Axial Fan Design Using Multi-Element Airfoils to Minimize

Noise
Hurtado, Mark1
Virginia Tech
635 Prices Fork Road, 445 Goodwin Hall, Blacksburg, VA 24061

Burdisso, Ricardo2
Virginia Tech
635 Prices Fork Road, 445 Goodwin Hall, Blacksburg, VA 24061

ABSTRACT
Axial fans are one of the most harmful sources of noise due to their close proximity
to occupied areas and widespread use. The prolonged exposure to hazardous noise
levels can lead to noise-induced hearing loss. Consequently, there is a critical need
to reduce noise levels from ventilation fans. Since fan noise scales with the 4-6th
power of the fan tip speed, minimizing the fan tip speed is an effective method to
reduce fan noise. However, reducing the fan speed results in a decrease in
aerodynamic performance. To this end, the present study uses multi-element airfoils
to increase the aerodynamic characteristics of the fan blades to enable lower fan
speeds and noise relative to fan blades with single element airfoils. Additionally, a
control vortex design method has been implemented to increase the effectiveness of
the blade outer sections, i.e. spanwise varying axial flow. The resulting blade
geometry has been shown to reduce noise levels while maintaining the same
volumetric flow rate.

Keywords: Multi-element, Fan, Noise

I-INCE Classification of Subject Number: 30

1. INTRODUCTION
Axial ventilation fans are used extensively to control the temperature, humidity,
and to remove and dilute contaminants from occupied areas. However, their close
proximity to occupied areas and widespread use are a significant threat to the immediate
and long term health and hearing. Long term exposure to harmful noise levels can lead to
irreversible hearing damage, i.e. noise-induced hearing loss (NIHL). Consequently, the
National Institute for Occupational Safety and Health (NIOSH) recommends an exposure
limit of 85 dBA for an 8-hour time-weighted average1. The short term exposure to high
noise levels can create physical and psychological stress, reduce work efficiency, and
contribute to workplace accidents and injuries from difficulty to hear warning signals2.
Consequently, there is a critical need to reduce noise form ventilation fans.
______________________________
1
[email protected]
2
[email protected]
 Ventilation fans are often loud and have poor aerodynamic performance due to
poor design features e.g. flat plates of constant chord blades, support struts too close to
the fan and so forth. The main driver in the design and fabrication of most ventilation fans
is lower cost. Therefore, there are opportunities to significantly reduce fan noise by better
design of the fan blades. Since fan noise scales with the 4 to 6th power of the fan tip
speed3, an effective approach to reducing fan noise is to reduce the fan tip speed, i.e.
reducing the fan tip speed by half can reduce noise levels by up to 18dB. However,
decreasing the fan tip speed typically corresponds to a decrease in the fan aerodynamic
efficiency, i.e. the design requires a trade-off between the aerodynamic efficiency and the
fan noise. Hence, the approach here is to reduce the fan tip speed while maintaining the
same volumetric flow rate by increasing the aerodynamic characteristics of the fan blades.
To this end, the blade spanwise chord and twist distribution are designed to maximize the
volumetric flow rate contribution of the outer radii, i.e. the axial flow velocity increases
from the fan hub to the tip. However, a non-uniform spanwise axial flow is susceptible to
radially outward flow that increases the fan tip losses4. Consequently, radial equilibrium
is implemented into the design process using a control vortex design (CVD) approach.
This allows for a reduction of the fan tip speed and noise while reducing the complexity
of the blade, i.e. the chord and twist distribution are reduced along the blade5. The
resulting blade geometry is ideally suited to implement multi-element airfoils that can
further improve the aerodynamic characteristics of the fan blades. Multi-element airfoil
configurations outperform single element airfoils by preventing flow separation and
subsequent stall6 that lead to higher lift characteristics. This is possible due to the
accelerated flow in the gap of the multi-element airfoil that increases the kinetic energy
of the flow on the suction surface of the subsequent airfoil7. Consequently, the use of
multi-element airfoils in a tandem configuration has been previously investigated to
improve the aerodynamic characteristics of compressor blades8,9,10. However, to the best
of our knowledge, they have not been used for the design of commercial portable low-
pressure axial ventilation fans. To this end, this paper demonstrates the use of multi-
element airfoils as a promising solution to increase the aerodynamic characteristics of the
fan blades to enable lower fan speeds and noise while maintaining the same volumetric
flow rate. The multi-element airfoil blade design presented here represents a “proof of
concept” design. Future work will focus on validating the results experimentally and
optimizing the multi-element airfoil geometry.

