dwellings contain central AC for a total of 82,475,000 central AC units.
Electric heat pumps are the primary heat source for 12% of American dwellings for a
total of 16,059,000 heat pumps. Figure 2 shows the growth of AC systems within the
5
U.S.

Figure 2. Air-conditioning equipment growth in the US (U.S. Census Bureau 2009).

US energy consumption for cooling is increasing as more homes are constructed
with air conditioning systems. Thus, the importance of central AC system performance is
also increasing.
Heating & Cooling Systems
The key components of HVAC systems are the equipment used to supply energy
for heating or extract energy for cooling, the fuel sources, and the method used to
distribute the heating and cooling throughout the house. In forced-air systems, heated and
cooled air is delivered by bulk convection into spaces that need heating or cooling.
Individual systems that serve multiple areas of the home are referred to as whole-house
systems, and those that serve only sections or single rooms are called local or non-
distributed systems. Either of these can provide heating only, cooling only, or both,
depending on the system type and climate requirements. Fuels can be fossil fuels (e.g.,
6
gas or oil), electricity, wood, or, in some cases, solar energy can be employed. Selecting
the most appropriate distribution system and equipment, combined with proper design
and installation, are critical to creating an effective and efficient HVAC system.
Central AC systems and heat pumps consist of a compressor and two coils made
of copper or aluminum tubing (typically one located inside and one outside) that are
surrounded by aluminum fins to aid heat transfer. Refrigerant, a chemical compound that
transfers heat as it changes from a liquid to a gas and back, flows back and forth between
the indoor and outdoor coils. The compressor, located outdoors, is a mechanical pump
that increases the refrigerant pressure, raising its temperature. In a conventional AC
system during cooling operation, the indoor evaporator coil serves to evaporate the
refrigerant, changing its phase from a liquid to a gas, which absorbs energy from the air
passing over the coil and thereby cools the airstream. Refrigerant is then piped to the
condenser unit and heat is rejected to the outside, which acts as a heat sink. Humidity is
also removed from the warm airstream as it passes over the cooler evaporator coil.
Moisture leaves the vapor phase and is removed from the air as liquid water and drained
to the exterior or a sewer system. Heat pumps operate similarly in the cooling mode.
In the heating mode, heat pumps operate in reverse. The outdoor condenser unit
acts as an evaporator coil whereby liquid refrigerant in the outdoor unit extracts heat from
the outdoor air and evaporates into a gas. The indoor coil now acts as a condenser and
rejects heat into the indoor environment as it condenses back into a liquid. A reversing
valve, near the compressor, can change the direction of the refrigerant flow for cooling as
well as for defrosting the outdoor coil in winter.

HVAC Motors
Electric motors are classified as either alternating current (AC) or direct current
(DC) motors. AC motors are further broken down by the number of phases and whether
they are synchronous or asynchronous (induction) motors. Approximately 90%
7
residential fans are permanent split-capacitor (PSC) (Sachs et al 2002), which do not have
flow controls to maintain specified airflow rates. These fractional horsepower, AC
induction motors usually have multiple speed windings, are low cost and tend to be 55%
efficient under full-load (Murray 2012), although electric efficiencies measured in-situ
are often much lower (Stephens et al 2010).
The most common alternative to PSC motors is electronically commutated motors
(ECM), also called brushless permanent magnet (BPM) blowers. Unlike PSC motors,
ECMs are able to adjust voltage levels to optimize motor torque as rotational speed is
reduced. For these reasons, ECM motors tend to be much more efficient than comparably
sized PSC motors, especially at low rotational speeds. They are also typically installed
with flow controls that allow systems to maintain airflow rates in the presence of higher
pressure drops. While ECM motors are more efficient, they also are more expensive due
to the permanent magnets employed and additional circuitry.
Different HVAC operation modes require different airflows, which require
varying blower motor speeds. Blower fan speeds for heat pump and air-conditioning
modes are typically higher than a fan-only mode where no cooling or heating is being
performed. Fan speeds are selected by the HVAC control system based on the particular
mode of operation. Motor efficiency may vary based on fan speed. For example, unlike
PSC motors, the efficiency of ECMs tends to increase as fan speed slows. Test results
suggest that, on average, ECM motors represent a 51 percent full-load efficiency
improvement over PSC motors (Murray 2012).

Types of Filters
Air filtration devices, including particle air filters, are used to protect HVAC
equipment from damage and coil fouling and to reduce concentrations of airborne
particulate matter in indoor environments. Filters are normally installed as part of a
forced-air HVAC system. The three primary types of filters are mechanical, pleated, and
8
electronic. The most common type used in homes, mechanical air filters, uses synthetic
fibers or fiberglass to remove particles as they pass through the filter media. Pleated air
filters are more effective than other mechanical air filters because they contain more fiber
per square inch than mechanical filters. Electronic air filters use electricity to attract
oppositely charged particles to metal fins. The particle removal efficiency of mechanical
and pleated filters often improves as they become dirty because smaller and smaller
particles are captured in the increasingly fine openings and loaded dust acts to increase
effective fiber area (Earnest et al 2001). The effectiveness of electronic filters typically
decreases over time without cleaning because the metal fins become ineffective when
loaded. Filters may be installed at the air handling unit, return register, or as separate
stand-alone equipment. This study focuses on filters installed at the air handling unit, as
is common in many homes.

Filter Efficiency
ANSI/ASHRAE Standard 52.2-2007 is used to address the ability of filtration
devices to remove particles from the airstream and measure their resistance to airflow
(ANSI/ASHRAE 2007). Here, efficiency refers to the ability of the filter to remove
particles from the airstream, represented as one minus the ratio of particle concentrations
measured downstream of the filter to that measured upstream of the filter. Filter testing is
conducted at flow rates between 472 CFM (0.22 m
3
/s) and 3,000 CFM (1.4 m
3
/s) in
Standard 52.2 laboratory tests. The test procedure for device efficiency uses laboratory-
generated potassium chloride particles and synthetic dust to simulate field conditions.
There are two types of pressure within duct systems: static and velocity pressure.
Static pressure is often simply called pressure. Every point in a fluid will have a static
pressure, which is the force per unit area at that point. The force is equal in all directions.
Velocity pressure is often called dynamic pressure. Velocity pressure is the pressure from
a moving fluid when it makes a direct hit on an object, such as a register at the end of a
9
duct. Pressure is measured in inches of water column (IWC) or Pascals (Pa). Total system
pressure is the sum of the static and velocity pressure at a given point in the HVAC
system. Pressure drop refers to the pressure difference across the filter or any other
component within the HVAC systems air stream (AC evaporator coil, humidifier, etc.).
Efficiency measurements for mechanical and pleated filters are performed at
several intervals during a simulated dust-loading procedure in Standard 52.2 to establish
a curve of efficiency as a function of dust loading. Measurements are taken at the
following points:
a. Before any dust is fed to the device;
b. After an initial conditioning step with a dust loading of 30 g or an increase in
10 Pa (0.04 in. of water) pressure drop across the device, whichever comes
first;
c. After the dust-loading increments have achieved an airflow resistance increase
of one-quarter, one-half, and three-quarters of the difference between the
beginning and the prescribed end point limit of airflow resistance; and
d. After the dust increment that loads the device to its prescribed end point
resistance limit (ANSI/ASHRAE 2007).

The result of this test is the filters Minimum Efficiency Reporting Value
(MERV) rating, which is a measure of efficiency of the filter at removing particles at
various size bins from 0.3 to 10 m. This efficiency-testing standard is also somewhat
helpful for defining when a filter is dirty. The device is assumed to be clean when the
filter resistance is equal to the initial resistance value in Table 12.1 (ANSI/ASHRAE
2007). A very dirty (fully loaded) filter has resistance equal to the final resistance value,
although translation to actual operational environments is difficult. A dirty filter has a
resistance value in the middle of the two extremes. Beyond these inferences, there is little
literature on what constitutes a dirty filter.
10
Low-efficiency filters are typically defined as having a MERV rating of under 4.
Medium-efficiency filters have MERV ratings of 5-10 and high-efficiency filters are
MERV 11 and above. Most residential HVAC filters are currently sold at 1-inch depths,
although higher efficiency filters (e.g., MERV 13+) are now being sold at 2-inch and 5-
inch depths, which have the benefit of providing more filter surface area for particle
collection while minimizing impacts on initial filter pressure drop.

Coil Fouling
Filters are used to protect HVAC equipment from damage and coil fouling and to
improve indoor environmental quality for occupants. Coil fouling effects cooling
capacity in two ways: by reducing the heat transfer coefficient and reducing airflow.
Siegel et al.

(2002) applied experimental and simulation results to estimate the impact of
coil fouling on AC system performance and capacity. They found that typical coils foul
enough within 7.5 years to double evaporator pressure drop. When a typical residential
system coil was fouled, the pressure drop increased by about 40%, the airflow reduced 5-
10%, and the efficiency and capacity of the AC decreased 2-4%. Although this is a
relatively minor decrease in efficiency and capacity, this is based on assumed correct
airflow. Residential systems often have low airflow and performance impacts can be
greater because air conditioner capacity is more sensitive to changes in low airflows
(Parker et al 1997).
Yang et al. (2004) studied the impact of different filter types on the performance
of packaged air conditioners under both clean and fouled conditions. They evaluated
three packaged systems: one 35-ton, one 5-ton, and one 3-ton unit. The units represent
typical systems for a medium to large commercial building, small commercial building,
and a small commercial or residential building, respectively. They introduced a set
amount of particles to foul the coils over time and measured the percent passing through
11
the filter. Field measurements as well as manufacturers data was collected and integrated
into a system simulation model, ACMODEL.
Somewhat surprisingly, Yang et al (2004) found that the heat transfer coefficient
could actually increase with limited fouling and that under extreme fouling the coefficient
reduction was minor. Capacity reduction was primarily the result of airflow reduction.
After introducing the equivalent of one years worth of dust loading (600 grams of dust),
the degradation in cooling capacity from coil fouling was relatively minor: 2-4% for the
35-ton system, 2-3% for the second 35-ton system, 5-7% for the 5-ton system, and 4-5%
for the 3-ton system. Depending on the AC unit size, the EER was reduced by 2-10%
because of fouling.

