Fan Performance Troubleshooting Guide - FE-100

Fan Performance Troubleshooting Guide - FE-100
FAN ENGINEERING
FE-100
Information and Recommendations for the Engineer
Fan Performance
Troubleshooting Guide
How Fan Systems Work
Before getting into troubleshooting it is worth our while
to look at how fans and duct systems work together. A
duct system’s design affects how a fan operates; how
the fan is selected affects how the duct system functions.
Fans operate on all kinds of different duct systems, ranging from no ductwork at all (mancoolers and panel fans)
to those with short simple runs of ductwork to systems
with long complicated runs of ductwork, and finally to
completely shut off systems (where the purpose is only
to generate pressure). The longer and more complicated
the ductwork gets, the more resistance to flow there is.
When fans are tested in a laboratory, they are tested
under conditions that simulate the extremes of no resistance, complete resistance (no flow at all), and everything
in between. Figure 1 shows the test points of a typical
fan operating at constant speed and at standard air density. The flow rate is plotted on the horizontal axis, the
fan static pressure is plotted on the lower vertical axis,
while the top vertical axis shows the brake horsepower.
The point labeled A1 simulates a duct system where there
is no resistance flow since the static pressure is zero.
This point is sometimes referred to as “wide open volume” or “free delivery.” Point A2 shows the brake horsepower measured at this point. Brake horsepower is the
power required at the fan shaft. It does not include
losses associated with the motor or belt drives.
Points B through F are generated by increasing the
resistance to flow. In the laboratory, closing a damper
does this. Connecting the test points generates the fan
performance curve. Point F is the point where the fan
delivers no airflow and is purely generating pressure. This
point is sometimes referred to as “shut off” or “no delivery.”
Once the performance of a fan at a given speed is
established, it is possible to predict the performance at
other speeds or other air densities using a set of equations known as the fan laws. The first fan law states that
the change in flow rate is proportional to the change in
speed. This can be written as:
QN = QO x
( )
RPMN
RPMO
49 IN. SINGLE WIDTH BACKWARD INCLINED
700 RPM
ρ = 0.075 lb/cu.ft
BRAKE HORSEPOWER
Trying to determine why a fan is not working the way
you want it to can sometimes be a very confusing and
frustrating experience. This document, based on our
experience, was put together to help fan users diagnose
and troubleshoot fan performance problems.
Figure 1. Test Points at Constant Speed and
Standard Air Density
30
E2
D2
B2
A2
10
F2
E1
D1
F1
4
3
C1
2
B1
1
0
A1
0
1
3
2
4
5
6
CFM in 10000's
The second fan law states that the change in pressure
is proportional to the square of the change in speed:
PN = PO x
( )
RPMN
RPMO
2
ρN
xρ
O
where:P=pressure (in. w.g.)
ρ=air density (lb/cu.ft)
The change in the power requirement is proportional to
the cube of the change in speed, as shown in the equation of the third fan law:
HpN = HpO x
where:Q =volumetric flow rate (CFM)
RPM=fan speed in revolutions per minute
N =new conditions
O =old conditions
C2
20
5
STATIC PRESSURE
Introduction
where:Hp =
©2010 Twin City Fan Companies, Ltd.
( )
RPMN
RPMO
3
ρN
xρ
O
fan power requirement (hp)
Figure 2 shows the same fan as Figure 1 with the
performance calculated at different fan speeds.
The fan laws show that the pressure and horsepower
change with density but the flow rate stays the same.
This is why fans are sometimes referred to as constant
volume machines. But because the pressure and horsepower change with air density, it is important to select
fans for the proper operating conditions. When testing
fans, it is important to correct the test data for density
effects and to make sure when comparing test data to
the design data that they are both at the same density.
Figure 3 shows the performance of the same fan as
Figure 1 at standard air density (0.075 lb/cu.ft) and at an
elevation of 5,000 feet above sea level where air weighs
0.059 lb/cu.ft.
How Systems Work
Figure 2. Test Points at Different Speeds
and Standard Air Density
where:
BRAKE HORSEPOWER
49 IN. SINGLE WIDTH BACKWARD INCLINED
ρ = 0.075 lb/cu.ft
700 RPM
650 RPM
600 RPM
550 RPM
10
500 RPM
STATIC PRESSURE
5
P1 = P2 x
P1
P2
Q1
Q2
=
=
=
=
600 RPM
550 RPM
500 RPM
2
1
0
0
1
2
3
4
5
6
CFM in 10000's
30000
= 4.5 in. w.g.