2. AIRFOIL DESIGN
A method for improving the aerodynamic performance of the blades is to
implement high lift airfoil geometries. However, most airfoils specifically designed for
high lift have been designed to operate at high Reynolds numbers11,12. At low Reynolds
numbers (105 < Re < 106) these high lift airfoils show a dramatic decrease in the maximum
lift coefficient that does not make them suitable airfoils to use for the design of the blades.
However, some high lift airfoils such as the S1223 have been designed to operate at low
Reynolds numbers. The airfoil was designed by Selig et al. by making use of a concave
pressure recovery with aft loading that delays separation of the turbulent boundary until
it is close to the trailing edge13. The resulting airfoil geometry ensued a complex shape
with a sharp trailing edge that is difficult to fabricate. Consequently, the use of multi-
element airfoils to achieve high lift has become increasingly more popular for UAV14 and
sail boat15 applications which operate at low Reynolds numbers. However, the available
literature of high lift multi-element airfoil configurations at low Reynolds numbers is
limited. To this end, the approach here is to conduct a trade study to design the multi-
element airfoil configuration at a chord Reynolds number of 200,000 using the
viscous/inviscid MSES/MSIS16 solver. The airfoil geometry used for this study is
restricted to the E214 airfoil geometry which has been shown to have good aerodynamic
characteristic at low Reynolds numbers5,17. Additionally, the number of airfoils is limited
to two. Consequently, hereafter the multi-element airfoil will be referred simply as a
tandem airfoil.

2.1 Airfoil trade study

The design parameters for the tandem airfoil used in the trade study are shown in
Figure 1. They are the chord ratio (c2/c1), airfoil angles (𝛿1 and 𝛿2 ), and gap size (defined
when 𝛿1 = 𝛿1 = 0). A total of 21 different airfoil configurations with different
parameters were investigated in the trade study. The selection of the best airfoil
configuration was based on (a) a high lift coefficient for angles of attack between 0-10
deg. and (b) a high lift to drag ratio at lift coefficients greater than 1.5. The geometric and
aerodynamic characteristics of the best resulting airfoil are shown in Table 1 for a chord
Reynolds number of 200,0000.

gap

c1 c2
𝛿1 𝛿2
Figure 1: Design parameters for the tandem airfoil geometry.

Table 1: Geometric and aerodynamic characteristics of the tandem airfoil design.

Parameter
Airfoil geometry E214 airfoil
c2 0.88
c1
Leading airfoil angle (𝛿1 ) 16.4 deg
Trailing airfoil angle (𝛿2 ) 13.9 deg
Gap size 0
Maximum lift coefficient 1.96
Maximum lift to drag ratio 73.5

The coefficient of lift and drag of the tandem airfoil are compared to a single
element airfoil with the same chord in Figure 2. Here it can be illustrated that the tandem
airfoil results in an increase of lift while also increasing the drag.
(a) (b)

Figure 2: (a) Lift and (b) drag polar for the tandem airfoil and single element airfoil for a chord Reynolds number of
200,000.
 Furthermore, Table 2 presents a comparison of the single and tandem airfoil at
their maximum lift to drag ratio, i.e. 77 and 75 respectively. As illustrated here, at this
point the tandem airfoil increases the lift coefficient by 132% while also increasing the
drag by 140%. However, the penalty caused by the increase in drag is more than offset
by the gain in the lift coefficient since the fan noise scales with the 4-6th power of the fan
tip speed.
Table 2: Tandem and single element airfoil characteristics at the maximum lift to drag ratio.

