Conference PaperPDF Available

DEVELOPMENT OF BELL MOUTH FOR LOW SPEED AXIAL FLOW COMPRESSOR TESTING FACILITY

Authors:

Abstract and Figures

A bell mouth is a tapered channel used to properly direct the flow at the intake of any air breathing engine. It does so by gradually converging the walls towards the centre to render a smooth entry into the engine. In this process, a boundary layer gets formed near the walls which grows from the upstream to downstream with reduction in dimensions. The growth of this boundary layer reduces the effective area for fluid to flow and hence mass flow through the intake. Therefore, with a view to get rid of this problem to some extent, the geometry/profile such as a bell mouth (decreasing area section) should be used to achieve the desired mass flow rate at the desired location. This paper aims to guide for designing such decreasing area sections used for low speed axial flow compressor test facility. Different prospective geometries have been modelled and studied for flow field analysis. The goal for this study was to achieve nearly uniform flow at the entry with shorted possible distance for placing axial flow fan in duct. Numerical study using Ansys CFX© was performed to understand flow field with different entry profile shapes. Air sucked from the atmosphere need to be analyzed in terms smooth entry through bell mouth. To achieve this goal two separate flow domains named entry and casing were made. Thicknesses of the boundary layer and the developing length have been identified as the key decision parameters for comparison of performance of different possible bell mouth geometries.
Content may be subject to copyright.
1
Proceedings of the Asian Congress on Gas Turbines
ACGT2016
14-16 November 2016, Indian Institute of Technology Bombay
Mumbai, India
ACGT2016 -18
DEVELOPMENT OF BELL MOUTH FOR LOW SPEED AXIAL FLOW
COMPRESSOR TESTING FACILITY
Apurva Tiwari
Institute of Technology, Nirma University
Ahmedabad, Gujarat, India
Arpit Lad
Institute of Technology, Nirma University
Ahmedabad, Gujarat, India
Sahil Patel
Institute of Technology, Nirma University
Ahmedabad, Gujarat, India
Chetan S. Mistry
Indian Institute of Technology Kharagpur
West Bengal, India
ABSTRACT
A bell mouth is a tapered channel used to properly direct the flow
at the intake of any air breathing engine. It does so by gradually
converging the walls towards the centre to render a smooth entry
into the engine. In this process, a boundary layer gets formed
near the walls which grows from the upstream to downstream
with reduction in dimensions. The growth of this boundary layer
reduces the effective area for fluid to flow and hence mass flow
through the intake. Therefore, with a view to get rid of this
problem to some extent, the geometry/profile such as a bell
mouth (decreasing area section) should be used to achieve the
desired mass flow rate at the desired location. This paper aims to
guide for designing such decreasing area sections used for low
speed axial flow compressor test facility.
Different prospective geometries have been modelled and
studied for flow field analysis. The goal for this study was to
achieve nearly uniform flow at the entry with shorted possible
distance for placing axial flow fan in duct. Numerical study using
Ansys CFX© was performed to understand flow field with
different entry profile shapes. Air sucked from the atmosphere
need to be analyzed in terms smooth entry through bell mouth.
To achieve this goal two separate flow domains named entry and
casing were made. Thicknesses of the boundary layer and the
developing length have been identified as the key decision
parameters for comparison of performance of different possible
bell mouth geometries.
Keywords: Bell Mouth, Jet Engine Testing, Boundary Layer,
Developing Length.
NOMENCLATURE
A0 Angle between horizontal and line joining centres,
degrees
BLT Boundary Layer Thickness, in mm
D1 Diameter of first circle, mm
D2 Diameter of second circle, mm
D3 Diameter of third circle, mm
a Semi-minor axis of ellipse, mm
b Semi-major axis of ellipse, mm
INTRODUCTION:
For development of experimental facility for axial flow
compressor, the designer must select the shape of the bell mouth
that reduces entry losses and provide high quality of flow at the
compressor inlet. The well designed bell mouth shaped entry is
one of the best possible option for the same. The bell mouth
shape facilitates to provide smooth entry of flow, reduces
downstream boundary layer thickness, reduces flow non-
uniformities and produce an even velocity profile at the
compressor rotor entry. Bell shaped intakes are used for several
applications such as jet engines, I.C. engines, industrial fans,
pumps, compressors, blowers, intake manifold where any kind
2
of fluid has to be sucked in for further use with avoidance of
unwanted disturbances, vibrations and pressure losses.
For any suction device to perform its fullest capacity, it must
be able to overcome the loss in the mass flow incurred while
sucking in air from any source. The loss in flow is because of
growth in boundary layer thickness and thus reduced effective
flow area. Sometimes, due to space constrain or design
requirements, the size of inlet duct need to be reduced in
dimensions. In such cases, the developing length of the flow in
the inlet needs to be reduced to achieve desired flow quality with
minimum possible distance. Thus, the shape and size of the bell
mouth shall play a vital role in improvement in the performance
of the engine.
Bell mouth is an important component used for intake of any
air breathing engines and has direct influence on performance of
engine. It is not been studied as an independent component in
detail so far as open literature concern. Blair et al. [1, 2] explored
the possibilities to optimize the shape of bell mouth used for I.C.
engine intake application. Based on the experimental and
numerical study, they recommended elliptical shaped bell mouth
performs better for their application. Ismael AR et al. [3] used
the bell mouth study for their diesel engine application and found
improvement in performance by 10% as compare to non-Bell
mouth intake of engine.