20000
Figure 4 shows these two points along with the system line that is defined by them. If the flow remains
turbulent (and it almost always does) and if the duct
configuration does not change (including the position of
any dampers), the pressure required for a given flow rate
will follow this line. Some variable air volume (VAV) systems require a minimum pressure at the end of the duct
runs in order for the VAV boxes to open. No flow occurs
until this pressure is reached. The system line equation
for this type of system is:
Figure 3. Test Points at Different Air Densities
P1 = P0 + (P2 - P0) x
49 IN. SINGLE WIDTH BACKWARD INCLINED
700 RPM
BRAKE HORSEPOWER
( )
2
P1 = 2 x
650 RPM
3
2
pressure at point 1
pressure at point 2
flow rate at point 1
flow rate at point 2
700 RPM
4
( )
Q1
Q2
This equation defines a system resistance curve, which
is usually referred to as a “system line.” As long as the
duct system remains constant (no dampers changing,
filters getting dirty, etc.) the pressure required to achieve
different flow rates will follow this line.
Example: A duct system requires 2 in. w.g. in order
to have a flow rate of 20,000 CFM. What pressure is
required if we want the flow rate to increase to 30,000
CFM?
Using the equation above:
30
20
It makes sense to say that more pressure is required if
you want to move more air through your duct system.
You can feel this when you try to blow more air through
a straw. How the airflow varies with pressure is a function
of how the air flows through the duct system.
For the vast majority of duct systems, the air flowing
in the ducts is turbulent. Turbulent flow is characterized
by rapid, random fluctuations of velocity and pressure.
With turbulent flow, the relationship between the flow rate
and the pressure is as follows:
( )
Q1
Q2
2
where P0 = pressure required to have flow
ρ = 0.075 lb/cu.ft
30
20
ρ = 0.059 lb/cu.ft
Figure 4. Two Performance Points Along the System Curve
10
STATIC PRESSURE
5
ρ = 0.075 lb/cu.ft
4
3
ρ = 0.059 lb/cu.ft
2
4
3
2
1
1
0
0
0
1
2
3
4
CFM in 10000's
2
STATIC PRESSURE
5
5
6
0
1
2
4
3
5
CFM in 10000's
6
Fan Engineering FE-100
Figure 5 shows this type of system passing through a
design point of 20,000 CFM at 2 in. w.g. and a VAV box
requirement of 1 in. w.g.
Figure 5. Variable Air Volume System Curve
5
System Effects
STATIC PRESSURE
4
Most of the time fans are tested with open inlets, and
designs that do not have an integral inlet bell will be
tested with one to simulate inlet ductwork. Figure 7
shows the streamlines the air follows as it enters a fan
with a well-designed inlet bell and with a straight run of
duct upstream of the fan. The airflow is smooth and the
velocity at the entrance is uniform.
3
2
1
0
Even though a fan is selected and sold to deliver a
specified flow rate at a specified static pressure and
speed, all the fan manufacturer can be sure of is that
the point of operation will lie somewhere on the fan curve
(sometimes even this is in doubt depending on the duct
geometry, which will be explained later). Where along the
fan curve the operating point lies depends entirely on the
ductwork. The fan will operate at the place where the
system line intersects the fan curve.
0
1
4
3
5
CFM in 10000's
2
6
Figure 7. Air Streamlines at Fan Inlets
INLET BELL
FAN
INLET DUCT
FAN
How Fans and Systems Work Together
A fan will operate at the point where the system line
intersects the fan curve. To generate the different test
points shown in Figure 1, the system resistance had to
change. Remember that at Point A1 there is no resistance
to flow and that the resistance to flow was increased as
we moved towards F1. As a result each test point has a
unique system resistance curve associated with it, as
shown in Figure 6. In fact, there are an infinite number
of system lines between points A1 and F1.
Figure 6. System Resistance Curves
BRAKE HORSEPOWER
49 IN. SINGLE WIDTH BACKWARD INCLINED
700 RPM
ρ = 0.075 lb/cu.ft
30
E2
C2
B2
A2
10
F2
E1
D1
F1
4
SYSTEM
LINES
3
C1
2
B1
1
0
A1
0
1
2
3
4
CFM in 10000's
3
Figure 8. Typical Outlet Velocity Profiles
20
5
STATIC PRESSURE
D2
The velocity profile coming out of even the most efficient fans is not uniform. Especially on centrifugal fans
there are regions of high velocity air coming out of the
impeller. Figure 8 shows typical outlet velocity profiles for
centrifugal and axial fans.