Airfoil AOA CL CD
E214-airfoil 8.25 1.80 0.024
Tandem airfoil 2 0.77 0.01

3. FAN DESIGN
A comparison between three fan designs is presented in this section to investigate
the use of tandem airfoils in the blade design. Two of the fans were designed using the
control vortex design methodology described by Hurtado et al.5 using a single element
airfoil and a tandem airfoil. The third fan is a baseline fan designed using the conventional
free vortex design and a single element airfoil. Therefore, the three fans will be referred
to as FVS (Free-vortex single), CVS (Control-vortex single), and CVT (Control-vortex
Tandem). Here the first two letters indicate the swirl velocity distribution (free vortex or
control vortex) while the last letter indicates the airfoil geometry (S for single element
airfoil and T for tandem airfoil). The ventilation fans have been designed to generate a
target volumetric flow rate of 1030 CFM. To this end, a multi-objective Genetic
Algorithm (GA) has been implemented to design the velocity profile that satisfies the
volumetric flow rate requirement. Next, an inverse design is implemented that uses the
classical blade-element/vortex formulation to design the blade geometry that results in
the desired velocity profiles. Lastly, the broadband noise of the three fans is computed
using the semi-empirical model developed by Mugridge and Morfey18.

3.1 Fan velocity profile design

The design of the velocity profile involves finding the hub-to-tip ratio ( v ) and the
tip tangential velocity (Mtan) that result in the target volumetric flow rate. The tip
tangential velocity is
r
M tan  t (1)
cair
where  is the angular velocity, rt is the tip radius, and cair is the speed of sound. The
hub-to-tip ratio defined as
r
v h (2)
rt
where rh is the hub radius. Additionally, the velocity distribution is also an important
parameter as it can be used to further improve the aerodynamic characteristics of the
blades. To this end, the velocity profile is designed to increase from the hub to the tip
while maintaining radial equilibrium using the control vortex design methodology
presented by Hurtado, et al.5. Here, radial equilibrium is incorporated into the design of
the velocity profile by solving the radial equilibrium equation assuming a swirl velocity
power law distribution defined as
v (r )  ar n (3)
where a is the swirl velocity coefficient and n is the swirl velocity exponent. The resulting
axial velocity distribution that maintains radial equilibrium is

va (r )  2a(r n 1  rhn 1 )  R(r ) * a 2 (n  1)  vh2 (4)

where r
R (r )  ln( ) 2 for n  0
rh

r 2 n  rh2 n
R(r )  for n  0
n
The velocity at the hub is defined as vh  rh 0.6 which is the minimum velocity while
maintaining a high aerodynamic efficiency19.
Therefore, the aim of the design of the velocity profile is to define the hub-to-tip
ratio ( v ), the tip tangential velocity (Mtan), and the velocity distribution (swirl velocity
coefficient a and exponent n). To this end, a Multi-objective Genetic algorithm (GA) from
the MATLAB toolbox20 has been implemented to investigate the trade-offs between hub-
to-tip ratio and tip tangential velocity for a target volumetric flow rate of 1030 CFM.
Hence the objective functions to be minimized are the tip tangential velocity (Mtan), and
the hub-to-tip ratio ( v ) defined in Equation (1) and Equation (2) respectively. Lastly, a
constraint function is used to limit the solutions to velocity distributions, i.e. swirl velocity
coefficient a and exponent n, that result in the target volumetric flow rate of 1030 CFM.
The non-dominating set of feasible solutions (Pareto front) are presented in Figure 3. Here
the minimum tip tangential velocity for a range of hub-to-tip ratios that result in the target
volumetric flow rate is shown. It can be illustrated here that as the hub-to-tip ratio
decreases, the required tip tangential velocity to generate the target volumetric flow rate
increases.