Rotating components like fans, compressors, turbines,
pumps etc. are running under specific operating conditions. This
puts change in operating conditions in terms of change of speed,
change of mass flow rate etc. The changes in such operating
parameters changes the growth of boundary layer along the
casing, tip vortex flow across the surfaces due to tip clearance
etc. The change in upstream flow conditions through entry makes
the flow behaviour more complex added to more 3-D flow
through the whole fluid flow passage. This has direct impact on
overall operating of the rotating components and possibly may
lead to deteriorated performance and drop off pressure. The
length of intake duct also plays an important role in terms of
space constraints, boundary layer growth and losses because of
wall friction.
Most of the attempts in this field of turbomachines
applications are in the form of analysis of input flow to fans or
jet engines. The efforts made by Andersen [4], Briley [5] and
Levy [6, 7] illustrated a clear pattern of merits and demerits of
the use of various Numerical methods in analysing the flow
stream inside an intake. Vakili [8] has compared the results of
numerical tests with experimental models.
Towne C.E. [9] analyzed the flow in the subsonic diffuser
section of a typical modern inlet design. They studied the effect
of curvature of the diffuser centerline and transitioning cross-
sections to determine the primary cause of flow distortion in the
duct. Their conclusion was based on reports of total pressure
values in the engine compressor face.
Anand [10] explored study of 3 different profiles, i.e.
circular, aerofoil and ellipse for jet engine inlet design. Changes
in co-efficient of discharge and mass flow rate were compared
with respect to changes in the pressure ratio. He found 3.5% co-
efficient of discharge over simplest bell mouth and consequently,
the elliptical profile was found to be the most efficient among all
other shapes.
Son et al [11] numerically studied the effects of bell mouth
geometries on the flow rate of centrifugal blowers with different
bell mouth geometries. They found the bell mouth radius had a
major effect on the flow rate of the centrifugal blower. Too small
radius would result in a noticeable loss in the flow rate because
of the appearance of a vortex under the bell mouth region.
For internal combustion engines intake, Blair et al [1, 2] made
extensive studies on the optimum shape of a bell-mouth air
intake through numerical and experimental procedure. Based on
numerous findings, they recommended an elliptical shape as the
best profile to utilize in bell-moth intakes.
The present study is related to design of bell mouth for
low speed axial flow compressor fan testing facility. The
experimental setup is planned to design with space constrain.
Three possible bell mouth shapes were explored to meet specific
requirement. Different methods named 3 circle method, aerofoil
shaped and elliptical shaped bell mouth geometries were
explored for application. The details numerical studies were
conducted with optimized dimensions using above three
methods. The parameter called boundary layer thickness and
fully developed flow is considered as a parameter for final
selection of bell mouth shape.
It is believed that the paper gives idea to future designers for
geometrical selection of various bell mouth geometrical
parameters and shapes for low speed testing facilities.
DESIGN PROCESS:
As it is known, bell mouth shape is a tapered channel with
decreasing dimensions downstream. To make the decreasing
section from atmospheric inlet to casing outlet, there are different
possibly to design the shape. In present study to meet the specific
requirements, three different design approaches were explored,
viz.:
a. 3 Circles profile
b. Aerofoil profile
c. Elliptical profile
(a) 3 Circles Profile:
In this method, the bell mouth profile was generated using a
tangential path to three individual circles. There are different
possible combinations for the same. Either by changing the
position of these three circles besides each other or change in
their respective radii gives the space to generate distinct profiles
for bell mouth.
As shown in Fig. 1, three circles are connected tangentially
to lead to a single continuous and smooth profile.
3
Fig. 1: Profile constructed by three circles
approach.
The primary parameters to control the shape of the profile are:
i. The diameters of the circles (D1,D2,D3)
ii. The inclination of the line joining them. (A0)
For present study, changes in the flow with variation in these
parameters have been explored to understand effects of changes
in these two parameters. This method finds difficulties:
1. The circles cannot be joined at every inclination of the
line to obtain a smooth curve,
2. Inclination of line.
Hence, this method needs selection for any arbitrary higher
inclination help in reducing the developing length. This incline
line will give higher dimensions for inlet diameter in the bell
mouth at the entry. The parametric study for the same is
explained in detail in the results and discussion section later.
(b) Aerofoil Profile:
In this approach, NACA aerofoils used to construct a bell
shaped inlet profile as leading edge dimensions of aerofoil. The
entire aerofoil profile is divided into parts such that the curvature
of the upper half near the leading edge forms the required shape
of bell mouth entry. The governing parameters which affect the
shape of the bell mouth are:
1. The amount of camber along the length of the aerofoil
2. Location of maximum camber.
Figure 2 shows variation in % camber, keeping other aerofoil
parameters constant. A high camber has generally been observed
smooth entry to flow with reduced boundary layer develop
length. The higher a value of % camber increase the diameter at
the entry to a very large extent. This puts limit in reduction of
overall dimensions of bell mouth at the entry.
Another important parameter is the location of the
maximum camber along the chord length of aerofoil.
Fig. 2: Selection of % camber for profile
Fig. 3: Location of the maximum thickness (in % of
chord)
Figure 3 shows that the profile essentially gets more thicken as
the location of the maximum thickness is increased. This will be
helpful in smooth guiding of the flow at the inlet. At the same
time it increases the longitudinal dimension of the bell mouth.
Therefore, an optimized value for each of these parameters for
every profile needs to be studied so as to obtain the best possible
4
configuration. The detailed study using the aerofoil approach is
explored in detail in the results and discussions section later.
(c) Elliptical Profile:
The need for smooth entry of flow with dimensions
constrain explore the need for the shape with minimum
constrains in terms of geometrical dimensions. Also the design
approach need to be simple and ease in manufacture. For
elliptical shape is one of the alternative for the same. As elliptical
shape has two limiting parameters,
1. Major axis dimension.
2. Minor axis dimension.
Fig. 4 Profile showing the elliptical region with the
straight throat
One-quarter of the ellipse shape can be consider as a shape for
bell mouth entry as shown in Fig 4. This method gives good
envelope for selection of only two dimensions which explore
possibilities small diameter at the entry and with small axial
length. Careful selection of these parameters will lead to the
requirements in terms of ease of design and manufacturing. The
entry edge is found to be sharp at the extreme ends of the inlet
which led to flow non-uniformities. The curved entry at the sharp
edges will lead the flow to be more uniform. The detailed study
for the same is explored in result and discussion section.
METHODOLOGY:
In order to explore the design envelope of bell-mouth for low
speed facility, the detailed numerical studies were conducted
for all three cases discussed earlier. The systematic numerical
approach was followed for the same.
Meshing:
In the current study, three dimensional geometry was selected so
as to capture the flow around the tip entry region of the bell
mouth. The flow enters from the atmosphere through this tip
region. The major study reported [1,2] considered 2-D geometry
and the flow inlet starts from inlet of bell mouth geometry itself.
This study was not considered with end wall effects and overall
suction of flow through the atmosphere. The end wall near the
edges also plays important role to study smooth entry from direct
atmosphere. Taking this criteria into consideration of actual flow
situation was generated with two separate domains named bell
mouth domain and spherical domain as separated domains as
shown in Fig. 5.
Fig. 5: Meshing of the model
In order to study the effect of various bell mouth shapes on the
flow physics, quarter of the whole approach was selected. This
consideration helps in minimum computational efforts. The
meshing was done using the integral “Meshing” module of
ANSYS© software. The inner volume has been divided into
larger element size than those at the boundaries since the criteria
in this study are to analyze the boundary phenomena near the
walls of bell mouth. Inflation layer option was adopted at the
curvature of the bell profile to facilitate the study of the
formation of boundary layer along the wall of the intake with
very fine mesh. Since the consideration for design of suction is
from open atmosphere, a k-ε turbulence model with low intensity
of 5% has been considered.
Boundary Conditions:
The used boundary conditions for the study are as shown in fig.7.
a. The Walls: The walls of the bell mouth on both sides
(internal and external surfaces), have been given no slip
condition to study viscous flow effects. The walls were
considered as smooth wall throughout the study.
b. Open Boundary: An open boundary was considered for
the spherical domain so that the fluid could be made to
enter the bell mouth symmetrically from all directions.
This feature provides detailed flow physics near the
entry of bell mouth as it captures the flow from large
volume as atmosphere.
5
c.
Outlet: For analysis purpose, an outlet velocity of 45
m/s was considered at the outlet of the bell mouth which
is flow velocity for compressor rotor downstream.
Fig. 7: Boundary Conditions at various locations
RESULTS & DISCUSSIONS:
The main focus for present computational study is to
understand the growth of boundary layer along the wall of bell
mouth. The boundary layer thickness was determined by plotting
the velocity along the lines at various sections of the bell mouth
(Fig. 8) for all the three cases. The plots for velocity along the
sections were plotted. After some location downstream of bell
mouth, there is no further improvements in velocity. The
boundary layer thickness was measured considering the 99.99%
freestream velocity profile. The difference between the distances
shown as the radius and the marked location gave the boundary
layer thickness. These lines are significant as they show the
velocity profile at the respective sections which is crucial to
study the nature of the flow in that region. Moreover, they form
the base as the deciding parameters for comparison between the
3 bell mouth shaped profiles.
Fig. 8: Various lines along which velocity is plotted
As discussed earlier, initially the 3-circle method was consider
for study. Circles of different diameters and inclinations were
tried for development of bell mouth shape. In order to capture
smooth entry of the flow at the mouth, the entry was made
slightly curved in proportion to the smallest diameter. The
comparatively good results are discussed in this paper. Figure 9
shows the velocity contours along the bell mouth shape. At the
entrance of bell mouth the flow is under stagnant condition. The
velocity slowly increases downstream. When it reaches the
converging section, it shows increase in velocity. Further
downstream at particular distance, there is no further increase of
velocity. The growth of boundary layer can be clearly seen from
the wall along the length. With initial profile, the boundary layer
thickness calculated as discussed earlier was 10 mm at a distance
of 500 mm from the entrance where the flow is fully developed.
The flow phenomenon can also be understood with pressure
contours as the entry section of the bell mouth act as a nozzle to
the flow. Further downstream where the velocity becomes
constant, the static pressure also becomes uniform.
Fig. 9: Velocity contour of initial 3 Circles Profile
Fig. 10: Pressure Contour of initial 3 Circles Profile
The requirement of the bell mouth is to achieve fully devloped
flow with minimum possible distance downstream so that the
required length of the intake is smaller. The velocity countours
near the curved entry region shows the flow is not attached
smothly at that region as shown in fig. 9 with the circle. This
indicates the possibilities for further improvement in geometry.
6
In order to overcome the difficulty the diameters of the first 2
circles (D
1
, D
2
), the distance between the centres (L) and the
angle of the line joining the centres was increased with respect
to the horizontal (A0).
A
number of geometries were studied and
the best results out of all trials for 3-circle method are as shown
in Fig. 11 and 12. The selction of all mentioned parameters gives
larger diameter at the entry section. For present study the
atmosphere domain size and shape was consider to be same to
enable comparative study.
Figure 11 shows velocity countours with modified geometry. It
shows that, with the intended change of geometry, the flow is
getting attached smoothly all the way from entry. With steep
decrease of diameter the flow gets acceleration. It is inetersting
to observe steep decrease of diameter acceleartes the flow down
stream but the flow is getting fully devloped in later part
downstream. This lead to the reduction of boundary layer
thickness to 6 mm compared to 10 mm for earlier case. This
geometry increase the distance at which flow is fully devloped
and is 650 mm compared to 500 mm from entrance. So, from the
BLT perspective, a steeper intake proves to be better but at the
same time, it also increases th edeveloping length.
Fig. 11: Velocity Contour of final 3 Circles Profile
Fig. 12: Pressure Contour of final 3 Circles Profile
Figure 12 shows the pressure contours for the said case. It shows
the bubble formed immediately after the end of curvature. This
shows the flow is unable to stay attached further downstream.
For flow to become fully developed, it travels a further distance.
The 3-circle method studied for different geometries clearly
indicates the selection of diameters of all three circles and
inclination line joining the center plays important role in shape
development. The target to achieve the fully developed flow
downstream the bell mouth geometry changes the dimensions
which may put constraints in terms of both fabrication
difficulties and cost.
To overcome the said difficulties, the aerofoil shaped profile was
considered for further study. The selection of dimensions are as
discussed earlier. The aerofoil leading edge shape was
considered as profile for the bell mouth. For smooth entry at the
edge of entry it was given smooth curvature. Figure 13 and 14
shows the velocity contours and pressure contours for initially
selected aerofoil shaped bell mouth using NACA 4215 profile.
As seen from Fig. 13, flow is accelerating straight from the edge
of bell shape. The length available for flow acceleration is found
to be lower. As a result, the flow downstream of bell shape is not
giving fully developed flow immediate downstream of bell
shape.
Fig. 13: Velocity Contour of initial Aerofoil Profile
Fig. 14: Pressure Contour of initial Aerofoil Profile
For aerofoil profile, the boundary layer thickness achieved was
6.6 mm at a distance of 390 mm from the entrance. It can be seen
7
from both velocity and pressure contours there is a presence of
flow disturbances as compare to relatively smooth flow. The
prime requirement of smooth fully developed flow at entry was
still missing. Even after achieving better results in terms of
boundary layer thickness and small entry length compare to 3
circle case, the initial aerofoil profile did not satisfy the basic
requirements of maintaining a smooth accelerating flow. So it
was decided to, change the camber and location of max. camber
to meet the requirements.
After number of attempts, the camber was increased and the
distance of max. camber were changed using NACA8212 profile
such that the bell mouth becomes steeper so that rapid
contraction can be obtained and at the same time it can be
verified for not having detached flow. For the selected profile,
the velocity and pressure contours are shown in Figs. 15 and 16
respectively. As seen from the figure with modification of
camber, flow is getting smooth acceleration through
downstream.
The flow non uniformities observed in earlier case vanished. The
selected profile shows the reduction of the boundary layer
thickness to 6 mm and fully developed flow downstream at a
distance of 350 mm.
Fig. 15: Velocity Contour of final Aerofoil Profile
Fig. 16: Pressure Contour of final Aerofoil Profile
It shows that increase of camber and decrease in the location of
maximum camber yielded better results in terms of flow
uniformities downstream of bell shape. This is because the
curvature of the resulting profile is much rapid as also observed
for 3-circle method. Any further increase of the camber or the
location of max camber bring the leading edge closer which will
give a very steep bell shaped entry. This may again lead to non-
uniformities in flow down stream of bell shape. The change also
brings the increase of frontal area of the bell mouth, increasing
its size, so further increase is not desirable.
For both 3-circle method and aerofoil shaped profile, to achieve
the small length with fully devloped flow downstream, it
requires comparatively higher diameter at the entry of bell shape.
This increases the fabrication difficulties and will be costly too.
To have a better, cost effective solution, the elliptical shaped
profile was further explored.