5
6
A length of straight duct is required at the outlet to
allow the high velocity (kinetic energy) to be converted to
static pressure (potential energy). To get full pressure
conversion, a duct at least as long as the 100% effective
duct length is required. AMCA defines the 100% effective
duct length to be a minimum of two and one half
equivalent duct diameters for outlet velocities up to 2,500
fpm, with 1 duct diameter added for each additional
1,000 fpm. Also it is important that the outlet connection
be smooth. In the test code, AMCA specifies that the
area of the outlet duct be between 95% and 105% of
the fan outlet area and that any transitions used have
sides that are sloped no more than 15° on converging
sections and 7° on diverging sections.
Duct fittings or fan installations that do not provide
smooth, uniform flow at the fan inlet or complete pressure
recovery at the fan outlet cause the fan to perform at a
Fan Engineering FE-100
level below its catalog rating. Elbows cause non-uniform
airflow at the fan inlet. More air enters the fan at the
outside of the elbow than at the inside. Figure 9 shows
what happens to a centrifugal fan’s performance curve
first when an elbow is located at its inlet, and second
when an elbow is located at the inlet and there is no
outlet duct. This loss in performance is known as system
effect.
Figure 10. Drop Box
Figure 11. Inlet Box
AIRFLOW
AIRFLOW
Figure 9. System Effects Curves
BRAKE HORSEPOWER
49 IN. SINGLE WIDTH BACKWARD INCLINED
700 RPM
ρ = 0.075 lb/cu.ft
30
20
10
5
STATIC PRESSURE
CATALOG FAN CURVE
with this type of inlet. Most fans can be purchased with
an optional inlet box as shown in Figure 11. These boxes
are designed to have minimal and predictable system
effects.
Often the duct configuration can cause an airflow pattern that is detrimental to fan performance. Figure 12
shows one example of a duct configuration that causes
the air to spin in the direction opposite the impeller rotation (counter rotating swirl). When this happens there may
be a slight increase in pressure produced by the fan, but
with a significant increase in horsepower.
Figure 12. Duct Configuration Causing
Counter Rotating Swirl
FAN WITH ELBOW ON INLET
4
FAN WITH ELBOW ON INLET
AND NO OUTLET DUCT
3
2
1
0
0
1
3
2
4
5
6
CFM in 10000's
Duct configurations that cause system effects should
be avoided whenever possible. To help in those situations
when it is unavoidable, AMCA Publication 201 contains
methods of estimating the system effects for some of the
most common situations. A series of tests were run in
AMCA’s laboratory. Although variations due to factors
such as fan type and design were observed, figures
showing system effect factors were put together to allow
fan users to estimate system effects.
The system effect factor must be added to the system
design pressure to select the fan and determine the
required speed and horsepower. The system effect will
cause the fan to generate only the system design pressure, not the pressure that the fan was selected for.
Similar duct configurations can cause the air to spin
in the direction of impeller rotation (pre-rotating swirl). This
reduces the pressure produced, and the power consumed
by the fan. Inlet vanes produce this effect to control flow
with horsepower savings. Like drop boxes, the severity of
the system effect caused by counter rotating and prerotating spin is impossible to predict. Avoid duct configurations that have the potential to cause spin.
Figure 13 shows how straightening vanes placed in the
elbows can minimize system effect. If there is room,
straighteners added to the duct just upstream of the fan
inlet will break up spinning air. Case Studies 2 and 4 in
the following section demonstrate how straighteners can
solve performance problems.
Figure 13. Straightening Vanes
Inlet System Effects
Elbows mounted near the inlet of a fan are the most
common inlet system effect. Elbows should be installed
at least three equivalent duct diameters upstream of a
fan inlet. Not only do elbows mounted closer cause a
system effect but they can also cause pressure fluctuations, increased sound levels, and structural damage to
the fan.
As depicted in the previous example, elbows at a
centrifugal fan’s inlet cause a system effect. Often the
elbow at the inlet is not a true elbow, but a drop box
as illustrated in Figure 10. The system effect for this type
of inlet elbow is nearly impossible to predict, as it is a
function of fan type, inlet size, and box width and depth.
Losses in flow rate as high as 45% have been observed
4
1D to 2D
D
Ductwork sized smaller than the fan inlet, butterfly
valves and belt guards placed too close to the fan inlet
are two examples of obstructions that cause system
effects. Fan inlets located too close to a wall cause nonuniform flow and a resulting system effect.
Fan Engineering FE-100
Table 1. Blast Area
FAN TYPE
BI & AF SWSI
BI & AF DWDI
INDUSTRIAL FANS
RADIAL TIP FANS
FC DWDI
BLAST AREA ÷ OUTLET AREA
0.80
0.77
1.00
1.00
0.61
The charts in Publication 201 show that tubeaxial fans
do not have a system effect curve for no outlet duct.