Pareto Front
Hub/tip
ratio for
design

Figure 3: Multi-objective GA Pareto front.

The hub-to-tip ratio and corresponding tip tangential velocity selected for the
design of the control vortex velocity profiles are 0.405 and 0.1 respectively. Here, the
swirl velocity coefficient and exponent for Equations (3) and (4) are 4.48 and -0.1922
respectively. The resulting axial and swirl velocities are shown in Figure 4. Additionally,
the free vortex axial and swirl velocity profiles to generate the target volumetric flow rate
are also shown in Figure 4. As illustrated here, the velocity distribution using the control
vortex design results in a lower swirl velocity and hence fewer losses near the hub of the
blade.

(a) (b)

Free Vortex
Free Vortex

Control Vortex

Control Vortex

Figure 4: (a) Axial and (b) swirl velocities for the free vortex and control vortex designs.

3.2 Fan geometry design

An inverse design method has been implemented to design the blade chord and
twist distribution for the FVS, CVS, and CVT fans that generate the velocity profiles
shown in Figure 4. The constraints imposed on the design of the blades are (a) a chord at
the hub less than the hub radius, (b) a chord at the tip greater than 40% of the hub radius,
and (c) a twist at the hub less than 85deg. These constraints were imposed to limit the
blade designs to realistic blade geometries that are practical to fabricate. The baseline
FVS fan was designed with a single element airfoil to generate the free vortex axial and
swirl velocity profiles shown in Figure 4. The design fan tip speed and the number of
blades for the FVS fan are 0.125 and 7 respectively. The resulting chord and twist
distribution for the blades of the FVS fan are shown in Figure 5.
(a) (b)

Figure 5: (a) chord and (b) twist distribution for the FVS, CVS, and CVT fan designs.

The CVS and CVT fans were designed to generate the control vortex axial and
swirl velocity profiles shown in Figure 4 with a single element airfoil and tandem airfoil
respectively. The chord and twist distribution for the CVS fan which results in a realistic
blade geometry while generating the control vortex axial and swirl velocity profiles are
shown in Figure 5. The design fan tip speed and number of blades for the CVS fan are
0.115 and 7 respectively. Similarly, the CVT fan was designed with 5 blades and a fan
tip speed of 0.1 to generate the control vortex velocity profiles. The resulting chord and
twist distribution are shown in Figure 5. As illustrated in Figure 5(a), the chord variation
along the blade span is smaller for the CVT fan relative to the CVS fan although it has
fewer blades due to the higher lift coefficient. Additionally, Figure 5(b) shows that the
twist of the CVT blade increases given that it operates at higher angles of attack using the
tandem airfoil configuration. However, the overall change in the twist from the hub to the
tip is the same as that of the CVS fan as illustrated in Figure 5(b). Furthermore, Table 3
shows that the CVT fan results in a 15% decrease in the fan tip speed relative to the CVS
fan while generating the same axial and swirl velocity profiles. Similarly, the CVS fan
design resulted in a 9% decrease in the tip speed relative to the FVS fan. Consequently,
it has been shown that the design of the velocity profile and the use of multi-element
airfoils can be used to enable lower fan speeds and noise while generating the same
volumetric flow rate.

Table 3: Fan design features.

Fan Design Airfoil Velocity Number of Fan tip speed

distribution blades
FVS E214 airfoil Free vortex 7 0.125
CVS E214 airfoil Control 7 0.115
vortex
CVT E214 tandem Control 5 0.1
airfoil vortex

The blade geometry of the three fan designs is shown in Figure 6. It is apparent
here that the low twist and chord variation of the blades designed using the control vortex
design approach is more suitable to implement multi-element airfoils. On the contrary,
the free vortex design approach makes it difficult to implement multi-element airfoils as
a small chord at the tip will result in airfoils with a very small chord that are difficult to
fabricate. Additionally, Figure 6 shows that the tandem airfoil blades of the CVT fan
reduce the total volume of the blades relative to the single element airfoil blades.
Therefore, reducing the amount of material required for fabrication of the blades.
Moreover, the tandem blades reduce the blade thickness noise associated with the
thickness of the blades.
FVS

CVS

CVT

Figure 6: Blade geometry comparison for the FVS, CVS, and CVT blade designs.
 The CAD geometry of the FVS, CVS, and CVT fans is presented in Figure 7.
Additionally, a rotating shroud has been added to the blade geometry to control the tip
clearance noise21. Additionally, Figure 7 illustrates the decrease in the solidity of the fan
when tandem airfoils are used instead of single element airfoil.