Pressures within HVAC Systems
For any fixed system, the relationship between airflow and system pressure
follows a quadratic relationship, presented in Equation 1 below.

P = k * Q
2

(1)
where
P = system pressure (IWC or Pa)
K = constant based on measured airflow and pressure
Q = system flow (CFM or m
3
/s)

Airflow within an HVAC system may be measured with a Pitot-tube traverse,
flow hood, flow grid, or an anemometer. Static pressure is measured with a static
pressure probe (simple Pitot-tube) and a manometer.

12
The static pressure drop across individual components of the HVAC system is
proportional to the square of the airflow rate and a coefficient of proportionality, C. The
coefficient C is constant for static system components and is based on the components
geometry. A filters coefficient changes over time as loading changes its geometry. The
evaporators coefficient will also change as moisture condenses and fouling occurs over
time. A systems total coefficient is the sum of all components connected in series with
duct branches connected in parallel (Equation 2) (Stephens et al 2010).

(2)

where
C
total
= coefficient of proportionality for the entire system
C
return
= coefficient of proportionality for the return duct
C
filter
= coefficient of proportionality for the filter
C
CC
= coefficient of proportionality for the cooling coil
C
HC
= coefficient of proportionality for the heating coil
C
S1
= coefficient of proportionality for the first supply duct branch
C
S2
= coefficient of proportionality for the second supply duct branch
C
S3
= coefficient of proportionality for the third supply duct branch

The power draw of an AHU fan can be described as a function of the required
pressure increase across the fan and the airflow rate, as shown in Equation 3 (Stephens et
al 2010).

(3)
Ctotal = Creturn + Cfilter +CCC + CHC + CS1 +
CS 2 + CS 3
CS 2 CS 3
13

where
W
fan
= power draw of fan
P
fan
= required pressure increase across the fan
Q
fan
=system airflow rate
fan
= efficiency of the fan
motor
= efficiency of the fan motor
C = coefficient of proportionality

In an HVAC system where the airflow rate remains constant through the use of an
ECM and the motors efficiency does not change with airflow rate, the power draw of the
fan is a linear function of the filter coefficient.
2
In this scenario, a 5% increase in total
pressure drop would cause a 5% increase in fan electric power draw. Most residential
HVAC fans use PSC motors, however, which do not adjust rotational speed to maintain
constant airflow rates. Thus, an increase of filter pressure drop will generally decrease the
airflow rate and also decrease the fan power draw for a PSC motor (Siegel et al 2007,
Stephens et al 2010).

Figure 3 represents a theoretical HVAC system without airflow controls
(Stephens et al 2010). System and fan curves are used to characterize HVAC system
performance based on pressure and airflow. The graph illustrates the relationship between
pressure and airflow. In this example, increasing the filter pressure drop by replacing the
mid-MERV filter with a high-MERV filter increases the total system pressure and
decreases the airflow rate (moving the working point from A to B). The reverse effect

2
The efficiency of many ECM fans, however, changes as a function of rotational speed.
14
occurs when decreasing the filter efficiency or pressure drop (moving from point A to
point C). In terms of the impacts on pressure and airflow, higher-efficiency filters and
loaded filters are conceptual equivalents in that they both describe ways to increase filter
pressure drop.

Figure 3. System and fan curves for medium-, high-, and low-pressure-drop filters
(Stephens et al 2010).

However, calculating the effect of filter pressure drop on air conditioner capacity,
efficiency, and overall power draw is far from straightforward. The intersection point
between the fan curve and the system curve determines airflow through an air
conditioner. The fan and its installation determine the fan curve. The flow resistance of
all the components throughout the system, including return duct, filter, coil, and supply
duct, determines the system curve. Increasing the pressure drop of the filter will have a
different effect on the system curve depending on the flow resistance of the rest of the
15
system. To further complicate matters, residential fan curves have different slopes at
different points, which means that changing the filter pressure drop of a system operating
at one point in the curve will have a different effect than changing the pressure at another
point on the same curve.
All filters restrict some amount of airflow. This restriction is quantified by
calculating the difference in pressure immediately upstream and downstream the filter,
which is known as pressure drop. One recent study in California measured filter pressure
drop in 34 split air conditioners with furnaces (Proctor 2012). They found that the typical
replacement filter had 0.282 in. w.c. of pressure drop. The pressure drops ranged from
0.075 to .792 in. w.c. (approximately 20 to 200 Pa). These pressure drops are higher than
expected and may negatively impact HVAC system performance. The high pressure
drops reduce evaporate airflow, which lowers the total EER and the sensible EER of the
machine (Proctor 2012).

Airflow within HVAC Systems
There are three common methods of measuring airflow across the evaporator coil:
temperature split, balometer (airflow capture hood), and TrueFlow plates or an
equivalent flow grid. The temperature split method was originally promoted by Carrier
Corporation and allows for a quick check to establish if airflow is likely within an
acceptable range (Carrier Corporation 1994). Airflow is qualitatively assessed based on
the dry-bulb temperature drop across the evaporator coil and the return plenum wet-bulb
temperature. First, the actual temperature split is calculated (return air dry-bulb
temperature minus the supply air dry-bulb temperature). Second, using the Carrier
Corporations table, identify the target temperature split using the return air wet-bulb
temperature and return air dry-bulb temperature. And finally, calculate the difference
between the target and actual temperature split (actual temperature split-target
temperature split). A difference of 3F is deemed probably acceptable while outside
16
that range requires further investigation.
Balometers, such as the Alnor Flow Hood and The Energy Conservatorys
FlowBlaster Capture Hood Accessory for the Duct Blaster, measure volumetric
airflow from diffusers, grilles and registers (Figure 4). They consist of a capture device
(hood) that sits over the HVAC register to direct airflow to the fan. The fan speed is
regulated based on the airflow coming out of the register.

Figure 4. The Alnor EBT Balometer. Image from http://www.alnor-usa.com

The TrueFlow plate is an example of a Pitot array or flow grid (The Energy
Conservatory 2006). The plate is installed in the filter slot at the air handler or large
return register. It uses multiple Pitot tubes to calculate an average velocity sampled over
large area. With the HVAC system running, it provides a measurement of total system
airflow. One advantage to the TrueFlow device is that it can provide airflow rates at the
air handler when the filter is installed at the air handler.

17

Figure 5. TrueFlow Air Handler Flow Meter (The Energy Conservatory, 2006).

Airflow through an HVAC system is limited to the capacity of the air handler fan.
For example, an air handler fan rated for 800 CFM simply cannot move any additional
air. The airflow through the system may be reduced due to undersized ducts, dirty filters,
and dirty evaporator coil. Downey and Proctor (2002) evaluated airflow testing from
13,258 HVAC systems and found that 21% of the residential and commercial systems
experienced low air flow across the evaporator coil. Low airflow was usually the result of
dirty filters, fouled coils, dirty blower wheels, or incorrect blower speed settings.
Reduced airflow impacts an air conditioners ability to cool and dehumidify air. The air
conditioners total cooling capacity consists of the ability to remove latent and sensible
heat. Reduced airflow reduces sensible heat transfer to the evaporator coil, which reduces
sensible cooling capacity. This leads to cooler coil surface temperatures, which increases
moisture removal. The reduced airflow allows longer contact time between the air and
cooler coil, which improves dehumidification and latent capacity. These interactive
effects result in a reduction of cooling capacity that is not directly proportional to the
change in airflow, although information is lacking on in-situ measurements of these
impacts in real residential environments.
18
Filters Impact on Energy Consumption
In large commercial and high-efficiency residential systems the fan and motor
controls typically maintain the required airflow rates regardless of pressure drop. These
ECMs will adjust airflow rates to compensate for changes in pressure drop. For this
reason, a greater pressure drop will generally lead to increased energy consumption
(Chimack and Sellers 2000, Fisk et al 2002).
The Chimack and Sellers (2000) study was conducted in an office building in
Hoffman Estates, Illinois with two nearly identical variable air volume (VAV) fan
systems to determine if premium air filters are financially sound investments for building
owners. Premium filters need to be replaced less, reducing maintenance costs, and are
less restrictive to airflow. One VAV fan system, the control group, operated with
standard bag-type air filters, while the second VAV operated with premium filters. Bag
filters use dry media that is arranged in a long stocking shape to extend their surface area
or to allow recovery of the collected material. Although bag filters are commonly used in
commercial HVAC systems, many are being replaced with rigid dry filters. Power draw
of the supply fan motors was monitored for 40 weeks. Static pressure drops across the
filters were routinely measured, although it was impossible to isolate the pre-filter and
final filters, so only a total filter system drop was recorded.
The total power draw of the system with high-MERV filters was approximately
21 percent less than the control system. The relationship of filter loading to time was
nearly linear with the prefilter capturing the majority of particles. The authors do not
provide the static pressure data nor did they measure airflow rates. A payback of
upgrading filters was calculated at 10-months to 2 years.
Nassif (2012) explored the impact of air filtration on energy consumption for a
typical air-conditioning system with a constant- or variable-speed fan.

HVAC
performance was modeled in eQuest, a software package that uses DOE 2.1-E hourly
building energy simulation engine, combined with a direct expansion AC (ACDX)
19
computer model (Brandemuehl and Andersen 1993). eQuest was used to evaluate the
annual energy consumption for the AC systems under a range of flow and capacity
conditions. ACDX was used to model the effect of reduced airflow rate on cooling
performance. A typical storefront building was modeled in each of Greensboro (NC),
Orlando (FL), New York (NY), and San Francisco (CA). For constant-speed AC systems,
as the filter gets dirty, the static pressure increases and the airflow rate drops. The total
system cooling capacity and sensible cooling capacity both drop with lower airflow,
whereas the latent cooling capacity increases. For example, the ACDX model simulated
a drop to 80% of designed airflow causing a reduction in the total system capacity to
96.1% and the sensible capacity to 92.1% of design value. The annual cooling energy use
was predicted to increase between 50 and 70 kWh per ton and the fan energy use
increased 60-80 kWh per ton in these commercial simulations.
For variable-speed AC systems, as dirty filters increase resistance, the fan
increases its speed to maintain a constant airflow and meet the sensible load requirements
of the building. Since there is no change in airflow rate there is little to no direct change
in cooling and heating energy use. Increased fan use generates more waste heat, which
slightly increases the cooling energy use and reduces heating energy use. The increase in
fan energy use ranged 15-20 kWh per ton (Nassif 2012).
Stephens et al. (2010) explored the theoretical, as well as measured, energy
implications of higher-pressure drop filters. The study analyzed the energy consumption
of filters in two air conditioning systems in a test house in Austin, Texas over the course
of four months. The studys results are summarized in Table 1.