Tests at AMCA indicated very little loss in performance
with tubeaxial fans without outlet ducts and with vaneaxial fans with outlet ducts as short as 50% effective
duct length.
Figure 14. Elbow Positions
Elbows are frequently
mounted close to fan outlets.
C
This placement will disrupt
the energy conversion from
WORST
velocity pressure to static
pressure. As with short ducts,
tests indicate that the factors
are negligible with tubeaxial
fans and greater with centrifugal fans than with vaneaxial fans. The position of the
B
elbow influences how great
BAD
the system effect is with centrifugal fans. When the elbow
follows the curvature of the
scroll, as in position B in
Figure 14, the system effect
is minimized, but when the
A
elbow goes against the curBEST
vature of the scroll as in
position C the effect is worsened. Not all system effects are bad. Frequently, cones
or rectangular transitions called evasés are added to fan
outlets to increase the amount of static pressure regain
over what is possible with a straight duct section that
matches the outlet area. The increase in pressure can be
estimated using the static pressure regain for expansion
calculation, or corrected performance data can be
obtained from the manufacturer.
Manufacturing and Testing Tolerances
Most fans are fabricated metal assemblies and there are
slight differences from one fan to the next. These variances could be due to manufacturing tolerances, welding
distortion, or assembly tolerances. These small differences
may have a slight effect on how the fans perform.
In addition to the differences due to manufacturing,
there are uncertainties in the performance of a fan due
to testing tolerances or errors. Even when tested in the
laboratory there are variations in the readings that can
contribute to test error. Sources of these variations can
be due to the instruments or due to the test setup.
Periodically instruments need to be calibrated. The
reading an instrument gives often drifts or changes over
long periods of time. When an instrument is calibrated,
5
Figure 15. Field Test vs. Catalog Performance
49 IN. SINGLE WIDTH BACKWARD INCLINED
700 RPM
ρ = 0.075 lb/cu.ft
RATED PRESSURE
5
50
TOLERANCE ALLOWED
40
4
3
30
RATED POWER
2
20
10
1
BRAKE HORSEPOWER
The most common cause of system effects on fan outlets
is an insufficient length of straight duct length that prevents the full conversion of velocity to static pressure. As
defined previously, the length of straight duct length
required is the 100% effective duct length. Many of the
charts in AMCA Publication 201 use the blast area, which
is the outlet area minus the area of the cut off. The blast
area is not normally included in catalog performance data
but is shown in Table 1 as a percentage of outlet area
for various fan types.
its measurements are compared to a standard that is
traceable to the National Bureau of Standards. In order
to get more accurate results, a calibration correction is
frequently applied to the reading the instrument gives. If
the period of time since the instrument was calibrated is
too long, the instrument is out of calibration, and there
is a greater chance for the instrument to give a reading
that varies slightly from the “true” reading. Because the
air flowing through fans is turbulent, the readings vary
over time. These variations must be averaged, either
electronically or visually. How accurately the averaging is
done affects the uncertainty of the reading.
Air density corrections usually have to be made.
Catalog performance is based on standard air density,
which is 0.075 lb/cu.ft. This corresponds with air at 70°F
at sea level and a relative humidity of 50%. Air density
calculations require readings of barometric pressure, wet
bulb temperature, dry bulb temperature, and elevation
above sea level. If the fan handles a gas other than air,
the composition of the gas must be known in order to
make the appropriate air density corrections. The care
and precision in taking these measurements and the
subsequent calculations affect the accuracy of the test
data.
An association of fan manufacturers, the Air Movement
and Control Association (AMCA), has set standards for
the industry. They have standards that specify how fans
are to be tested in the laboratory. They also have established a certified ratings program. Part of the certified
ratings program requires that a fan line’s performance is
tested periodically in an independent laboratory. The test
results must be within a certain tolerance of the catalog
rating. This tolerance, shown in Figure 15, takes into
account the manufacturing variances and the variances
expected when testing a fan in two laboratories.
STATIC PRESSURE
Outlet System Effects
SYSTEM LINE
0
0
1
4
2
3
CFM in 10000's
5
6
0
There is a tendency to compare the performance along
lines of either constant pressure or constant volume. This
is not the correct way. Notice that the tolerance for flow
and pressure is based on a system line. This is done
because that is how fans and duct systems interact. The
tolerance varies with the location of the fan curve, but
corresponds to a change in fan speed of approximately
3%. If the fan test results are within the shaded region,
the fan is operating within acceptable limits to its performance curve. The tolerances shown in Figure 15 are
based on laboratory testing. Field testing is rarely as
accurate as laboratory testing.