Fig. 17: Velocity Contour of initial Ellipse Profile
Fig. 18: Pressure Contour of initial Ellipse Profile
Figure 17 and 18 shows the velocity and pressure contours for
elliptical shaped bell mouth having a/b ratio of 2. It shows
smooth entry flow from the edge. With comparatively small
diameter at the entry, it gives improvement both with respect to
downstream boundary layer thickness, developing length and the
uniformity of the flow. The flow disturbance at the entry point
8
due to the existence of a sharp edge at the entry is noticeable.
This flow disturbance need to be addressed for further possible
smooth entry flow. For this profile, the boundary layer thickness
was found to be 7 mm with fully devloped flow occuring at 471
mm from the bell mouth entrance.
Fig. 19: Velocity Contour of final Ellipse Profile
Fig. 20: Pressure Contour of final Ellipse Profile
The selection of entry shape is a key factor for smooth entry of
flow with minimum disturbance to the flow. The entry profile
was modified with trials such that smoother and gradual entry
will give fully developed flow with reduced boundary layer
thickness. The developing length also had to be decreased.
Figure 18 and 19 shows the velocity and pressure contours for
such modified profile having a/b ratio of 2.67. It can be clearly
seen from the figure the smooth entry of flow with minimum
entry length for fully devloped flow. The boundary layer
thickness was further reduced to 5.4 mm, at a distance of 342
mm by reducing the dimensions of the semi-major and semi-
minor axes of the ellipse. The semi-major and semi-minor axes
of the final elliptical profile selected were 200 mm and 75 mm
respectively.
Figure 21 shows a comparison between the velocity plots of
respective profiles at the entry section. The 3 circle profile
renders the lowest velocity among the three in the region where
normalized radius = 1. The other profiles have relatively greater
velocities in that region which shows that the others are much
quicker in accelerating the flow initially. Gradually, as the flow
reaches further downstream, following the same trend, the
elliptical profile accelerates it the most, thereby delivering the
lowest developing length as compared to the others. So, it gets
further clarified that the elliptical profile performs better and also
has the least size, i.e. lower frontal area due to the reduced semi-
minor axis and also the axial length was small as compared to
others, the developing length being lower.
It is more interesting to observe for elliptical profile the flow is
highly accelerated compare to both the other cases. Then
gradually develops along the span. This is the indication of small
area available at the entry which facilitates greater acceleration
of flow and flow manages to attain lesser developing length. It
shows the improvement in velocity profile compared to both 3
circles and aerofoil shaped profiles.
Fig. 21: Comparison of various velocity profiles at
entry sections
The graph in Fig. 22 shows a comparison in the velocity profiles
at the location of fully developed flow length for the elliptical
profile. This plot will help in understanding the reason for
achieving the short length using elliptical profile shape. Figure
clearly shows, the both aerofoil and elliptical profile renders the
freestream velocity = 45 m/s but the three circles profile does not
at specific location. The flow is still under development and need
more distance for getting developed. The velocity profile
generated by the aerofoil is very close to achieved by elliptical
profile. This shows that its shape is effective to produce the
desired result from the velocity perspective. But it is important
to notice for similar developing lengths, the dimensions of
elliptical profile are smaller and showing easiness in cost
effective fabrication and this adds to the advantage compare to
aerofoil shaped profile. Hence by various technical and practical
considerations, the elliptical profile suits the desirable
characteristics in terms of achieving goals.
9
Fig. 22: Velocity Profile of all profiles
(at developed length of Elliptical profile)
The comparative study for all the three approaches for bell
profile selection clearly indicates, the elliptical profile performs
the better compared to the other two approaches with suitable
selection of the geometrical parameters and hence was selected
for further study.
MESH SENSITIVITY:
For the selected elliptical profile, mesh sensitivity
analysis was performed to understand the effect of selection of
mesh elements on the boundary layer growth and length of fully
developed flow. To carry out mesh sensitivity analysis, the
number of elements were changed progressively and the
important parameter called boundary layer thickness and the
distance of fully developed flow from entry were measured. The
studied mesh and observed parameters are as shown in the table
below.
.
No. of
Elements BLT
(in mm) Distance from entry of fully
developed flow
(in mm)
10,41,625 5.2 342
13,85,467 5.4 346
17,06,983 5.4 347
11,69,512 5.4 346
The bell mouth was designed for low speed axial flow
compressor setup. In order to carryout performance analysis of
axial flow compressor, it will run with different mass flow
conditions. It became necessary to understand the inflow
conditions through bell mouth under this variable flow
conditions. To carry out this study, the outlet velocity was
changed. The exit velocities were varied in the range of 45 m/s,
35 and 55 m/s. All these changes were applied to the final
elliptical profile and compared as shown below.
Fig. 23: Comparison of velocity profiles at fully
developed length of elliptical profile with varied mass
flow rates
Figure 23 shows the velocity profile at the fully developed flow
region for elliptical profile. The profiles near wall region shows
marginal variation near wall and remains almost constant over
the span. With the increased velocity, the developing length was
decreased by 9.3% and an increase of 7.6 % was observed on
decreased velocity. The boundary layer thickness was found to
be increased by a nominal 0.8 mm in increased velocity and
decreased by 0.3 mm in the decreased velocity case. These
clearly indicates, the elliptical shaped bell mouth performs
effectively for varied flow condition also. Therefore, the profile
is selected for the development of experimental setup from the
flow conditions point of view too.
This paper aims to give basic design idea for design bell mouth
to be used for low speed axial flow compressor setup with
predefined constrains to achieve smooth flow with fully
developed flow within shortest possible length after bell shape.
CONCLUSION:
This paper discusses three different methods for design and
development of bell mouth profiles. Different geometrical
parameters were compared mutually to investigate the flow
physics involved to study the nature of the flow under normal
test conditions mentioned. The purpose of the numerical study is
to achieve best optimize shape and size of the bell mouth for
experimental facility development. Three different approaches
10
were considered for the study. The higher curvature at the entry
improves the performance of the bell mouth with regard to the
lesser boundary layer thickness and developing length achieved
but it also poses a drawback that the overall dimension at the
entry gets increased. Some of the important conclusions drawn
out of the study are as below:
1. Three Circles Method: The profile derived from this
approach has limitation by selection of diameters of
three circle and line joining the center of these circles.
By parametric study it shows the requirements to be
fulfill by selecting the steeper inclination. This steeper
inclination leads to increase the diameter at the entry. It
has complications for manufacturing and is also not
cost effective for specific application.
2. Aerofoil profile: The amount of controlling parameters
for profile development are more. By changing various
aerofoil parameters – camber, position of maximum
camber and thickness, better results can be obtained.
Parametric study showed, bell mouth profile to be better
than the 3 circle profile as it led to much lower
boundary layer thickness and it also reduces the
required length of the bell mouth.
3. Elliptical profile: This method has proven the
advantages in terms of minimum controlling
parameters but still very good control of flow
distribution. The shaped developed using this method
showed the least boundary layer thickness and the
minimum length required for the bell mouth. So far as
the three methods are concerned, the best results under
aerodynamic considerations are the elliptical profile
with the front slightly extended backwards on the same
profile to facilitate smooth entry. Also, such shaped
profile is compact, relatively simpler to manufacture
and cost effective too.
REFERENCES
[1] Blair GP, “Design and simulation of four stroke
engines”, Society of Automotive Engineers, SAE
reference R-186. 1998; 813.
[2] Blair GP, Lau HB, Cartwright A, Ragunathan BD,
Mackey DO, “Coefficients of discharge at the apertures
of engines”, SAE International Off-Highway Meeting,
Milwaukee; 1995
[3] Ismail AR, Bakar RA, Semin , “The effect of discharge
coefficient of intake and exhaust valve for direct
injection engine”, International Conference on
Mechanical Engineering, 2007.
[4] Anderson BH., “Three dimensional viscous design
methodology for advanced technology aircraft
supersonic inlet systems”, NASA 1M 83558, AIAA
Paper 84-0192, 1984.
[5] Briley WR, McDonald H., “Analysis and computation
of viscous subsonic primary and secondary flows”, AIM
Paper 79-1453, 1979.
[6] Levy R., Briley WR., McDonald H., “Viscous
primary/secondary flow analysis for use with non-
orthogonal coordinate systems”, AIAA Paper 83-0556,
1983.
[7] Levy R, McDonald H, Briley WR, Kreskovsky JP., “A
Three-dimensional turbulent compressible subsonic
duct flow analysis for use with constructed coordinate
systems”, AIM Paper 80-1398, 1980.
[8] Vakili A, Wu JM, Hingst WR, Towne CE.,
“Comparison of experimental and computational
compressible flow in an SDuct”, AIM Paper 84-0033,
1984.
[9] Towne CE., “Computation of viscous flow in curved
ducts and comparison with experimental data”, AIM
Paper 84-0531; 1984.
[10] Rishab Anand, “Design a bellmouth intake for a gas
turbine turbo shaft engine- Rolls Royce/Alison 250 C-
10 d”, University of the Highlands and Islands Perth
College UHI.
[11] Pham Ngoc Son, Jaewon Kim and E. Y. Ahn, “Effects
of bell mouth geometries on the flow rate of centrifugal
blowers”, Journal of Mechanical Science and
Technology, 25 (9) (2011) 2267~2276.
... The selection of restrictor with venturi shape based on previous study which found that venturi shape has coefficient of discharge 0.975, higher than orifice shape 0.6. The venturi shape also has the lower pressure loss than the orifice shape 2,16,17) . Modify the convergent and divergent length of venturi restrictor can produce the lower value of pressure loss inside the restrictor 18,19) . ...
... In the case of a bell mouth, if the axial distance at the shroud from the blade LE to the bell mouth is tight, it can significantly affect the inlet flow and the TLF, which directly correlates with the performance [24][25][26]. However, in order to stabilize the inlet flow near the shroud sufficiently [27], the axial distance of the axial fan in this study was secured to be at least 0.4 times the radius of shroud. Although the effects of the hub cap and bell mouth could not be completely excluded, the axial fan for the experimental test was designed to minimize these effects, as validated in the section below. ...