(a) (b) (c)

Figure 7: CAD geometry of the (a) FVS, (b) CVS, and (c) CVT fan designs.

The performance of the three fan designs at their respective design speed is
presented in Table 4. As illustrated here, the CVT fan reduces noise levels by 5.6 dB
relative to the baseline FVS fan and 4.2 dB relative to the CVS fan. However, as
illustrated in Table 4 the CVT fan increases the power consumption by 5.7% relative to
the CVS fan. The increase is the CVT fan power consumption is associated with the
increase in the drag of the tandem airfoil relative to the single element airfoil as shown in
Table 2.

Table 4: Fan performance comparison at the fan design speed.

Fan design Fan speed [rpm] Power [watts] Sound power level [dBA]
FVS 4338 158 59.9
CVS 4000 165 58.5

CVT 3500 175 54.3

Additionally, the volumetric flow rate as a function of the fan speed is presented
in Figure 8. As illustrated here, the CVT fan design outperforms the CVS and baseline
FVS fan designs for all fan speeds.

Design volumetric
flow rate

Figure 8: Volumetric flow rate vs fan speed for the FVS, CVS, and CVT fan designs.
 Furthermore, mechanical power consumption is presented in Figure 9 as a
function of the volumetric flow rate. As illustrated here, all fans have a similar power
consumption (within 10%) for all volumetric flow rates shown, i.e. from 200 CFM to the
design volumetric flow rate of 1030 CFM.

Figure 9: Mechanical Power vs Volumetric flow rate for the FVS, CVS, and CVT fan designs.

The sound power level as a function of the volumetric flow rate is presented in
Figure 10. As illustrated here, the CVT fan results in lower noise while generating the
same volumetric flow rate as the CVS and FVS fan designs. Consequently, the use of a
tandem airfoil for the design of the fan blades has been shown to be an effective approach
to significantly reduce noise levels while maintaining the same volumetric flow rate and
similar power consumption.

Figure 10: Sound Power Level vs Volumetric flow rate for the FVS, CVS, and CVT fan designs.

4. CONCLUSIONS
The use of multi-element airfoils has been explored as a method to enable lower
fan speeds and noise. The design of the multi-element airfoil has been accomplished
through a trade study using the viscous/inviscid MSES/MSIS solver at a chord Reynolds
number of 2x105. The resulting airfoil was shown to increase the lift coefficient by 132%
while also increasing the drag by a similar amount. However, the penalty caused by the
increase in drag is more than offset by the gain in the lift coefficient. This is apparent
from the 4.2 dB noise reduction in the fan designed using a multi-element airfoil relative
to a single element airfoil. Additionally, a control vortex design approach was
implemented to further increase the aerodynamic characteristics of the blades. The
resulting blade geometry was shown to be ideally suited to implement multi-element
airfoils. A total noise reduction of 5.6 dB was shown from the combination of the control
vortex design and the multi-element airfoil blade relative to a baseline single element free
vortex blade design. Consequently, it has been shown that the design of the velocity
profile and the use of multi-element airfoils can be used to enable lower fan speeds and
noise while maintaining the same volumetric flow rate and a similar fan power
consumption.

5. ACKNOWLEDGEMENTS
This study was sponsored by the Alpha Foundation (Award Number: AFC215FO-
71) for the Improvement of Mine Safety and Health. The views, opinions and
recommendations expressed herein are solely those of the authors and do not imply any
endorsement by the Alpha Foundation, its Directors and staff.