Fans in the laboratory are tested under ideal conditions. The inlet conditions provide for smooth entry of air
Fan Engineering FE-100
into the fan and straight ductwork on the discharge
allows static pressure regain. Such ideal conditions in the
field are the exception rather than the norm and as a
result most fans in the field do not perform as well as
they do in the laboratory. The laboratory test setup must
meet strict requirements. Duct connections must exactly
match the fan. If transitions are used, they must have
very low expansion or contraction angles. When using a
pitot tube test setup, the ductwork upstream of the traverse plane must have a straight run of at least ten duct
diameters and a flow straightener that meets specific
construction standards. If flow nozzles are used, they
must be manufactured to tight dimensional tolerances and
in many cases must have flow settling screens upstream
of them. These screens ensure that the flow entering the
nozzles is uniform to a high degree.
Field testing instruments are typically not manufactured
with the same precision, or as accurate or calibrated with
the same frequency as laboratory instruments. So when
comparing field test data to the catalog level, it is reasonable to use a testing tolerance greater than that shown
in Figure 15.
Troubleshooting Guide — Case Studies
The rest of this newsletter consists of case studies that demonstrate some of the
most common field performance problems and the logic that was followed in figuring out what the problem was.
Case Study 1
System Resistance Higher Than Anticipated
( )
( )
( )
RPMnew = 1196 x
This case study demonstrates one of the most common
fan performance problems. A 36" tubeaxial fan was
selected and installed to deliver 15,000 CFM at a static
pressure of 1.70 in. w.g. After the fan was installed, an
air performance test indicated a flow rate of 13,600 CFM
with a static pressure of 1.68 in. w.g. and 6.8 BHP.
Further investigation revealed that four additional
elbows and 10 feet of additional duct length was in
order to clear some obstructions, as shown in the figure
below. The additional ductwork increases the resistance
to flow resulting in a new system line as shown on the
curve. We would expect the fan to operate where the
new system curve intersects the catalog curve, at 14,080
CFM and 1.80 in. w.g. The difference between the measured performance and the expected performance is due
to a combination of test error and system effect due
to the elbow located upstream of the fan.
To get the design flow rate, we can use the fan laws
to determine the increase in fan speed required:
HPnew = 6.8 x
SPnew = 1.68 x
15000
= 1319
13600
1319 3
= 9.12
1196
1319 2
= 2.04
1196
Figure 16a. System as Designed
AIR
FLOW
Figure 16b. System as Installed
AIR
FLOW
Figure 17. Measured and Expected Performance Curves
Customer:
Job Name:
Fan Tag:
Model: Tubeaxial Fan
9
4.5
7
3.5
SP
3.0
2.0
1.5
MEASURED
CFM & SP
1.0
0
6
BHP
EXPECTED
CFM & SP
AS INSTALLED
SYSTEM CURVE
0
4
BHP:
Outlet Vel.:
6.83
2,093
Density:
0.075
8
MEASURED
CFM & HP
2.5
1.7
1,196
5
SP REQ'D. FOR
DESIGN FLOW
DESIGN CFM & SP
4
3
AS DESIGNED
SYSTEM CURVE
2
H O R S E P O W E R
4.0
0.5
6
C U R V E
15,000
SP:
RPM:
B R A K E
S T A T I C
P R E S S U R E ( I N. W. G.)
P E R F O R M A N C E
CFM:
1
8
12
16
20
C F M ( i n 1, 0 0 0 's)
24
28
0
CURVE #1
Fan Engineering FE-100
Case Study 2
System Effects Due to Spinning Air and an
Obstructed Inlet
Two plenum fans were selected to operate in parallel to
deliver 80,000 CFM (40,000 CFM each) at a static pressure of 1 in. w.g. The system installation is at 5,200 feet
above sea level and handles room temperature air.
Figure 18 is a sketch of the duct system and Figure 19
shows the performance curve. Air balance test reports
indicated that the airflow was approximately 31,000 CFM
through each fan, with a static pressure of 1 in. w.g.
Because of the high elevation, air density corrections
were made to all static pressure and velocity pressure
readings. Notice that the density shown in Figure 19 is
at the high elevation.
An investigation of the duct system revealed two
problems. One was that the duct geometry caused the
air in the vertical duct section below the fan inlets to
spin in the direction of wheel rotation. The streamlines
sketched in Figure 18a show the general pattern of the
airflow.