20
Table 1 Results from Stephens et al. (2010) when moving from a low-MERV filter
Low-MERV
Replacement

Airflow
Change
Fan Power
Draw
Outdoor Unit
Power Draw
Total Cooling
Capacity
High-MERV - 7% and 11% + 3-4% - <1% -3-4%
Mid-MERV - 3% and 8% +0-2% - <1% and +<1% -1-2% and -6-7%

Compared to low-MERV filters, high-MERV filters decreased airflow rates by
approximately 7% and 11% and mid-MERV filters decreased airflow rates by
approximately 3% and 8% in each of the two systems. With high-MERV filters, the fan
power draw actually increased approximately 34% and the power draw of the outdoor
unit decreased less than 1%. With mid-MERV filters, the fan power draw increased
between 0%2% and power draw of the outdoor unit decreased less than 1% in one
system and increased less than 1% in the other. Total capacity and system efficiency
decreased approximately 3%4% in the presence of high-MERV filters and 1%2% in
one system and approximately 6%7% in the other system in the presence of mid-MERV
filters. The daily energy consumption did not significantly differ between low- and high-
MERV filter installations (Stephens et al 2010).
Nassif

(2012) and Stephens et al (2010) acknowledge that elevated pressures
within the ductwork will increase duct leakage, but no known studies currently explore
this relationship or account for this in energy calculations.
Palani et al (1992) used a standard split system air conditioner with a short-tube
orifice expansion device installed within a laboratory (bench test) to evaluate the
impact of reduced evaporator airflow on system performance. Reduction in evaporator
airflow was simulated using a plywood restriction board to cover the supply air duct. The
plywood board was pre-drilled at several places to allow airflow from 100 CFM to 1000
CFM. The study found that at 90% reduced evaporator air flow rate, the total power
consumption decreases by 17% and the EER decreases by 71%. Reduced airflow also
resulted in increased moisture removal and a reduced SHR.
21
Rodriguez et al (1996) also looked at the impact of reduced evaporator airflow on
HVAC system performance. They tested two residential sized air conditioners in a
psychrometric room. They reduced evaporator airflows between 0 and 50% below
manufacturers recommendations. One unit used a TXV and the other a short tube orifice.
For the TXV system, capacity and coefficient of performance (COP) actually improved
with a 10% reduction in airflow, but then decreased significantly at larger airflow
reductions. A 20% reduction in airflow resulted in approximately 1% drop in capacity,
2% reduction in COP, and 1% increase in power consumption. At that same reduced
airflow the orifice systems capacity reduced 10%, COP dropped 8%, and power
consumption increased by 1%. Capacity and COP also decreased as outdoor temperatures
increased. The impact was greater on the orifice controlled system.

22
CHAPTER 3
METHODOLOGY

This section details the process for: a) selecting test homes and air conditioning
equipment; b) performing measurements at different stages of simulated filtration
loading; and c) data analysis. Figure 6 below outlines the research process.

Figure 6 Research process workflow.

Research Design
As HVAC filters become loaded with dirt, dust, and other pollutants, the airflow
through the entire duct system becomes restricted, increasing the pressure drop across the
filter. This restriction of airflow causes cooling systems to run longer to provide adequate
cooling if they have PSC blowers. If they have ECM blowers, system runtime should not
change drastically for most loading conditions, although the fan will draw more power to
overcome the additional pressure drop. Therefore, any energy savings from filter
Review techncial
specification on
filters and HVAC
equipment
Assess potential
impact of filters on
HVAC performance
Design filter loading
simulation
methodology
Develop process for
measuring filter
impacts in-situ
Test and record
data on 6 HVAC
systems
Analyze data using
regression analysis
Report
23
replacement depends on the extent of filter loading (i.e., filter pressure drop) and fan
blower type, either permanent split-capacitor (PSC) or electronically commutated motors
(ECM). This study documents the impact of simulated filter loading in-situ on a small,
but generally representative, sample of Atlanta area homes.

Test House Selection
This pilot study assessed homes within the Atlanta metro-region. Since the 1970s,
the Atlanta region has experienced remarkable growth with a sustained influx of new
residents. According to the Atlanta Regional Commission (ARC) nearly 80% of all
homes have been built since 1970 (ARC unknown). Although the exact number of homes
with central air conditioning in the metro-region is unknown, for decades the majority of
new homes have been built with AC (U.S. Census Bureau 2009). Figure 7 below shows
the growth of air conditioning in the south.

Figure 7. Percent of new homes containing central air conditioners in the south region
(U.S. Census Bureau 2009).

Four homes within the Atlanta metro-region were selected for testing. The homes
all contain central forced-air AC or air source heat pump systems located in
24
unconditioned spaces such as basements, attics, or crawlspaces. HVAC systems in the
Southeast US are commonly located within unconditioned spaces. The selected homes
ranged in age from nearly 90 years old to approximately 65 years old, and the HVAC
systems were between 1 and10 years old. The HVAC systems ranged from 2 to 4 tons.
They include Rheem, Trane, Ruud, and Carrier brand equipment. Lance Beaton with VIS
VIVA Energy Consulting, an Atlanta based home performance and HVAC contracting
company, confirmed that these brands represent some of the most common HVAC
systems sold in the Atlanta area (Beaton 2013). Thus the HVAC systems were
considered representative of those installed in Atlanta homes over the past 10 years.
A total of 5 AC systems were assessed for this initial phase of testing. Four homes
had one system each and one home had two systems. Tested systems were selected to
include both horizontal and vertical configurations. All systems had filters located at the
air handling unit (AHU). A range of filter sizes was included. The testing methodology
compares the HVAC system as-is to induced, or simulated, loading conditions. The as-
is conditions serve as the baseline to measure the impact of filter loading. Because our
testing included only the HVAC system itself and not its ability to cool any particular
space, housing characteristics such as building structure, thermal envelope, shading from
sunlight did not impact the testing methodology. Table 2 summarizes the test system
characteristics.

25
Table 2. Test system characteristics
Test System Home Characteristics HVAC Characteristics
Year Built Size (ft
2
) Year Manufactured Size (tons) Location
1 1950s 1,700 2009 2.5 Vented Crawlspace
2 1939 2,014 2012 2.5 Unconditioned Basement
3 1939 2,014 2012 2 Attic
4 1927 1,750 Approx. 2003 3.5 Unconditioned Basement
5 (5- Right) 1950s 1,850 2007 4 Unconditioned Basement
6 (5-Left) 1950s 1,850 2007 4 Unconditioned Basement

Filter Loading Simulation
This study simulated filter loading by installing the TrueFlow airflow metering
device and partially taping off the face at 3 different increments, similar to the procedure
in Palani et al. (1992). This resulted in measurements at five discrete static pressures and
simulated loading conditions: no filter, TrueFlow measurement, TrueFlow Taped #1,
TrueFlow Taped #2, and TrueFlow Taped #3. Examples of the TrueFlow taped for the
three consecutive measurements are below (Figures 8 - 10). The exact amount and
placement of the tape varied between homes to take into account the measured
performance of individual HVAC systems. The TrueFlow was not used to measure
airflow once taped.

26

Figure 8 An example of a partially taped filter (TrueFlow Taped 1)

Figure 9. An example of a partially taped filter (TrueFlow Taped 2).

27

Figure 10. An example of a partially taped filter (TrueFlow Taped 3).

Data Collection
The following information about each HVAC system was collected and
measured: AC indoor and outdoor coil make and model numbers, furnace (if present)
make and model number, and presence of mechanical ventilation. Testing was performed
during periods when the outdoor temperature was above 60F to allow for the safe
operation of air conditioners.
Measurements were taken after the HVAC system reached approximately steady
state. Steady State refers to the point at which the system is running at peak efficiency
and the evaporator coil is fully cooled. Since AC units cycle on and off as they maintain
their desired temperature, it is important to properly identify steady state. Steady state
was identified based on a constant difference of air characteristics (pressure, temperature,
and relative humidity) across the coil. Palani et al (1992) found that the test bench
28
reached steady state within 4 minutes after switching on the unit. Measurements in this
work were taken after at least 10 minutes of operation, once steady state was achieved.
Measurements were recorded with the filter removed, TrueFlow installed, and the
TrueFlow partially taped at three different intervals.
Three HVAC system configurations were tested. Figure 11 below shows a vertical
HVAC system with one central filter. Figure 12 is a horizontal HVAC system with one
central filter.

Figure 11. Vertical HVAC system with one central filter. Image courtesy Terrel Broiles.

29

Figure 12. Horizontal HVAC system with one central filter. Image courtesy Terrel
Broiles.

The Figure 13 below shows the vertical HVAC system with two return plenums
each with a central filter. The return plenums were tested independently. This figure
shows the set up for testing the left side only.

30

Figure 13. Vertical HVAC system with two central filters. Each side of the system was
tested independently. This shows the configuration to test the left side. Image courtesy
Terrel Broiles.