Article
Full-text available
An axial flow fan, which is applied for ventilation in underground spaces such as tunnels, features a medium-large size, and most of the blades go through the casting process in consideration of mass production and cost. In the casting process, post-work related to roughness treatment is essential, and this is a final operation to determine the thickness profile of an airfoil which is designed from the empirical equation. In this study, the effect of the thickness profile of an airfoil on the performance and aerodynamic characteristics of the axial fan was examined through numerical analysis with the commercial code, ANSYS CFX. In order to conduct the sensitivity analysis on the effect of the maximum thickness position for each span on the performance at the design flow rate, the design of experiments (DOE) method was applied with a full factorial design as an additional attempt. The energy loss near the shroud span was confirmed with a quantified value for the tip leakage flow (TLF) rate through the tip clearance, and the trajectory of the TLF was observed on the two-dimensional (2D) coordinates system. The trajectory of the TLF matched well with the tendency of the calculated angle and correlated with the intensity of the turbulence kinetic energy (TKE) distribution. However, a correlation between the TLF rate and TKE could not be established. Meanwhile, the Q-criterion method was applied to specifically initiate the distribution of flow separation and inlet recirculation. The location accompanying the energy loss was mutually confirmed with the axial coordinates. Additionally, the nonuniform blade loading distribution, which was more severe as the maximum thickness position moved toward the leading edge (LE), could be improved significantly as the thickness near the trailing edge (TE) became thinner. The validation for the numerical analysis results was performed through a model-sized experimental test.
Thesis
Full-text available
Cavitation and stall (saddle), which are considered to be unstable phenomena and profound problems in our field, were analyzed along with methodologies for improvement or suppression. The analysis was mainly based on computational fluid dynamics, and each experimental test and fast Fourier transform were conducted in parallel. For cavitation of a mixed-flow pump, the following variables were considered that could affect the local flow pattern or pressure peak at the blade inlet: thickness (blockage), incidence angle, leading edge shape (ellipse ratio), and inlet diameter. The thickness was defined as a concept of blockage. As the blockage increased, the cavitation characteristics improved while the stagnation point near the leading edge gradually moved closer to the suction surface. However the best efficiency point moved to a lower flow rate point so that the rated performance in non-cavitation state was affected. In the case of different incidence angles, which were designed to exhibit the same specifications including the best efficiency point and rated performance, the cavitation characteristics were improved with a smaller incidence angle, and the cavity blockage near the leading edge was reduced. On the other hand, a larger incidence angle caused pressure fluctuations and cavity oscillations. Here, the amount (volume) of generated cavities could not be an absolute indicator to determine the cavitation characteristics. The leading edge shape was defined as an ellipse ratio. As the ellipse ratio increased, the cavitation characteristic improved due to the lower cavity blockage, and the rated performance such as the best efficiency point was almost maintained. The square-shaped leading edge, which was defined with the ellipse ratio of zero, obtained the best cavitation characteristics, but showed significant degradation in rated performance. As a parallel focus, an in-depth analysis of cavitation surge and pressure gain was presented with the tendency of head drop slope. For the next, the inlet diameter was reduced while maintaining the incidence angle, and the cavitation characteristics were improved with a narrower inlet diameter. This design method allowed to maintain the rated performance including the best efficiency point. In this dissertation, the cavitation characteristics of a pump were analyzed to be more influenced by the shape of vapor, which showed a specific tendency to the geometry, rather than being determined by the hydrodynamic parameters such as the inlet velocity (dynamic head). Although this can be understood as a relatively unusual case, it is encouraging that the results are well correlated with each other. The analysis for the saddle of a mixed-flow pump followed the subject of cavitation-related inlet diameter reduction. Most mixed-flow pumps obtain a saddle-like Q-P curve with fairly strong backflow, rotating stall, and reattaching flow owing to the increased incidence angle at low flow rates. These unstable flow patterns were suppressed under the design method for inlet diameter reduction, and the saddle improved without pressure fluctuations. Regarding the stall of an axial fan, an anti-stall fin (ASF) was presented as a simple and original method. An ASF-attached axial fan stably recovered performance degradation in the stalling flow rates without unstable flow patterns and pressure fluctuations. The sensitivity analysis for design variables was conducted through the 2k full factorial design method, and the optimization was performed using the response surface method. Subsequently, each one-factor analysis was performed for various design parameters that ASF can be derived aerodynamically, and the functional limitations were suggested with normalized dimensions. Meanwhile, all results included the underlying mechanisms to identify each cause.
Article
The aerodynamic and aeroelastic study of flow through turbomachine blade passage is the most essential requirement for the development of fuel-efficient gas turbine engines. The use of various cascade tunnels is very common to predict the flow behaviour within the blade passage and at the exit of cascade blades. The present work focuses on the innovative method for the development of annular sector cascade wind tunnel of exit Mach number 0.62 for such applications, using the computational study. The flow behaviour within the tunnel was numerically analysed using commercial software, ANSYS CFX®. To meet the special requirements/constraints like available laboratory space and height of the test section, the numerical study was explored for various design iterations. The use of a wide-angle diffuser with inclination was one of the challenges which were addressed using a porous screen. The modelling for the screens and honeycomb section are the most challenging aspects, which were resolved using porous media option with porosity and permeability as an input. The numerical method discusses to define the outlet flow domain as a half-spherical shape with opening boundary condition to simulate the actual exhaust flow condition. The cost-effective honeycomb section, with square-shaped cells, which are easy to manufacture and clean, worked effectively to straighten the flow by reducing velocity fluctuations in the lateral and longitudinal direction. The settling chamber screens reduced the streamwise velocity fluctuations, and accelerating flow in the contraction duct reduced the spatial variation. Near about less than 7% RMS change in the contraction duct exit velocity was achieved as it is the need to test the turbine blades. It is strongly believed that the various design aspects will help future designers for selecting various geometrical parameters for the design and development of such cascade tunnels.
Article
Full-text available
An approximate analysis is presented which is applicable to nonorthogonal coordinate systems having a curved centerline and planar transverse coordinate surfaces normal to the centerline. The primary flow direction is taken to coincide with the local direction of the duct centerline and is hence normal to transverse coordinate planes. The formulation utilizes vector components (velocity, vorticity, transport equations) defined in terms of local Cartesian directions aligned with the centerline tangent, although the governing equations themselves are expressed in general nonorthogonal coordinates. For curved centerlines, these vector quantities are redefined in new local Cartesian directions at each streamwise location. The use of local Cartesian variables and fluxes leads to governing equations which require only first derivatives of the coordinate transformation, and this provides for the aforementioned ease in using constructed coordinates.
Article
In the present study, the effects of bell mouth geometries on the flow rate of centrifugal blowers were numerically simulated using a commercial computational fluid dynamics (CFD) program, FLUENT. A total of 12 numerical models were prepared by combining the different values of bell mouth radii and gaps between the bell mouth and the upper fan case; the cross section of bell mouth was chosen as a circular arc. All models were meshed with hexahedra elements using the Gambit software. The frozen rotor method combined with a realizable k-epsilon turbulence model and nonequilibrium wall function was used to simulate the three-dimensional flow inside the centrifugal blowers. CFD investigation showed that the bell mouth radius had a strong effect on the flow rate, which can vary by more than 5% with different bell mouth models, whereas the effect of the gap between the bell mouth and the upper fan case on the flow rate was weaker. On the basis of CFD results, experiments were carried out for five typical bell mouths to verify the effect of the bell mouth radius, the gap between the bell mouth and the upper fan case on the flow rate. CFD results were validated by the good agreement between the CFD results and the parallel experimental results.
Article
A new approximate flow analysis, designed to enable numerical solution as an initial value problem, is developed for a wide class of viscous subsonic flows at high Reynolds number and in straight or smoothly curved three-dimensional flow geometries. The analysis is coordinate-independent and corrects an a priori known inviscid primary flow for viscous and thermal effects, secondary flows, total pressure distortion, internal flow blockage and pressure drop. Computed results include laminar solutions for three-dimensional boundary layer flow, fully viscous flow in circular arc ducts, and also flow in a curved duct shaped like a turbine blade passage.
Article
A broad program to develop advanced, reliable, and user oriented three-dimensional viscous design techniques for supersonic inlet systems, and encourage their transfer into the general user community is discussed. Features of the program include: (1) develop effective methods of computing three-dimensional flows within a zonal modeling methodology; (2) ensure reasonable agreement between said analysis and selective sets of benchmark validation data; (3) develop user orientation into said analysis; and (4) explore and develop advanced numerical methodology. Previously announced in STAR as N84-13190
Article
A three dimensional analysis for fully viscous subsonic internal flow is evaluated. The analysis, designated PEPSIG, solves an approximate form of the Navier-Stokes equations by an implicit spatial marching procedure. Results of calculations are presented for laminar flow through two different circular cross-sectioned 180 degree bends, and for laminar and turbulent flow through circular and square cross-sectioned 22.5 to 22.5 degree S-ducts. Quantitative comparisons with experimental data are shown for all cases. Special emphasis is placed on verifying the ability of the analysis to accurately predict the distored flow fields resulting from pressure-driven secondary flows. Previously announced in STAR as N84-13404
Article
This paper describes experimental measurements of secondary flow in a constant area, circular cross-section 30-30 deg S-duct, and compares the results obtained with the computations performed using the PEPSIG code, a parabolized Navier-Stokes code. The flow entering the duct was turbulent, with entrance Mach number of 0.6, and the boundary layer thickness at the duct entrance was 10 percent of the duct diameter. The duct mean radius of curvature to the duct diameter was 5.077. Flow parameters were measured at six stations along the length of the duct. These measurements were made using a five-port cone probe. At least ten radial traverses were made at each station on both sides of the symmetry plane. Wall static pressures along three azimuth angles of zero, 90, and 180 deg along the duct were measured. Plots presenting the secondary velocity field as well as contour plots of the total and static-pressure fields have been obtained. Strong secondary flows were observed in the first bend, and these continued into the second bend with the formation of new vorticity in the opposite sense in the second bend. The flow exiting the duct contained two pairs of counter-rotating vortices. The computational results are in general agreement with the experiments. However, it appears that the computations underestimate the extent of the pressure distortion, due to simplifications made in the pressure field calculations.
The effect of discharge coefficient of intake and exhaust valve for direct injection engine
  • A R Ismail
  • R A Bakar
  • Semin
Ismail AR, Bakar RA, Semin, "The effect of discharge coefficient of intake and exhaust valve for direct injection engine", International Conference on Mechanical Engineering, 2007.