The other problem was airflow monitoring probes
mounted in the throats of the fan inlets. The throat of
a fan inlet is one of the most performance-sensitive
parts of a centrifugal fan. The velocity in the throat can
be 10,000 fpm or higher, so obstructions can result in
significant pressure losses. In addition, subtle changes
in the velocity profile at the throat can have significant
effects on fan performance.
To correct the problem with the spin, flow straighteners were added at the lower portion of the vertical duct
sections. These straighteners eliminated most of the spin
and as a result the airflow increased to 34,000 CFM.
The next step was to remove the probes from the
inlets. When this was done, the performance increased
to 41,000 CFM with a static pressure of 1 in. w.g.
Removing the probes increased the flow rate by 25%!
Another thing to be careful with when measuring the
performance of plenum fans is the location of the
static pressure tap in the fan outlet plenum. The high
velocity exiting the fan impeller swirls around in the
plenum. To get an accurate static pressure reading, you
need to find a “dead spot” which is not affected by the
swirling, turbulent air. Typically, calm areas are found in
the corners, but it may take some searching to find a
good spot. In this case, the static pressure in the plenum varied by 20% depending on location.
Notice that the last test point is above the catalog
performance. This could be due to test error, but it also
could be real. The performance for this fan is based on
tests done on a size 365. Typically, larger fans perform
better than smaller fans.
Figure 18. Duct System
SPIN
SPIN
Figure 18a.
SECTION A-A
AIRFLOW PATTERN BEFORE STRAIGHTENER WAS ADDED
ADDED FLOW STRAIGHTENERS
Figure 18b.
SECTION A-A
LOCATION OF PROBES
ROOF
Figure 18c.
A
A
ADDED FLOW STRAIGHTENERS
PLENUM
OPENING
PLENUM
OPENING
ELEVATION VIEW
Figure 19. Performance Curves With and Without Straighteners
Customer:
Job Name:
Fan Tag:
Model: Plenum Fan
32
SP
3.5
28
3.0
24
20
2.5
BHP
16
2.0
WITH STRAIGHTENERS
WITH PROBES IN INLET
1.5
12
WITHOUT STRAIGHTENERS
WITH PROBES IN INLET
1.0
BHP:
Outlet Vel.:
WITH STRAIGHTENERS
WITH PROBES REMOVED
0.5
8
Density:
870
17.44
N/A
0.0619
CORRECTED FOR
ALTITUDE 5,200
H O R S E P O W E R
4.0
RPM:
C U R V E
40,000
1
B R A K E
S T A T I C
P R E S S U R E ( I N. W. G.)
P E R F O R M A N C E
CFM:
SP:
4
SYSTEM CURVE
0
7
0
10
20
30
C F M ( i n
40
1, 0 0 0 's)
50
60
0
CURVE #2
Fan Engineering FE-100
Case Study 3
Air Handling Unit Factory Test Low on Air
An air handling unit manufacturer performed a factory air
performance test on one of their units utilizing a DWDI
fan. A short length of duct, as shown in the figure, was
added to the outlet of the unit to determine the flow
rate using a pitot tube traverse. The dimensions of the
air handling unit outlet and the duct section match the
dimensions of the fan outlet.
The fan was selected to deliver 20,000 CFM at 4 in.
w.g. static pressure. Restriction was added to the outlet
of the duct by closing a control damper until the pressure rise across the fan was 4 in. w.g. and a pitot tube
traverse was performed.
The results of the pitot tube traverse indicated that
the flow rate was 16,070 CFM. The air handling unit
manufacturer called the fan manufacturer, upset that the
fan was 20% low on flow rate.
Key to solving this problem is to remember that fans
and duct systems interact along system lines and there
are tolerances associated with field tests and catalog
performance.
The curve shows the catalog performance and the
measured test point. Also shown are system lines drawn
through the design point and the measured point.
The difference between the measured performance
and the point where the “as tested” system line intersects the catalog curve is 3%. This falls within AMCA’s
certified ratings program check test tolerance which
takes into account manufacturing variations and laboratory test accuracy. A portion of the difference could be
due to inaccuracies associated with the test setup. The
AMCA laboratory test code requires 10 equivalent duct
diameters and a flow straightener between the fan outlet and the pitot tube traverse plane. Since this test
setup does not have this there may be more inaccuracies in the test data.
In either case, the first fan law tells us that by
increasing the fan speed by 3% the measured test point
will move to the catalog curve. After the increase in fan
speed, the duct restriction was readjusted to maintain
the 4 in. w.g. across the fan. The flow rate was
rechecked and found to be very close to 20,000 CFM.