Fan Curve Determination Procedure
A series of measurements were first conducted to establish fan curves and to
better understand the relationship between airflow and filter pressure drop in the test
systems. Measurements were recorded within the supply plenum, return plenum, and
between the evaporator coil and filter. The measurements include static pressure before
and after the filter; static pressure before and after the evaporator coil; temperature before
and after the evaporator coil; and relative humidity before and after the evaporator coil.
Table 3 below summarizes the location and type of measurements recorded.
31
Table 3. Measurement type and location
Measurement Location Measurements Recorded
Return plenum Dry-bulb temperature (C), relative humidity,
and static pressure (Pa)
Supply plenum Dry-bulb temperature (C), relative humidity,
and static pressure (Pa)
Coil side of filter Static pressure (Pa)

The testing equipment is presented below in Table 4.
Table 4. Testing instrumentation
Measurement Units Equipment Accuracy
Pressure Pa (IWC) Energy Conservatory
DG-700
1% of reading or 0.15 Pa
(0.0006 IWC)
Temperature C (F) Fieldpiece ARH4 1F for readings 32F to
113F
Relative Humidity %RH Fieldpiece ARH4 2.5%@77F(25C), 10%
to 90% RH
Airflow m3/h (cfm) Energy Conservatory
TrueFlow Plate
7% of reading

The filter face was partially sealed with tape to restrict overflow, simulating filter
loading and higher MERV equipment. Airflow rates were calculated in the field using
Trueflow and the measured supply pressure. System measurements were required with
the filter removed, TrueFlow installed, and the TrueFlow partially sealed at 3 intervals to
achieve a total airflow reduction of 40%. Since there is no known research using tape to
simulate filter loading, this process was carefully evaluated and documented.
The TrueFlow Manual provides equations for estimating adjusted airflow
(Equation 4) based on measured supply plenum pressure (The Energy Conservatory
2006).

(4)

where
Q
reference
= volumetric flow rate of air with measuring device installed, m3/h (cfm)
32
Preference
= supply plenum pressure with measuring device installed, Pa (in. w.c.)
Poperating
= operating supply plenum pressure, Pa (in. w.c.)
Sensible, latent, and total capacity will be calculated for each of the 5 system
conditions based on the following equations used by Stephens et al. (2010). The sensible
capacity, q
sensible
(kW, kBtu/hr), is calculated using the Equation 5.

q
sensible
= Q
fan
(C
p
T) (5)
where
Q
fan
= volumetric flow rate of air flowing through the cooling coil, m
3
/s (ft
3
/hr)
= air density, assumed constant, 1.2 kg/m
3
(0.075 lbm/ft
3
)
C
p
= specific heat of air, assumed constant, 1.005 kJ/(kgK), (0.24 Btu/[lbmF])
T = temperature difference across the cooling coil, K (F)

The latent capacity, q
latent
(kW, kBtu/hr), is calculated using Equation 6.
q
latent
= Q
fan
(Wh
fg
) (6)
where
Q
fan
3
/s (ft
3
/hr)
= air density, assumed constant, 1.2 kg/m3 (0.075 lbm/ft
3
)
W = humidity ratio difference across the cooling coil, kg/kg (lbm/lbm)
h
fg
= latent heat of vaporization for water, assumed constant, 2257 kJ/kg (970 Btu/lb)

The total capacity, q
total
(kW, kBtu/hr), is the sum of the sensible capacity
(Equation 5) and latent capacity (Equation 6). Total capacity is expressed in Equation 7.

q
total
=Q
fan
(C
p
T +Wh
fg
) (7)

where
Q
fan
3
/s (ft
3
/h)
= air density, assumed constant, 1.2 kg/m
3
(0.075 lbm/ft
3
)
C = specific heat of air, assumed constant, 1.005 kJ/(kgK), (0.24 Btu/[lbmF])
T = temperature difference across the cooling coil, K (F)
W = humidity ratio difference across the cooling coil, kg/kg (lbm/lbm)
h
fg
= latent heat of vaporization for water, assumed constant, 2257 kJ/kg (970 Btu/lb)

33
Data Analysis
The data was analyzed to assess the relationship between filter pressure drop and
airflow rates; filter pressure drop and temperature differences across the coil (T); filter
pressure drop and absolute humidity differences across the coil (W); airflow rate and
temperature differences across the coil (T); and airflow rate and absolute humidity
differences across the coil (W). Finally, the relationship between airflow rate and
sensible, latent, and total capacities were also evaluated. Regression analysis was used to
determine the strength of each relationship. The results from simulated filter loading were
also compared to previous studies to estimate the impact of reduced airflow rates on air-
conditioner energy consumption.
Finally, because the nature of data collection in the field can result in periodic
improper measurements, a method of calculating an observations Z-score was used to
identify whether or not an observation was a statistical outlier in the dataset. A Z-Score is
a statistical measurement of an observations relationship to the mean in a group of
scores. The z-score calculation is below in Equation 8.

i
=

i

s

(8)

where
i
= sample value
= sample mean
s = standard deviation

A Z-score of 0 means the score is the same as the mean. The Z-score can also be
positive or negative, indicating whether it is above or below the mean and by how many
standard deviations. A Z-score of 3.5 means the observation is at least 3.5 standard
deviations from the mean, which serves as a common criterion for identifying outliers in
34
datasets like this one. The Z-score analysis was eventually used to remove one outlier
within the data set (see pages 44-55 for additional information).
35
CHAPTER 4
FIELD RESULTS & DATA ANALYSIS

Test System Descriptions
Field testing was performed in July, August, and September 2013. Ambient air
temperature ranged from approximately 64F to 86F during the test periods. Twenty
initial test systems were excluded from analysis for not meeting key criteria. Cause for
exclusion included the following: system contained a known electrically commutated
motor (ECM) blower; not enough room to install measurement equipment on either side
of the filter; and filter located at a central return register and not at the air handler. Four
out of the six chosen test systems were known permanent split capacitor (PSC) blowers;
information could not be found for test systems 5 and 6, but from the subsequent airflow
and pressure data suggest that they were likely ECM blowers or at least variable speed
blowers.

Test System Locations
Four houses within the Atlanta metro-region were selected for testing (Figure 14).
The homes all contain central forced-air AC and/or air source heat pump systems located
in unconditioned spaces (unconditioned basement, attic, or crawlspace). The houses
ranged in age from nearly 65 to approximately 90 years old, and the HVAC systems
ranged from 1 to 10 years old. The HVAC systems ranged from 2 to 4 tons. They
include Rheem, Trane, Ruud, and Carrier brand equipment, which represent some of the
most common HVAC systems sold in the Atlanta area (Beaton 2013). Thus the HVAC
systems were considered generally representative of those installed in Atlanta homes over
the past 10 years.

36

Figure 14 Location of the 4 test homes in the Atlanta metro-region.

A total of five AC systems were assessed for this initial phase of testing. Two of
the systems were located within the same home. HVAC systems included both vertical
and horizontal configurations. All systems had filters located at the air handler. A range
of common filter sizes was included, from 14 x 20 to 20 x 24.

Test System 1
Test system 1 is located in an unconditioned crawlspace. The house was built in
the 1950s and has an area of approximately 1,700 ft
2
. The 2.5-ton capacity system is
horizontally configured (Figure 15).

37

Figure 15 Test system 1 is located in an unconditioned crawlspace and is horizontally
configured. The return plenum is on the right and supply plenum to the left.

Test System 2
Test system 2 has a 2.5-ton capacity and is located in the unconditioned basement
of a home built in 1939 (Figure 16). The home has an area of 2,014 ft
2
. The system was
installed in 2012.
38

Figure 16 Test system 2 is vertically configured and located in an unconditioned
basement. The return plenum is on the bottom and supply plenum above.

Test System 3
Test system 3 has a 2-ton capacity and is located in the unconditioned attic of the
same home as test system 2. The system is configured horizontally and was installed in
2012.

Test System 4
Test system 4 has 3.5-ton capacity and is located in an unconditioned basement
(Figure 17). The home was built in 1927 and is approximately 1,750 ft
2
. The system was
installed in approximately 2003.

39

Figure 17 Test system 4 is vertically configured and located in an unconditioned
basement. The return plenum is on the left side and supply plenum above.

Test Systems 5 and 6
Test systems 5 and 6 are actually the same HVAC system. Figure 18 below shows
how the system has two return plenums. Each plenum has its own filter and was tested
separately. The system has 4-ton capacity, located in an unconditioned basement, and
was installed in 2007. The house is approximately 1,850 ft
2
and built in the 1950s.

40

Figure 18 Test systems 5 and 6. The photo shows the left side return plenum set up for
testing.

Analysis
The following sections describe measured impacts of (i) filter pressure drop on
system airflow rates, (ii) airflow rate reductions on temperature and absolute humidity
differences across the coils, and (iii) airflow rate reductions on sensible and latent cooling
capacity.

Filter Pressure Drop & System Airflow
The relationship between induced filter pressure drop and system airflow rates
was first evaluated in each system, as shown in Figure 19 and Table 5. The absolute value
of the pressure drop across the filter was compared to the reduction in airflow (as a
41
percentage of total system flow without filter installed) for each induced pressure drop
condition. The induced filter pressure drop ranged from 16 Pa to 231 Pa. At least 75 Pa
was achieved in each system, although systems 2 and 3 did not reach 100 Pa at their
maximum level of filter blockage (which was nearly 75% blocked). Airflow rates ranged
from 100% of total system airflow (i.e. airflow remained unchanged) to as low as 61%
(i.e., a 39% reduction in airflow at the largest filter pressure drop of 231 Pa). Overall, as
filter pressure drop increased, airflow rates generally decreased, particularly for the
known PSC blowers. The two systems that were apparently ECM blowers responded to
increased filter pressure drop by nearly maintaining airflow rates until reaching a
maximum pressure and rapidly decreasing in flow, which is consistent with other ECM
data (Murray 2012).

Figure 19 Relationship of induced filter pressure drop and system airflow rate measured
in the test systems.
50%
60%
70%
80%
90%
100%
110%
0 100 200 300
%

o
f

m
a
x

f
l
o
w
Filter Pressure Drop (Pa)
HVAC System 1
HVAC System 2
HVAC System 3
HVAC System 4
HVAC System 5-Right
HVAC System 5-Left
42
Similarly, Figure 20 shows the same airflow and filter pressure data measured at
each site and simulated loading condition, combined across all data points and fit with a
linear regression. Regression analysis shows an intercept near 1.02, a slope of -0.0014,
and a coefficient of determination (R
2
) of 0.71, suggesting that with reasonable certainty,
each additional 10 Pa in filter pressure drop resulted in a ~1.4% decrease in system
airflow rates in these systems.

Figure 20 System airflow rates normalized to no filter conditions versus filter pressure
drop.