6. REFERENCES
1. NIOSH, “Occupational noise exposure”, Cincinnati, OH, (1998).
2. X. H. Lin et al., "The Progress of the Study on Fan Noise Reduction Technology",
Advanced Materials Research, Vols. 591-593, pp. 2056-2059, (2012)
3. Cudina, Mirko, "Noise generated by a vane-axial fan with inlet guide vanes", Noise
Control Engineering Journal 39.1 (1992)
4. Janos Vad, and Csaba Horvath, "The Impact of the Vortex Design Method on the Stall
Behavior of Axial Flow Fan and Compressor Rotors", Paper presented at the 2008 ASME
Turbo Expo, June 9, 2008 - June 13, 2008, Berlin, Germany, (2008)
5. Hurtado, Mark, and Ricardo Burdisso, "Low speed control vortex axial fan design for
minimum noise", INTER-NOISE and NOISE-CON Congress and Conference
Proceedings. Vol. 258. No. 7. Institute of Noise Control Engineering (2018)
6. Roy, Bhaskar, et al., "Low speed cascade studies of highly cambered single and tandem
compressor blading", International Journal of Turbo and Jet Engines 12.2 (1995)
7. Roy, Bhaskar, Aditya Mulmule, and Srivatsava Puranam, "Aerodynamic Design of a
Part-Span Tandem Bladed Rotor for Low Speed Axial Compressor", 27th AIAA Applied
Aerodynamics Conference (2009)
8. McGlumphy J, Ng W, Wellborn SR, Kempf S, “3D Numerical Investigation of Tandem
Airfoils for a Core Compressor Rotor”, ASME. J. Turbomach, (2010)
9. Saha, U. K., and Bhaskar Roy, "Experimental investigations on tandem compressor
cascade performance at low speeds", Experimental thermal and fluid science 14.3 (1997)
10. Bammert, K., and H. Beelte, "Investigations of an axial flow compressor with tandem
cascades", Journal of Engineering for Power 102.4 (1980)
11. Maughmer, Mark D., and Dan M. Somers, "Design and experimental results for a
high-altitude, long-endurance airfoil", Journal of Aircraft 26.2 (1989)
12. Althaus, D., and Wortmann, F. X., “Stuttgarter Profilkatalog I”,Vieweg, Brunswick,
Germany, (1981)
13. Selig, Michael S., and James J. Guglielmo, "High-lift low Reynolds number airfoil design",
Journal of aircraft 34.1 (1997)
14. Lakshmi, G. S., P. Balmuralidharan, and G. Sankar, "High Lift Two-Element Airfoil Design
for MALE UAV Using CFD", (2018)
15. Katz, Joseph, "Lift and Drag Measurements of Tandem, Symmetric Airfoils", 31st AIAA
Applied Aerodynamics Conference. 2013.
16. Drela, Mark, "MSES: A multielement airfoil design analysis system”, A Research Program
by: MIT Computational Aerospace Sciences Laboratory (2007)
17. Selig, M. S., J. F. Donovan, and D. B. Fraser, "Airfoils at low speeds, Soartech 8", edited by
HA Stokely, SoarTech Publ., Virginia Beach, VA (1989)
18. Mugridge, B. D., and C. L. Morfey, "Sources of noise in axial flow fans", The Journal
of the Acoustical Society of America 51.5A (1972)
19. Hurtado, Mark Pastor, Daniel Wu, and Ricardo Burdisso, "Design of low speed rim
driven ventilation fan for minimum noise", INTER-NOISE and NOISE-CON Congress
and Conference Proceedings. Vol. 254. No. 2. Institute of Noise Control Engineering,
(2017)
20. Chipperfield, A. J., and P. J. Fleming, "The MATLAB genetic algorithm toolbox",
(1995)
21. Longhouse, R. E, "Control of tip-vortex noise of axial flow fans by rotating shrouds",
Journal of sound and vibration 58.2 (1978)