A 3% increase in fan speed along with an adjustment
in system resistance resulted in an increase in flow rate
of 20%. This case study demonstrates that comparing
tested performance with cataloged performance along a
constant pressure line can make fan performance appear
to be a lot worse than it really is.
Figure 20. Air Handling Unit With Added Duct
PITOT TUBE
TRAVERSE PLANE
CONTROL
DAMPER
4"
Figure 21. Air Handling Unit Measured and Expected Performance Curves
Customer:
Job Name:
Fan Tag:
Model: DWDI Fan
3%
5
20
DESIGN
PERFORMANCE
MEASURED
PERFORMANCE
BHP
SP
4
16
DESIGN
SYSTEM
CURVE
3
12
2
8
AS TESTED
SYSTEM CURVE
4
1
0
5
10
15
20
C F M ( i n
25
30
1, 0 0 0 's)
35
40
4
1,038
BHP:
Outlet Vel.:
16.35
1,775
Density:
0.075
H O R S E P O W E R
24
6
0
8
C U R V E
20,000
SP:
RPM:
B R A K E
S T A T I C
P R E S S U R E ( I N. W. G.)
P E R F O R M A N C E
CFM:
0
CURVE #3
Fan Engineering FE-100
Case Study 4
System Effect Due to Fan Mounted at the
Discharge of a Cyclone Dust Collector
A backward curved industrial fan was selected to deliver 15,000 CFM at a static pressure of 18 in. w.g. on a
dust collection system that used a cyclone as an air
cleaner. A field performance test indicated that the
actual performance was 12,800 CFM, 13.1 in. w.g., and
39.4 BHP.
Cyclones clean by spinning the air and using the
centrifugal force acting on the dust particles to separate
them from the air. The air usually enters tangentially at
the top of the cyclone, spins in one direction down
along the outside, and then returns up and out the top
of the cyclone, spinning in the opposite direction. The
dust is flung to the outside of the cyclone walls, and
slides down to where it is collected at the bottom of
the cyclone.
The spinning air exiting the cyclone affects the fan
performance. In this case, the spin was in the direction
of wheel rotation, lowering the static pressure, flow rate,
and horsepower. If the spin were opposite wheel rotation, static pressure and flow rate would increase with
a dramatic increase in horsepower and reduction of fan
efficiency.
To correct the problem, straighteners were added to
the inner cylinder at the top of the cyclone. They prevented the spinning air from entering the fan and
allowed it to operate closer to its cataloged performance
level.
After installation of the straighteners, the performance
was measured again and found to be 14,500 CFM, 17.7
in. w.g. and 53 BHP. Notice that the measured flow rate
and static pressure are above the system line defined
by the original test point. This is due to the added
resistance of the straightener. This point is also slightly
below the cataloged performance curve, but certainly
within an acceptable tolerance for a field test. The
horsepower point is above the cataloged curve. This is
likely due to belt drive losses and inaccuracies in determining horsepower from amp and voltage data.
Figure 22. Cyclone Flow Patterns
FLOW PATTERNS
AIR
A
A
FLOW
SECT. A-A
BEFORE
FLOW
STRAIGHTENER
SECT. A-A
AFTER
Figure 23. Cyclone Performance Curve With and Without Straightener
Customer:
Job Name:
Fan Tag:
Model: Backward Curved Industrial Fan
C U R V E
20
60
BHP
MEASURED CFM & HP
SP WITH STRAIGHTENER
50
MEASURED CFM & SP
WITH STRAIGHTENER
MEASURED CFM & HP
WITHOUT STRAIGHTENER
16
40
12
30
MEASURED
CFM & SP
WITHOUT
STRAIGHTENER
8
4
0
20
10
SYSTEM CURVE
0
4
8
16
12
C F M ( i n
24
20
1, 0 0 0 's)
28
32
RPM:
1,789
BHP:
Outlet Vel.:
52.8
2,336
Density:
0.075
H O R S E P O W E R
24
15,000
18
B R A K E
S T A T I C
P R E S S U R E ( I N. W. G.)
P E R F O R M A N C E
CFM:
SP:
0
CURVE #4
9
Fan Engineering FE-100
Case Study 5
Industrial Fan Without an Inlet Bell
that all of the points fall on one system resistance line
since the duct system remained constant and we were
only modifying the fan.
As a final test, a new screen was made and mounted closer to the inlet of the bell mouth where the air
velocity is much lower. This screen had an immeasurable
effect on fan performance.
An industrial fan was selected to deliver 4,600 CFM at
a static pressure of 18.8 in. w.g. and ordered with a
plain inlet and an inlet screen. The fan was installed on
a system with an open inlet and ductwork connected on
the outlet. A performance test indicated that the actual
performance was 3,988 CFM at 14.8 in. w.g. static pressure.