Table 5 Filter pressure drop and system airflow
Filter pressure drop (Pa) Percent of no filter flow (%)
HVAC
System
Name
No
Filter
True
Flow
True
Flow
Taped
1
True
Flow
Taped
2
True
Flow
Taped
3
No
Filter
True
Flow
True
Flow
Taped
1
True
Flow
Taped
2
True
Flow
Taped
3
1 0 49.6 85.6 144.7 180.4 100% 93% 89% 77% 69%
2 0 27.8 47.3 58.3 76.3 100% 98% 96% 94% 97%
3 0 16.3 31.9 57.5 81.9 100% 100% 98% 95% 90%
4 0 51.7 125.3 181.2 230.8 100% 88% 82% 73% 61%
5 (Right) 0 125.2 127.5 132.2 141.9 100% 97% 95% 91% 83%
6 (5-Left) 0 36.0 57.7 86.4 115.8 100% 100% 99% 99% 99%

y = -0.0014x + 1.02
R = 0.71
0%
20%
40%
60%
80%
100%
120%
0 50 100 150 200 250
P
e
r
c
e
n
t

o
f

n
o

f
i
l
t
e
r

f
l
o
w
43
Test Systems 5 and 6 could not be verified as having an ECM blower, although
evidence suggests that they were at least variable speed blowers. These two systems
responded to increased filter pressure drop by nearly maintaining airflow rates until
reaching a maximum pressure and rapidly decreasing in flow. System 5 (left return
plenum) maintained a nearly identical airflow rates regardless of filter pressure drop, at
least until ~115 Pa. System 6 (right plenum of system 5) maintained airflow rates within
90% of the no filter flow until rapidly decreasing above 125 Pa.

Filter Pressure Drop & Temperature Difference Across the Coil
The relationship between induced filter pressure drop and the dry bulb
temperature difference across the evaporator coil (T) was evaluated in each system, as
shown in Table 6. The difference in temperature (T) before passing the evaporator coil
(before-filter) and after passing the coil (in the supply plenum) was measured. Values of
T at each flow and pressure drop condition were normalized to those measured with no
filter installed and compared to the absolute value of the pressure drop across the filter.
The change in T ranged from 152% of total system airflow (i.e. the difference in
temperature increased) to 97% (i.e., a 3% reduction in T).

Table 6 Filter pressure drop and coil T
Pfilt (Pa) Percent of no filter T (%)
HVAC
System Name
No
Filte
r
Tru
e
Flo
w
True
Flow
Taped 1
True
Flow
Taped 2
True
Flow
Taped 3
No
Filter
True
Flow
True
Flow
Taped 1
True
Flow
Taped 2
True
Flow
Taped 3
1 0 49.6 85.6 144.7 180.4 100% 94% 100% 108% 123%
2 0 27.8 47.3 58.3 76.3 100%
102
% 107% 107% 110%
3 0 16.3 31.9 57.5 81.9 100%
108
% 110% 113% 113%
4 0 51.7 125.3 181.2 230.8 100%
102
% 111% 152%* 138%
5 (5-Right) 0
125.
2 127.5 132.2 141.9 100% 97% 98% 98% 97%
6 (5-Left) 0 36.0 57.7 86.4 115.8 100%
107
% 109% 111% 114%
* Data point is statistically an outlier and removed from analysis.
44

Importantly, the maximum change in T (System 4 with 181.2 Pa filter pressure
drop) was identified visually to be a likely outlier in the dataset. To determine whether or
not this was the case, a Z-score was calculated for each of the 30 data points for the
relative change in T across all systems, as shown here in Equation 9:

i
=

i

s

(9)

where
i
= sample value
= sample mean
s = standard deviation of the sample

Z-scores for each data point are shown in Table 7 on the next page.

45
Table 7 Z-scores for each T measurement
Filter pressure
drop (Pa) Percent of no filter T (%) Z-score
0 100% -0.62
0 100% -0.62
0 100% -0.62
0 100% -0.62
0 100% -0.62
0 100% -0.62
49.6 94% -1.14
27.8 102% -0.47
16.3 108% -0.01
51.7 102% -0.42
125.2 97% -0.89
36.0 107% -0.09
85.6 100% -0.62
47.3 107% -0.03
31.9 110% 0.19
125.3 111% 0.28
127.5 98% -0.80
57.7 109% 0.09
144.7 108% 0.06
58.3 107% -0.03
57.5 113% 0.46
181.2 152% 3.60
132.2 98% -0.80
86.4 111% 0.27
180.4 123% 1.26
76.3 110% 0.19
81.9 113% 0.46
230.8 138% 2.49
141.9 97% -0.89
115.8 114% 0.54

The apparent outlier measured during the TrueFlow Taped 2 condition at System
4 was shown to have a Z-score of 3.6, suggesting it was indeed an outlier with a score
greater than commonly used criteria of Z = 3.5. Therefore, this data point is removed
from further analysis. The remaining maximum change in T is therefore 138% (at the
highest filter pressure drop measured). The exact cause for the outlying data is unknown,
but it was likely caused by experimental error in the field. No combination of airflow and
pressure measurements was removed, however, because there were no outliers with Z-
scores greater than 3.5.
46
Figure 21 shows the remaining relative T data versus absolute filter pressure
drop measured in each of the test systems.

Figure 21 Relative T across the coil versus filter pressure drop for each test system.

Similarly, Figure 22 shows the same T and filter pressure drop data measured at
each site and induced loading condition, combined across all data points and fitted with a
linear regression. Regression analysis shows an intercept near 1.00, a slope of +0.0008
with a coefficient of determination (R
2
) of 0.30, suggesting that with only moderate
certainty, each additional 10 Pa in filter pressure drop resulted in a ~0.8% increase in
system T in these systems. Similarly, if filter pressure drop reaches 200 Pa, coil T is
expected to increase by approximately 16% in these test systems, on average.

60%
70%
80%
90%
100%
110%
120%
130%
140%
150%
160%
0 50 100 150 200 250
%

o
f

n
o

f
i
l
t
e
r

T
HVAC System 1
HVAC System 2
HVAC System 3
HVAC System 4
HVAC System 5-Right
HVAC System 5-Left
47

Figure 22 Relationship of filter pressure drop and T across all systems.

These data suggest what other laboratory and field studies have shown: as airflow
rates are reduced in the presence of larger filter pressure drops, sensible capacity will not
decrease linearly with flow because the temperature difference across the coil increases
slightly and supply air is delivered at a lower temperature (Nassif 2012, Yang et al 2004).
This relationship is explored further in a subsequent section.

Filter Pressure Drop & Absolute Humidity
The relationship between induced filter pressure drop and changes in the
difference in absolute humidity across the coil (W) was evaluated in each system, as
shown in Table 7. The change in W across the coil (as a percentage of W without filter
installed) was compared to the absolute value of the pressure drop across the filter for
each induced pressure drop condition. Because W was calculated using the dry bulb
temperature and relative humidity measurements in each location, data from True Flow
Taped 2 in System 4 was again excluded as an outlier. The resulting changes in W
y = 0.0008x + 1.00
R = 0.29
60%
70%
80%
90%
100%
110%
120%
130%
140%
150%
0 50 100 150 200 250
P
e
r
c
e
n
t

o
f

n
o

f
i
l
t
e
r

T
48
ranged from a minimum of 41% (i.e., a 59% reduction in W) to a maximum of 100%
(i.e. W did not change). These data are shown in Figure 23 for each test system.

Table 8 Filter pressure drop and absolute humidity differences across the coil
Pfilt (Pa) Percent of no filter W (%)
HVAC
System
Name
No
Filter
True
Flow
True
Flow
Taped 1
True
Flow
Taped 2
True
Flow
Taped 3
No
Filter
True
Flow
True
Flow
Taped 1
True
Flow
Taped 2
True
Flow
Taped 3
1 0 49.6 85.6 144.7 180.4 100% 41% 48% 82% 98%
2 0 27.8 47.3 58.3 76.3 100% 82% 68% 77% 82%
3 0 16.3 31.9 57.5 81.9 100% 84% 81% 81% 71%
4 0 51.7 125.3 181.2 230.8 100% 86% 77% 132%* 100%
5 (Right) 0 125.2 127.5 132.2 141.9 100% 81% 86% 95% 90%
6 (5-Left) 0 36.0 57.7 86.4 115.8 100% 100% 100% 100% 100%
*This data point is statistically an outlier and removed from analysis.

Figure 23 Relationship of filter pressure drop and W for each test systems.
40%
50%
60%
70%
80%
90%
100%
110%
120%
130%
140%
0 50 100 150 200 250
%

o
f

n
o

f
i
l
t
e
r

W
HVAC System 1
HVAC System 2
HVAC System 3
HVAC System 4
HVAC System 5-Right
HVAC System 5-Left
49
Figure 24 shows the same W and filter pressure data measured at each site and
induced loading condition, combined across all data points and fitted with a linear
regression. Regression analysis shows a slope near zero and a coefficient of
determination (R
2
) of only 0.01, suggesting that there was no observed correlation
between filter pressure drop and W.

Figure 24 Relationship of filter pressure drop and W for all systems.

Airflow and Temperature Differences Across the Coil
Because airflow rates and pressure drops were not necessarily directly correlated,
this section explores the relationship of airflow rate (instead of pressure drop) and the dry
bulb temperature difference across the evaporator coil (T) in each system, as shown in
Table 9 and Figure 25. The difference in temperature (T) before passing the evaporator
coil (pre-filter) and after passing the coil (in supply plenum) was measured. Values of
(T) at each flow and pressure drop condition were normalized to those measured with
no filter installed and compared to the absolute value of the pressure drop across the
filter. The change in T ranged from 152% of total system airflow (i.e. the difference in
y = 0.0003x + 0.86
R = 0.01
0%
20%
40%
60%
80%
100%
120%
140%
0 50 100 150 200 250
P
e
r
c
e
n
t

o
f

n
o

f
i
l
t
e
r

W
50
temperature increased) to 97% (i.e., a 3% reduction in T). The reduced airflow ranged
from 100% of total system airflow (i.e. airflow remained unchanged) to 61%.

Table 9 Fractional airflow and changes in T across the evaporator coil
Percent of no filter flow (%) Percent of no filter T (%)
HVAC
System
Name
No
Filter
True
Flow
True
Flow
Taped 1
True
Flow
Taped 2
True
Flow
Taped 3
No
Filter
True
Flow
True
Flow
Taped 1
True
Flow
Taped 2
True
Flow
Taped 3
1 100% 93% 89% 77% 69% 100% 94% 100% 108% 123%
2 100% 98% 96% 94% 97% 100% 102% 107% 107% 110%
3 100% 100% 98% 95% 90% 100% 108% 110% 113% 113%
4 100% 88% 82% 73% 61% 100% 102% 111% 152%* 138%
5 (Right) 100% 97% 95% 91% 83% 100% 97% 98% 98% 97%
6 (5-Left) 100% 100% 99% 99% 99% 100% 107% 109% 111% 114%

Figure 25 Relationship of airflow and T for each system.

80%
90%
100%
110%
120%
130%
140%
50% 60% 70% 80% 90% 100% 110%
P
e
r
c
e
n
t

o
f

n
o

f
i
l
t
e
r

T
Percent of no filter flow
HVAC System 1
HVAC System 2
HVAC System 3
HVAC System 4
HVAC System 5-Right
HVAC System 5-Left
51
Similarly, Figure 26 shows the same T and system airflow data measured at each
site and induced loading condition, combined across all data points and fit with a linear
regression. Regression analysis shows an intercept near 0.80, a slope of 1.81, and a
coefficient of determination (R
2
) of 0.44, suggesting that with moderate certainty, each
additional 10% decrease in system airflow resulted in a ~9.8% increase in system T in
these systems.

Figure 26 Relationship of airflow and T for all systems.

rates are reduced in the presence of larger filter pressure drops, temperature difference
across the coil increases slightly and supply air is delivered at a lower temperature
(Nassif 2012, Yang et al 2004). This data also shows that there is a stronger relationship
between T and airflow as opposed to T and induced filter pressure drop (R
2
= 0.44 vs.
0.29). This is plausible because induced filter pressure drop affects airflow differently in
each system, and T is directly impacted by airflow (and not necessarily pressure drop).
y = -0.7982x + 1.8115
R = 0.44
40%
60%
80%
100%
120%
140%
160%
50% 60% 70% 80% 90% 100% 110%
P
e
r
c
e
n
t

o
f

n
o

f
i
l
t
e
r

T

52
Airflow and Absolute Humidity Differences Across the Coil
The relationship between system airflow and changes in the difference in absolute
humidity across the coil (W) was evaluated in each system, as shown in Table 10.
Because W was calculated using the dry bulb temperature and relative humidity
measurements in each location, data from True Flow Taped 2 in System 4 was again
excluded as an outlier. The resulting changes in W ranged from a minimum of 41% (i.e.
a 59% reduction in W) to a maximum of 100% (i.e. W did not change). The reduced
airflow ranged from 100% of total system airflow (i.e. airflow remained unchanged) to
61% (i.e. a 39% reduction in airflow). These data are shown in Figure 27 for each test
system.

Table 10 System airflow and W across the evaporator coil

Percent of no filter W (%) Percent of no filter flow (%)
HVAC
System
Name
No
Filter
True
Flow
True
Flow
Taped 1
True
Flow
Taped 2
True
Flow
Taped 3
No
Filter
True
Flow
True
Flow
Taped 1
True
Flow
Taped 2
True
Flow
Taped 3
1 100% 41% 48% 82% 98% 100% 93% 89% 77% 69%
2 100% 82% 68% 77% 82% 100% 98% 96% 94% 97%
3 100% 84% 81% 81% 71% 100% 100% 98% 95% 90%
4 100% 86% 77% 132% 100% 100% 88% 82% 73% 61%
5 (Right) 100% 81% 86% 95% 90% 100% 97% 95% 91% 83%
6 (5-Left) 100% 100% 100% 100% 100% 100% 100% 99% 99% 99%
*This data point is statistically an outlier and removed from analysis.

53

Figure 27 The relationship of airflow and absolute humidity for each system.

Similarly, Figure 28 shows the same W and system airflow data measured at
each site and simulated loading condition, combined across all data points and fit with a
linear regression. Regression analysis shows an intercept near 0.98, a slope of 0.07, and a
2
) of 0.02, suggesting no correlation between system
airflow and W.

50%
60%
70%
80%
90%
100%
110%
0% 50% 100% 150%
%

o
f

n
o

f
i
l
t
e
r

W
Percent of no filter air flow
HVAC System 1
HVAC System 2
HVAC System 3
HVAC System 4
HVAC System 5-Right
HVAC System 5-Left
54

Figure 28 Relationship of airflow and absolute humidity for all systems.

Airflow and System Capacity
The change in sensible, latent, and total capacities (as a percent of system
capacity with no filter installed) was compared to the reduction in airflow (as a percent of
total system flow without filter installed), as shown in Table 11. The reduced airflow
ranged from 100% of total system airflow (i.e. airflow remained unchanged) to 61% (i.e.
the airflow decreased by 39%). The change in sensible capacity ranged from 108% to
80%.

y = -0.0722x + 0.9815
R = 0.02
40%
50%
60%
70%
80%
90%
100%
110%
120%
130%
140%
60% 65% 70% 75% 80% 85% 90% 95% 100% 105%
P
e
r
c
e
n
t

n
o

f
i
l
t
e
r

W

55
Table 11 Relationship of airflow and sensible capacity for all systems.
Percent of no filter flow (%) Percent no Filter Sensible Capacity (kBTU/hr)
HVAC System
Name
No
Filter
True
Flow
True
Flow
Taped 1
True
Flow
Taped 2
True
Flow
Taped 3
No
Filter
True
Flow
True
Flow
Taped 1
True
Flow
Taped 2
True
Flow
Taped 3
1 100% 93% 89% 77% 69% 100% 87% 89% 84% 86%
2 100% 98% 96% 94% 97% 100% 99% 103% 100% 106%
3 100% 100% 98% 95% 90% 100% 107% 107% 108% 102%
4 100% 88% 82% 73% 61% 100% 90% 92% 111% 85%
5-Right 100% 97% 95% 91% 83% 100% 94% 93% 89% 80%
5-Left 100% 100% 99% 99% 99% 100% 106% 108% 110% 113%

Similarly, Figure 29 shows the same sensible capacity and system airflow data
measured at each site and simulated loading condition, combined across all data points
and fit with a linear regression. Regression analysis shows an intercept near 0.36, a slope
of 0.67, and a coefficient of determination (R
2
) of 0.53, suggesting that with moderate
certainty, each 10% reduction in airflow resulted in a ~7% decrease in sensible capacity
in these systems. If one was concerned with keeping sensible capacity at least 80% of the
no filter sensible capacity, these data suggest that airflow rates should not be able to
decrease more than approximately 30%.

Figure 29 Relationship of airflow and sensible capacity for all systems.
y = 0.6678x + 0.3595
R = 0.53
60%
70%
80%
90%
100%
110%
120%
60% 70% 80% 90% 100% 110%
S
e
n
s
i
b
l
e

C
a
p
a
c
i
t
y

(
k
B
t
u
/
h
r
)
Percent of no filter airflow
Percent no Filter Sensible
Capacity (kBTU/hr)
Linear (Percent no Filter
Sensible Capacity
(kBTU/hr))
56
The change in latent capacity ranged from 100% to 38% (Table 12). Figure 30
shows the same latent capacity and system airflow data measured at each site and
simulated loading condition, combined across all data points and fit with a linear
2
) of 0.23, suggesting that with relatively low certainty, a
10% decrease in airflow resulted in a ~10.4% decrease in system latent capacity in these
systems.

Table 12 Relationship of airflow and latent capacity for all systems.
Percent of no filter flow (%) Percent no Filter Latent Capacity (kBTU/hr)
HVAC System
Name
No
Filter
True
Flow
True
Flow
Taped
1
True
Flow
Taped
2
True
Flow
Taped
3
No
Filter
True
Flow
True
Flow
Taped
1
True
Flow
Taped
2
True
Flow
Taped
3
1 100% 93% 89% 77% 69% 100% 38% 43% 63% 68%
2 100% 98% 96% 94% 97% 100% 80% 65% 72% 79%
3 100% 100% 98% 95% 90% 100% 84% 79% 77% 64%
4 100% 88% 82% 73% 61% 100% 76% 64% 97% 61%
5-Right 100% 97% 95% 91% 83% 100% 79% 82% 86% 75%
5-Left 100% 100% 99% 99% 99% 100% 100% 99% 99% 99%

Figure 30 Relationship of airflow and latent capacity for all systems.

y = 0.8355x + 0.0396
R = 0.23
40%
50%
60%
70%
80%
90%
100%
110%
40% 60% 80% 100% 120%
L
a
t
e
n
t

C
a
p
a
c
i
t
y

(
k
B
t
u
/
h
r
)
Percent no Filter Latent
Capacity (kBTU/hr)
Latent Capacity (kBTU/hr))
57
Finally, the change in total capacity ranged from 108% to 69% (Table 13). Figure
31 shows the same latent capacity and system airflow data measured at each site and
simulated loading condition, combined across all data points and fit with a linear
2
) of 0.36, suggesting that with moderate certainty, each
additional 10% reduction in airflow resulted in a ~6.7% decrease in total cooling capacity
in these systems.

Table 13 Relationship of airflow and total capacity for all systems.
Percent of no filter flow (%) Percent no Filter Total Capacity (kBTU/hr)
HVAC System
Name
No
Filter
True
Flow
True
Flow
Taped
1
True
Flow
Taped
2
True
Flow
Taped
3
No
Filter
True
Flow
True
Flow
Taped
1
True
Flow
Taped
2
True
Flow
Taped
3
1 100% 93% 89% 77% 69% 100% 69% 72% 76% 79%
2 100% 98% 96% 94% 97% 100% 93% 91% 92% 98%
3 100% 100% 98% 95% 90% 100% 98% 97% 97% 88%
4 100% 88% 82% 73% 61% 100% 85% 81% 106% 76%
5-Right 100% 97% 95% 91% 83% 100% 89% 89% 88% 78%
5-Left 100% 100% 99% 99% 99% 100% 104% 105% 106% 108%

Figure 31 Relationship of airflow and total capacity for all systems.
y = 0.6409x + 0.3321
R = 0.36
50%
60%
70%
80%
90%
100%
110%
120%
50% 60% 70% 80% 90% 100% 110%
T
o
t
a
l

C
a
p
a
c
i
t
y

(
k
B
t
u
/
h
r
)
Percent no Filter Total
Capacity (kBTU/hr)
Total Capacity (kBTU/hr))
58
rates are reduced in the presence of larger filter pressure drops, total cooling capacity is
reduced. Rodriguez et al. (1996) showed that a 10% reduction in airflow caused capacity
to decrease by approximately 7% in a laboratory setting. Stephens et al (2010) found that
decreasing airflow rates by approximately 7% results in only a 3-4% reduction in total
capacity in field measurements in a controlled test house setting. These are some of the
first data to systematically report changes in capacity in response to simulated filter
loading conditions in real residences.
In other words, as airflow decreased so did sensible, latent, and total capacity,
although these relationships were not linear. Because reductions in cooling capacity can
be linked to increased system runtimes, this research can be used to inform decisions
about maximum filter loading values that should inform filter replacement schedules.
Once a maximum acceptable reduction in sensible capacity is established, this data can be
used to identify the airflow and filter pressure drop thresholds, which can impact future
decisions about filter replacement timing.
Data herein can be used to inform filter replacement strategies in U.S. homes.
Once a maximum acceptable reduction in sensible capacity is established, this data can be
used to identify the airflow and filter pressure drop thresholds, which can impact future
decisions about filter replacement timing. Taking values relative to no filter conditions
(for which 100% flow is impossible in most systems because even the lowest efficiency
filter will induce more pressure drop than no filter conditions), one could define, say, a
20% reduction in sensible capacity relative to no filter conditions as the maximum
allowable value. This sensible capacity reduction is an admittedly arbitrary value that
should be verified in energy simulations and field tests. According to the data in Figure
29, sensible capacity is reduced to 80% of no filter sensible capacity when airflow is
reduced by approximately 30% to only 70% of the no filter airflow. According to Figure
20, a 30% reduction in airflow would occur at a filter pressure drop of approximately 225
59
Pa. Similarly, if only a 10% reduction in sensible capacity was set as the criteria, a
maximum of approximately 15% reduction in airflow (relative to no filter) would be
allowed, which would be reached a filter pressure drop of approximately 125 Pa.
Although these criteria should be further refined with simulations and larger field studies,
the methods and results herein suggest that enough information can be gathered to
establish maximum filter pressure drops allowed in residential HVAC systems, and
particularly in those that rely on blowers without sophisticated flow controls.

60
CHAPTER 5
CONCLUSION
This study sought to: (i) develop and apply a methodology for simulating filter
loading in-situ; (ii) measure the impact of simulated filter loading on air conditioner (AC)
performance in-situ; and (iii) provide a greater understanding of when a filter is dirty
and in should be replaced. Five central AC systems in the Atlanta metro-region were
evaluated using TrueFlow from The Energy Conservatory. Filter loading was simulated
by installing the TrueFlow airflow metering device and partially taping off the face at
three different increments. This resulted in measurements at five discrete static pressure
conditions: no filter, TrueFlow, TrueFlow Taped #1, TrueFlow Taped #2, and TrueFlow
Taped #3. The results strongly suggest that this methodology is an accurate means of
simulating filter loading in-situ and provides important HVAC performance data. Results
from each major section are summarized below.

Filter Pressure Drop & System Airflow
The relationship between induced filter pressure drop and system airflow rates
was first evaluated in each system. The absolute value of the pressure drop across the
filter was compared to the reduction in airflow (as a percentage of total system flow
without filter installed) for each induced pressure drop condition. Overall, as filter
pressure drop increased, airflow rates generally decreased, particularly for the known
PSC blowers. The two systems that were apparently ECM blowers (or at least variable
speed blowers) responded to increased filter pressure drop by nearly maintaining airflow
rates until reaching a maximum pressure and rapidly decreasing in flow, which is
consistent with other ECM data.

61
Filter Pressure Drop & Temperature Difference Across the Coil
The relationship between induced filter pressure drop and the dry bulb
temperature difference across the evaporator coil (T) was evaluated in each system. The
difference in temperature (T) before passing the evaporator coil (pre-filter) and after
passing the coil (in supply plenum) was measured. With moderate certainty, it was
concluded that as induced filter pressure drop increased, the T also increased in these
systems. These data suggest what other laboratory and field studies have shown: as
airflow rates are reduced in the presence of larger filter pressure drops, sensible capacity
will not decrease linearly with flow because the temperature difference across the coil
increases slightly and supply air is delivered at a lower temperature. This section also
further indicated the validity of this methodology for simulating filter loading in-situ.

Filter Pressure Drop & Absolute Humidity
The relationship between induced filter pressure drop and changes in the
difference in absolute humidity across the coil (W) was evaluated in each system. The
change in W across the coil (as a percentage of W without filter installed) was
compared to the absolute value of the pressure drop across the filter for each induced
pressure drop condition. There was no observed correlation between filter pressure drop
and W.

Airflow and Temperature
The relationship of airflow and the dry bulb temperature difference across the
evaporator coil (T) was evaluated in each system. The difference in temperature (T)
before passing the evaporator coil (pre-filter) and after passing the coil (in supply
plenum) was measured. These data suggest what other laboratory and field studies have
shown: as airflow rates are reduced in the presence of larger filter pressure drops,
temperature difference across the coil increases slightly and supply air is delivered at a
62
lower temperature. This data also indicates that there is a stronger relationship between
T and airflow rates compared to induced filter pressure drop because induced filter
pressure drop did not affect airflow similarly in each system and it is airflow rates that
directly impact T.

Airflow and Absolute Humidity
The relationship between system airflow and changes in the difference in absolute
humidity across the coil (W) was evaluated in each system. There was no observed
correlation between system airflow and W.

Airflow and System Capacity
The change in sensible, latent, and total capacities (as a percent of system
capacity with no filter installed) was compared to the reduction in airflow (as a percent of
total system flow without filter installed). Overall as airflow decreased in response to
larger filter pressure drops, so did sensible, latent, and total capacity, although impacts
were not necessarily linear.

Analysis Summary
Based on a relatively strong linear regression of system airflow versus induced
filter pressure drop, airflow rates were reduced 10% from no filter airflow at around 90
Pa and reduced 20% from no filter airflow at around 150 Pa. This data can inform
decisions about filter replacement and be used to educate builders, contractors, and
homeowners about the importance of routine HVAC maintenance. Once a maximum
acceptable reduction in sensible capacity is established, this data can be used to identify
the airflow and filter pressure drop thresholds. For example, a 10% reduction in sensible
capacity would result in approximately a 10% increase in AC run time, all is being equal.
If a 10% reduction in sensible capacity is selected as the threshold for filter replacement,
63
then the data indicates filters should be replaced at around 80-85% of no filter airflow,
which is around a filter pressure drop of 150 Pa.

Recommendations
This pilot study provided valuable proof of concept for an approach to simulating
filter loading in-situ. Taping the face of the TrueFlow was identified to work consistently
well to simulate filter loading. In the future, the study should be expanded to a greater
number of central air conditioners and continuous measurements should be recorded
using data loggers. One challenge was establishing when the AC reached steady state.
Greater certainty may be possible by recording continuous measurements and waiting
longer between each simulated filter condition. Measuring the air handler and condenser
power draw would also provide potentially valuable data. Additionally, the in-situ AC
performance data can be used to evaluate and improve current computer models of
HVAC performance.

64
APPENDIX A
TEST SYSTEM 1
No Filter True
Flow
True
Flow
Taped 1
True
Flow
Taped 2
True
Flow
Taped 3
Supply Measurements
Relative humidity, (fractional) 0.89 0.92 0.93 0.92 0.93
Dry-bulb temperature, t (C) 12.80 13.20 12.50 11.50 10.00
Humidity ratio, W (kgw/kgda) 0.01 0.01 0.01 0.01 0.01
Return Measurements
pressure wrt outside 67.50 58.30 53.80 40.10 32.60
Filter Measurement
Pressure (Coil Side) 67.50 107.90 139.40 184.80 213.00
Pfilt (pa) 0.00 49.60 85.60 144.70 180.40
True Flow Pressure (Pa) - 62 - - -
True Flow Air Flow (CFM) - 906 - - -
Adjusted airflow (CFM) 975 906 870 751 677
Percent of no filter flow 100% 93% 89% 77% 69%
deltaW 0.002 0.001 0.001 0.002 0.002
Percent of no filter deltaW 100.00% 40.60% 47.91% 81.70% 98.16%
deltaT 9.50 8.90 9.50 10.30 11.70
Percent of no filter deltaT 100.00% 93.68% 100.00% 108.42% 123.16%
System Capacity
Qfan (m3/s) 0.46 0.43 0.41 0.35 0.32
Air density (p, kg/m3) 1.2 1.2 1.2 1.2 1.2
Specific heat of air (C, kJ/(kgK) 1.005 1.005 1.005 1.005 1.005
deltaT (K) 9.50 8.90 9.50 10.30 11.70
deltaW (kg/kg) 0.002 0.001 0.001 0.002 0.002
Hfg (kJ/kg) 2257 2257 2257 2257 2257
Sensible Capacity (kW) 5.27 4.59 4.71 4.40 4.51
Latent Capacity (kW) 2.98 1.13 1.28 1.88 2.04
Total Capacity (kW) 8.25 5.72 5.98 6.28 6.55
Sensible Capacity (kBTU/hr) 17.99 15.66 16.06 15.03 15.39
Latent Capacity (kBTU/hr) 10.18 3.84 4.35 6.41 6.94
Total Capacity (kBTU/hr) 28.16 19.50 20.41 21.44 22.34
Sensible Capacity (tons) 1.50 1.30 1.34 1.25 1.28
Latent Capacity (tons) 0.85 0.32 0.36 0.53 0.58
Total Capacity (tons) 2.35 1.62 1.70 1.79 1.86
Percent no filter Sensible
Capacity
100.00% 87.07% 89.28% 83.57% 85.59%
Percent no filter Latent Capacity 100.00% 37.73% 42.77% 62.97% 68.22%
Percent no filter Total Capacity 100.00% 69.24% 72.47% 76.12% 79.31%
65
TEST SYSTEM 2
No Filter True
Flow
True
Flow
Taped 1
True
Flow
Taped 2
True
Flow
Taped 3
Supply Measurements
pressure wrt house 79.80 78.10 75.00 72.40 69.10
Return Measurements
pressure wrt house 62.10 59.20 57.20 54.30 53.30
Filter Measurement
Pressure (Coil Side) 87 104.5 112.6 129.6
Pfilt (pa) 0 27.8 47.3 58.3 76.3
Percent of no filter flow 100.00% 97.64% 95.97% 93.51% 96.73%
deltaW 0.002 0.002 0.002 0.002 0.002
deltaT 11.00 11.20 11.80 11.80 12.10
System Capacity
Qfan (m3/s) 0.33 0.32 0.31 0.30 0.32
Air density (p, kg/m3) 1.2 1.2 1.2 1.2 1.2
deltaT (K) 11.00 11.20 11.80 11.80 12.10
deltaW (kg/kg) 0.002 0.002 0.002 0.002 0.002
Hfg (kJ/kg) 2257 2257 2257 2257 2257
Capacity
100.00% 99.41% 102.95% 100.31% 106.41%

66
TEST SYSTEM 3
No Filter True
Flow
True
Flow
Taped 1
True
Flow
Taped 2
True
Flow
Taped 3
Supply Measurements
pressure wrt attic 133.20 130.30 125.80 118.90 107.60
Return Measurements
pressure wrt attic 35.10 34.80 33.40 31.90 28.40
Filter Measurement
Pressure (Coil Side) 51.1 65.3 89.4 110.3
Pfilt (pa) 0 16.3 31.9 57.5 81.9
True Flow Pressure (Pa) - 16.7 - - -
deltaW 0.003 0.003 0.003 0.003 0.002
deltaT 12.00 12.90 13.20 13.60 13.60
Percent of no filter deltaT 100.00% 107.50
%
110.00% 113.33% 113.33%
System Capacity
Qfan (m3/s) 0.225 0.224 0.219 0.214 0.202
Air density (p, kg/m3) 1.2 1.2 1.2 1.2 1.2
deltaT (K) 12.00 12.90 13.20 13.60 13.60
deltaW (kg/kg) 0.003 0.003 0.003 0.003 0.002
Hfg (kJ/kg) 2257 2257 2257 2257 2257
Capacity
100.00% 107.04
%
107.30% 108.04% 101.94%
67
TEST SYSTEM 4
No Filter True
Flow
True
Flow
Taped 1
True
Flow
Taped 2
True
Flow
Taped 3
Supply Measurements
Return Measurements
Filter Measurement
Pressure (Coil Side) - 93.4 161.6 210 251
Pfilt (pa) 0 51.7 125.3 181.2 230.8
deltaW 0.002 0.002 0.002 0.003 0.002
deltaT 8.10 8.30 9.00 12.30 11.20
System Capacity
Qfan (m3/s) 0.68 0.60 0.56 0.50 0.42
Air density (p, kg/m3) 1.2 1.2 1.2 1.2 1.2
deltaT (K) 8.10 8.30 9.00 12.30 11.20
deltaW (kg/kg) 0.002 0.002 0.002 0.003 0.002
Hfg (kJ/kg) 2257 2257 2257 2257 2257
Capacity
100.00% 90.47% 91.52% 111.41% 84.96%

68

TEST SYSTEM 5
Supply Measurements No Filter True
Flow
True
Flow
Taped 1
True
Flow
Taped 2
True
Flow
Taped 3
pressure wrt outside 63.5 62.8 62.5 62 59.9
Return Measurements
Filter Measurement
Pressure (Coil Side) 0 125.2 127.5 132.2 141.9
Pfilt (pa) 0 8.4 15.9 30.8 57.3
True Flow Pressure (Pa) - 16.2 - - -
deltaW 0.002 0.002 0.002 0.002 0.002
deltaT 9.20 8.90 9.00 9.00 8.90
System Capacity
Qfan (m3/s) 0.30 0.29 0.28 0.27 0.25
Air density (p, kg/m3) 1.2 1.2 1.2 1.2 1.2
deltaT (K) 9.20 8.90 9.00 9.00 8.90
deltaW (kg/kg) 0.002 0.002 0.002 0.002 0.002
Hfg (kJ/kg) 2257 2257 2257 2257 2257
Capacity
100.00% 94.12% 93.03% 88.68% 80.10%
69
TEST SYSTEM 6
No Filter True
Flow
True
Flow
Taped 1
True
Flow
Taped 2
True
Flow
Taped 3
Supply Measurements
Return Measurements
Filter Measurement
Pressure (Coil Side) - 118.6 133.6 155.2 176.4
Pfilt (pa) 0 36 57.7 86.4 115.8
deltaW 0.002 0.002 0.002 0.002 0.002
Percent of no filter deltaW 100% 100.0% 100.0% 100.0% 100.0%
deltaT 9.100 9.700 9.900 10.100 10.400
System Capacity
Qfan (m3/s) 0.49 0.49 0.49 0.49 0.49
Air density (p, kg/m3) 1.2 1.2 1.2 1.2 1.2
deltaT (K) 9.10 9.70 9.90 10.10 10.40
deltaW (kg/kg) 0.0020 0.002 0.002 0.002 0.002
Hfg (kJ/kg) 2257 2257 2257 2257 2257
Capacity
100.00% 106.35% 108.05% 109.73% 112.73%
70
REFERENCES
Alnor Instruments. Retrieved from http://www.alnor-usa.com
ANSI/ASHRAE (2007). ANSI/ASHRAE Standard 52.2. Method of testing general ventilation
air-cleaning devices for removal efficiency by particle size. Atlanta, USA: American
Society of Heating, Refrigerating and Air-Conditioning Engineers.
ASHRAE. 2005. Chapter 35, Duct Design. 2004 ASHRAE HandbookHVAC Systems and
Equipment. Atlanta: American Society of Heating, Refrigerating and Air-Conditioning
Engineers, Inc.
Atlanta Regional Commission (unknown). Appendix: Housing. Retrieved October 17, 2013 from
http://www.atlantaregional.com/land-use/housing.
Beaton, L. (personal communication, October 18, 2013).
Brandemuehl MJ, Gabel S, Andersen I (1993). A Toolkit for Secondary HVAC System Energy
Calculation. Atlanta, USA: American Society of Heating, Refrigerating and Air-
Conditioning Engineers.
Carrier Corporation. 1994. Charging procedures for residential condensing units. 020-122.
Syracuse, NY. Carrier Corporation.
Chimack, M., and D. Sellers. 2000. Using extended surface air filters in heating ventilation and
air conditioning systems: Reducing utility and maintenance costs while benefiting the
environment. In the Proceedings of the 2000 ACEEE Summer Study on Energy Efficiency
in Buildings 3:7788.
Downey, T., and J. Proctor. 2002. What can 13,000 air conditioners tell us? In the Proceedings of
the 2002 ACEEE Summer Study on Energy Efficiency in Buildings 1:5367.
Earnest, G.S., Chen, D., and Pui, David. Y. H. An experimental study of filter loading with liquid
aerosols (EPHB 218-05r). To be published in American Industrial Hygiene Association
Journal.
Fisk, W., D. Faulkner, J. Palonen, and O. Seppanen. 2002. Performance and costs of particle air
filtration technologies. Indoor Air 12(4):22334.
71
International Code Council, International Residential Code for One- and Two-Family Dwellings
(ICC, 2012).
Murray, Matt, Fitzpatrick, Shawn. Residential HVAC Electronically Commutated Motor Retrofit
Report. Advanced Energy, February 3
rd
, 2012.
Nassif, Nabil. The impact of air filter pressure drop on the performance of typical air-
conditioning systems. Building Simulation, Vol. 5, No. 4. (1 December 2012), pp. 345-
350.
Palani, M., D. ONeal, and J. Haberl. 1992. The effect of reduced evaporator air flow on the
performance of a residential central air conditioner. In the Proceedings of the 1992
Symposium on Building Systems in Hot-Humid Climates, pp. 2026.
Parker, D., J. Sherwin, R. Raustad, and D. Shirey III. 1997. Impact of Evaporator Coil Airflow in
Residential Air-Conditioning Systems. ASHRAE Transactions 103(2):395-405.
Proctor, J. 2012. Residential AC Filters. ASHRAE Journal.
Rodriguez, A., D. ONeal, M. Davis, and S. Kondepudi. 1996. Effect of reduced evaporator
airflow on the high temperature performance of air conditioners. Energy and Buildings
24:195201.
Sachs, H., T. Kubo, S. Smith, and K. Scott. 2002. Residential HVAC fans and motors are bigger
than refrigerators. In the Proceedings of the 2002 ACEEE Summer Study on Energy
Efficiency in Buildings1:26172.
Siegel, J., Walker, I., and Sherman, M., 2002. Dirty Air Conditioners: Energy Implications of
Coil Fouling. Proceedings of the 2002 ACEEE Summer Study on Energy Efficiency in
Buildings, 1 (2002), 287-300.
Stephens, B., Novoselac, A., and Siegel, J.A. 2010. The effects of filtration on pressure drop and
energy consumption in residential HVAC systems (RP-1299). HVAC&R Research 16(3),
273-294
The Energy Conservatory. 2006. TrueFlow Air Handler Flow Meter Operation Manual.
Minneapolis, MN: The Energy Conservatory.

72
U.S. Census Bureau, Current Housing Reports, Series H150/09, American Housing Survey for
the United States: 2009, U.S. Government Printing Office, Washington, DC, 20401.
Printed in 2011.
U.S. Energy Information Administration, Annual Energy Review 2011, DOE/EIA-0384 (2011),
September 2012.
Yang, Li; Braun, James E.; and Groll, Eckhard A., "The Impact of Evaporator Fouling on the
Performance of Packaged Air Conditioners" (2004). International Refrigeration and Air
Conditioning Conference. Paper 687.

HVAC Thesis

Uploaded by

Copyright:

Available Formats

HVAC Thesis

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

HVAC Thesis

Uploaded by

Copyright:

Available Formats

How does filter loading affect system airflow and pressure drop across the filter?

How does filter loading affect temperature differences across the system?

THE IMPACT OF FILTER LOADING ON RESIDENTIAL HVAC

You might also like