The key to solving this problem is to see how the
fan was tested for its catalog ratings. Industrial fans are
frequently used in material or dust conveying systems
and are designed with relatively small inlets so that
conveying velocities are maintained. This reduces the
effective open inlet and fan performance drops. When
an inlet bell was added to the fan and the screen was
left in place, the performance increased to 4,329 CFM
at 16.7 in. w.g. With a ducted inlet or with a bell mouth,
the velocity profile at the fan inlet is uniform, as shown
in the figure below.
This is still below the expected performance level and
is due to the screen located at the fan inlet. The screen
reduces performance similar to the probes described in
Case Study 2. When the screen was removed the performance increased to 4,503 CFM and 18.1 in. w.g.
Figure 25 shows the measured performance points, the
catalog performance curve, and a dashed line that
shows the AMCA check test tolerance. Performance
points that lie above the tolerance curve meet AMCA’s
check test tolerance. As you can see, the last point is
within AMCA’s tolerance on the catalog curve. Notice
Figure 24a.With Open Inlet
Figure 24b.With Inlet Bell
Figure 25. Performance Curves With and Without Inlet Bell
Customer:
Job Name:
Fan Tag:
Model: Industrial Fan
S T A T I C
P R E S S U R E ( I N. W. G.)
P E R F O R M A N C E
30
CATALOG SP
25
BHP:
Outlet Vel.:
23.4
6,970
Density:
0.075
MEASURED
CFM & SP
WITHOUT
BELL WITH
SCREEN
10
0
18.8
3,450
MEASURED CFM & SP
WITH INLET BELL
WITHOUT SCREEN
MEASURED CFM & SP
WITH INLET BELL
WITH SCREEN
25
4,600
SP:
RPM:
DESIGN
POINT
20
5
10
C U R V E
CFM:
SYSTEM CURVE
0
1
2
3
4
C F M ( i n
AMCA TOLERANCE
5
6
1, 0 0 0 's)
7
8
9
CURVE #5
Fan Engineering FE-100
Air Performance Troubleshooting Guide
START
Obtain all job info.:
• Detailed description of problem or issue
• Customer contact #, email address
• TCF SO#, Fan, Arrg, special width, dia.,
etc.
• Drive and motor details
• Generate fan curve with design data or
get fan curve from file
• Get drawings and pictures of installation
• Make sure it is a decent selection
Field Measurements:
• Superpose field measurement
data (if any) - verify units (KW
vs. HP, etc.
• Ensure amp measurement is not
thru VFD
• Determine system resistance
curve and performance shortfall
• Determine velocity fluctuations at
fixed measurement location
along with minimum velocity
measured (not less than 400
fpm)
Basic Checks:
• Fan dimensions (OR, IR, overlap)
• Rotation and leaks
• Flex connectors collapsed in fan's inlet
• Wall clearance ˜ 0.5 D (plug or plenum fans)
• Clearance between fans in parallel ˜ D, where D = Blade OD
• Note if floor mount or centrally located in plenum
• Tip clearance, blade angle on axial fans
• Remove any inlet obstructions like Volu Probe, screens, belt guards, etc.
• Open inlet/discharge dampers fully
• Note any restricted discharge or inlet conditions that can magnify system effect (elbows on
inlet or outlet)
• If parallel fans, were they started and ramped to speed simultaneously (for fans operating
near peak pressure)?
• Does system resistance curve match up for 1 fan vs. 2 fans in parallel for flow, pressure
measurements?
• Are VAV box controls disabled?
A
11
No
Are results
within
tolerance by
increasing
speed or
lowering
system
tolerance?
Yes
STOP
Fan Engineering FE-100
Air Performance Troubleshooting Guide (cont'd.)
A
Piezometer Ring Measurements (refer to
published literature ES-105):
• Install a minimum of two piezo taps at
inlet funnel throat 90o apart - prefer
four taps for average
• Locate high pressure tap on front panel
(if plenum fan), housing side (if housed
fan), or in duct (if ducted inlet fan)
Are piezo
results
within
tolerance
requirements?
Yes
STOP
No
Other potential remedies:
• Inlet vortex breaker
• Inlet baffles/Vanes for uniform flow into
fan inlet
• Reselect larger dia. fan to reduce inlet
velocities and ensure fitup, improve
efficiency...
• etc.
Twin city fan & blower | www.tcf.com
5959 Trenton Lane N | Minneapolis, MN 55442 | Phone: 763-551-7600 | Fax: 763-551-7601
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement