Professional Documents
Culture Documents
CENTRIFUGAL FAN
06311A0368
06311A0380
06311A03B4
06311A03B5
06311A0368
06311A0380
06311A03B4
06311A03B5
ACKNOWLEDGEMENT
We would take immense pleasure to acknowledge with gratitude, the help & support
extended during the course of our project entitled DESIGN AND ANALYSIS OF A LOW
SPEED CENTRIFUGAL FAN from all people who have helped in the successful
completion of this project.
We also thank our parents, class mates and friends for the kind support given by them at all
stages of the project.
ABSTRACT :
The current project is aimed to design a low specific speed centrifugal fan.
Fans belong to the family of turbo machines and they move air or gas continuously
at desired velocity by action of a rotor. Flow investigation of the fan is planned to
be carried out by using ANSYS-CFX software for different designed off design
points of operation. The performance of the fan generated from the CFD analysis at
the design point will be compared with that of the designed data assumed for
calculation. This will also be compared with the best efficiency point of operation.
For the analysis, an Auto CAD drawing and a 3-D model the fan impeller and casing
are developed for the designed fan. This is followed by the generation of Grid and
aerodynamic analysis using the available CFD solver. The work is concluded by identifying
possible zones of improvements in the design of impeller and casing and suggest suitable
modifications.
Area
Impeller Width
Absolute velocity
dP
Diameter
Impeller diameter
Energy
Speed in rpm
nsh
Shape number
nq
Specific speed
Pressure
Gas constant
Radius
Rc
Blade speed
Specific work
Number of blades
5
GREEK LETTERS:
Flow coefficient
Efficiency
Density
Angular Velocity
bl
Blade or Impeller
b
h
m
t
u
Blade or Vane
Hydraulic
Meridional
Tip
Tangential or Peripheral component
CONTENTS
1 INTRODUCTION
10
2 LITERATURE SURVEY.
2.1 Specific work and static pressure rise
2.2 Impeller
2.2.1 Slip
2.2.2 Inlet Vane angle
2.2.3 Pre whirl
2.2.4 Impeller outlet angle
2.2.5 Impeller outlet diameter
2.2.6 Effect of Viscosity
2.2.7 Inlet passage
2.2.8 Effect of surface roughness
2.2.9
Volute casing
15
2.5 Applications
30
4 EXTRACTION OF COORDINATES.
38
5 CFD THEORY
42
6 ANSYS CFX..
51
7 METHODOLOGY.
55
71
8.1 General
8.2 Variation of flow parameters in the chosen impeller
8.3 Results
8.4 Pictorial analysis
8.5 Graphs
9 CONCLUSIONS.
92
10 SUGGESTION..
92
11 REFERENCES
93
1. INTRODUCTION
1.1 Introduction
to Turbo machines
Turbomachines used for the compression of gases are classified under radial, axial or
mixed flow types depending on the flow through the impeller. In a radial or centrifugal
machine, the pressure increase due to the centrifugal action forms an important factor in its
operation. The energy is transferred by dynamic means from the impeller to the fluid. The
fluid because of centrifugal action is continuously thrown outwards making way for fresh
fluid to be inducted in because of the reduced local pressure. Another characteristic feature
of the centrifugal impeller is the angular momentum of the fluid flowing through the
impeller is increased by virtue of the impeller outer diameter being significantly larger than
the inlet diameter. In axial flow machines, a large mass of gas is set in motion by the
rotating impeller and is made to move forward because of the aerodynamic action of the
blades. A mixed flow machine encompasses the properties of both the above types.
Depending on the pressure rise attained, these machines are named as fans and blower or
compressors. There is however no distinct demarcation among the different types. Fans
handle gases in large volumes without appreciable density variation. Pressure ratio
attainable is of the order of 1.05. They are invariably single stage machines.
Blowers cover pressure ratios from 1.05 to about 4. They are made
either as single
of the material.
1.2 FAN :
A fan can be defined as a volumetric machine, which, like a pump, moves a
quantity of air or gas from one place to another. In doing this, it overcomes
resistance to flow by supplying the fluid with the energy necessary for continued
motion. Physically essential elements of a fan are a bladed impeller (rotor) and a
housing to collect the incoming air or gas and direct its flow. Fans, Blowers or
Compressors all move air, but at different pressures. At any point in the flow of air
through the impeller, a pressure head obtains the centripetal acceleration, so that
the static pressure of the air increases from the eye to the tip of the impeller.
I
b. S-Casing
c.
U-Casing
The above said fans have different characteristics suitable for specific applications. If the
requirement is to blow air in large volume rate capacity, but relatively low-pressure gain,
axial flow fans may be suited by contrast a fan required to blow air through filtrate system
offering a high flow resistance will have a relatively small volume flow rate capacity with
high pressure rise.
Air or gas enters the impeller of the fan axially through the suction chamber. This gas
flows through the flow passage between the impeller blades while impeller rotates. The
action of the impeller swings the gas from a smaller radius to a larger radius and delivers the
gas at a high pressure and velocity to the casing. Due to impeller rotation centrifugal force
also contributes to the stage pressure rise. At the exit of the impeller a spiral shaped casing
known as scroll or volute collects the flow from impeller which can further increase the
static pressure of air.
Forward Curved Centrifugal Fans
In forward curved centrifugal fans the blades are inclined in the direction of motion.
This type of fan is best suited for application requiring high volume flow at low to medium
pressure rise. This type is sometimes referred to as a Volume Blower. It can compete with
tube axial and guide vane axial fans for some duties. Its efficiency is less than axial fans.
Radial Discharge Centrifugal Fans
This type of fan is mainly suited for handling of air borne particles. In this type of fan
blades tend to be self-cleaning in moderately dirty conditions and in efficient units with
curved heel blades is thus often used for draught induction in the boilers. Because of
tolerance these fans are suitable for handling particulate matter in filtration duties.
Back-bladed Centrifugal Fans
In backward curved centrifugal fans, the blades at the impeller are inclined away from
the direction of motion. The static pressure rise in the rotor results from the centrifugal
energy and the diffusion of the relative flow. The stagnation pressure rise and stage work
13
depends on the whirl components (Cu , Cu ) of the absolute velocity vectors C and C
1
2
1
2
respectively. These impellers are employed for lower pressure and lower flow rates.
CROSS-FLOW FANS
In this type of fans the air enters the impeller through peripheral segment other than
through hub. These fans are used where convenience is more important than efficiency.
These fans are suitable for low-pressure rise applications. The applications of cross flow
fans are domestic fan assisted heaters, handhold hair dryers and air curtain.
14
2. LITERATURE SURVEY
The specific work is developed in the impeller only through the energy transfer to the
fluid through the vanes and is given by Euler's equation
W = U2C2 U1C1
W= specific work developed by the stage (N.m/Kg)
U1 = impeller speed at start of vane
U2 = impeller tip peripheral speed
C1 and C2 are the components the absolute velocity in the tangential direction at points just
before the inlet to the impeller vane and the exit from the impeller vane respectively.
The above Equation can be rewritten as:
W = (U22 U12 + C12-C22+W02-W32)/2
As the flow energy of the fluid comprises the pressure energy, the kinetic energy and that due to
the geodetic head, the energy at any section of the passage (except where energy is being
added) can be written as:
E = P/ + C2/Z + g.h
15
between 25 to 30.
16
17
2.3VELOCITY TRIANGLES:
The three velocities that make a velocity triangle are namely
i Blade speed U
ii Absolute velocity C
iii Relative velocity W
Generally the blade speed is taken as the base of the triangle, the direction of U1 and U2 follow the
direction of rotation of impeller and W and C's direction vary depending on that and such that
W=C-U (In vectorial notation) is satisfied
18
2.4 Impeller
19
The impeller forms the major component in the whole machine where the actual
energy transfer to the fluid takes place.
to the fluid cannot be expected due to the limited number of vanes. The vane
thickness, the viscous effects, the relative circulation, return flows and the effect
due to bends make the velocity and pressure distribution far from uniform. The
actual flow deflection is less than that obtained when the flow truly follows the
vanes. The difference between the vane angle and the actual flow angle is
accounted by the introduction of a factor called slip factor.
2.4.1 Slip
In the case of vane congruent flow, the specific work of the machine is given by
W = U2 C2U - U1 C IU
The peripheral components of velocity just outside the impeller are different from those
just within. This difference in specific work is due to the slip in the impeller that is the flow
does not wholly follow the impeller vanes. The energy transfer obtained in practice is less
than that calculated assuming the flow is one - dimensional and that the fluid outlet angle
equals the impeller vane angle due to the relative eddy and nonuniform velocity profile at
the impeller.
20
between 25 to 30.
21
For a given specific work, the peripheral speed will come down or if the
rotating speed is also fixed, the diameter comes down. But an increase in 2 could
cause adverse effects at the vane boundary.
When the
boundary layer is not in equilibrium with the pressure gradient across the channel, a
flow normal to the through flow may arise which will alter the desired potential
flow pattern and cause direct losses as a result of the partial dissipation of the
energy absorbed from the through flow to create the secondary motion.
22
all around the periphery and discharge it into the exit flange. A spiral casing can be
used with or without a diffuser ring.
= constant
23
24
25
26
27
vehicle is a well-known example. The air sucked through the radiators cools the circulating
water as well as the engine. For this propeller fans are used and driven by the engine
through belt transmission drive. For cooling the electric motors, fans are generally mounted
on the extension of their shafts.
Steel Plants
In steel plant applications large and small fans are used. One or more high-pressure
blowers are also employed to supply blast furnace gases to the steam boilers. In such cases
impellers must be able to operate at high temperatures and speed. Main blast furnace
blowers are required to develop high pressures and therefore they apply many centrifugal
stages.
Other applications include pneumatic transport of granular materials, centrifugal
separators, furnace and drying equipment. The miniature fans are used in much equipment
for component cooling.
29
: 11000 cum/hr
ii.
300 mmwc
iii.
Approximate total pr
315 mmwc
iv.
250mair
v.
262.5mair
vi.
Specific work
: 2575.125 m^2/s^2
vii.
: 3.06 m^3/s
viii.
Reference density
: 1.2 kg/m^3
ix.
x.
Operating speed
: 980 rpm
xi.
Reference pressure
xii.
Reference temperature
20 deg.C
11.6 KW
30
{tan=(b1-b2)/(r2-r1)}
3.3 AUTO CAD DESIGN OF THE FAN IMPELLER:Computer-aided design (CAD) is the use of computer technology for the design of
objects, real or virtual. CAD often involves more than just shapes. As in the manual drafting
of technical and engineering drawings, the output of CAD often must convey also symbolic
information such as materials, processes, dimensions, and tolerances, according to
application-specific conventions.
CAD may be used to design curves and figures in two-dimensional space; or curves,
surfaces, and solids in three-dimensional objects. It is an important industrial art extensively
used in many applications, including automotive, shipbuilding, and aerospace industries,
industrial and architectural design, prosthetics, and many more. CAD is also widely used to
produce computer animation for special effects in movies, advertising and technical
manuals.
CAD has become an especially important technology within the scope of computer-aided
technologies, with benefits such as lower product development costs and a greatly shortened
design cycle. CAD enables designers to lay out and develop work on screen, print it out and
save it for future editing, saving time on their drawings.
AutoCAD software is used to design a two-dimensional model of the impeller fan and it
is also used in extraction of co-ordinates. The process is explained in detailed steps with the
assist of figures below
35
6) The enclosed arc is the median of the blade and it is shown in figure 3.
7) Taking 2mm off set on either side of the blade median curve, two identical curves are drawn. The top
curve is the pressure side and the bottom curve is the suction side. (Figure 4)
36
8) After obtaining one blade mirroring is used, where the numbers of blades are specified as 10 and
angle as 360.
9) To generate the side view the taper is considered and the following figure is generated
37
4. EXTRACTION OF COORDINATES
4.1 Method of Extraction
The coordinates of blade, hub and shroud are extracted from the 2-d diagram of fan impeller.
Coordinates are used in generating a 3D figure in turbo grid.
A series of coordinates are absorbed from a 2D cad diagram,
i.
The CAD diagram is first simplified to represent one blade passing through one of the axis.
ii.
Further more the area between the inner radius and outer radius are divided at a series of
equal intervals.
iii.
For example a series of concentric circles are drawn considering the center of the impeller as
shown in the fig.
iv.
These lines intersect the blade profile at both pressure and suction side and also intersecting
the axis as shown.
v.
Considering the geometrical x axis as y axis an geometrical y axis as x axis, using the crock
screw thumb rule the meridional geometrical x axis represents z axis.
vi.
Now, considering the intersection point on the blade profile the perpendicular distance from
x and y as shown in fig., the x and y coordinates are absorbed.
vii.
For the similar point the circle passing through the intersection also passes through the
geometrical X axis as seen in fig., a perpendicular is drawn to the meridional diagram.
viii.
From the meridional diagram, as defined earlier the geometrical x axis is the z axis, from
this the perpendicular intersection the meridional diagram at both hub and shroud the "zhub" and "z-shroud" coordinates are extracted, as the representation uses the crock screw
thumb rule the values of z is considered negative.
ix.
And for the leading edge a series of concentric circles with a difference of "2mm" are drawn
and coordinates are generated for the x, y, z-hub, z-shroud.
x.
As the value of z is generated for both hub and shroud, by varying the values of z,
profile.curve coordinates are generated as a set for hub using the z-hub coordinates, and a
set for the z-shroud.
38
275
275
275.9
275.899
-0.1864
276.689
276.689
277.6588
277.6571
0.9718
305
300.2754
53.4761
345
323.6533
119.0809
385
341.0639
178.6069
425
353.9198
235.2993
465
363.2347
290.3198
505
369.43
344.3058
545
372.7376
397.6074
585
373.1085
450.4228
625
371.1548
502.8609
665
366.3716
554.9747
665
362.0027
557.8342
625
366.9361
505.9475
585
369.2318
453.7542
545
368.8591
401.2081
505
365.7523
348.21
465
359.6106
294.5772
425
350.7593
239.9852
385
338.2689
183.8454
345
321.3768
125.1295
305
298.8742
60.8211
277.6588
277.4731
10.1544
276.689
276.5684
8.1688
275.9
275.8225
6.5400
275
275
39
275
275
104.000
275.9
275.899
-0.1864
103.859
276.689
276.689
103.736
277.6588
277.6571
0.9718
103.584
305
300.2754
53.4761
99.308
345
323.6533
119.0809
93.051
385
341.0639
178.6069
86.795
425
353.9198
235.2993
80.538
465
363.2347
290.3198
74.282
505
369.43
344.3058
68.026
545
372.7376
397.6074
61.769
585
373.1085
450.4228
55.513
625
371.1548
502.8609
49.256
665
366.3716
554.9747
43.000
665
362.0027
557.8342
43.000
625
366.9361
505.9475
49.256
585
369.2318
453.7542
55.513
545
368.8591
401.2081
61.769
505
365.7523
348.21
68.026
465
359.6106
294.5772
74.282
425
350.7593
239.9852
80.538
385
338.2689
183.8454
86.795
345
321.3768
125.1295
93.051
305
298.8742
60.8211
99.308
277.6588
277.4731
10.1544
103.584
276.689
276.5684
8.1688
103.736
275.9
275.8225
6.5400
103.859
275
275
104.000
40
In generating the hub.curve and shroud.curve file from the meridional view, a series of
horizontal lines intersection both hub and shroud lines are drawn, and the coordinates for
these intersection points are considered as shown in fig.
Hub curve
X
-195
40
-155
80
-115
120
-75
160
-35
195
275
345
665
Shroud curve
X
249.9640
-129.6601
250.4691
-124.6601
252.0511
-119.6601
254.964
-114.6601
259.964
-109.6601
271.0792
-104.9638
665
-43
41
5. CFD THEORY
5. CFD THEORY:
CFD is playing a strong role as a design tool as well research tool. In CFD, the fundamental
equations of fluid mechanics are based on the following universal laws of conservation:
1. Conservation of mass
2. Conservation of momentum
3. Conservation of energy.
Mass is conserved
Continuity equation
Momentum equation
Energy conserved
Energy equation
Where,
x, y, z
Cartesian Coordinates
u, v, w
L.H.S
R.H.S
Density
Where,
= Fluid density
43
in X direction
in Y direction
in Z direction
44
Where,
e
internal energy
V^2/2
K
Kinetic Energy
Dh Dp Q
.q
Dt Dt t
computer power and enhanced graphics software, it is now much easier for CFD analysts to
create animations of their data. These often help in understanding complex flow phenomena
that are sometimes difficult to see from static plots.
5.2.1. K-Epsilon Model:
Boussinesq suggested that the apparent turbulent shearing stresses might be related to the
rate of mean strain through an apparent scalar turbulent or "eddy" viscosity. For the general
Reynolds stress tensor the Boussinesq assumption gives
ui u j
x
xi
j
ui u 'j T
'
2 ij T uk k
3
xk
Where T is the turbulent viscosity, k is the kinetic energy of turbulence given by,
ui'u 'j
k
2
By analogy with kinetic theory, by which molecular (laminar) viscosity for gases be evaluated
with reasonable accuracy, we might expect that the turbulent viscosity can be modeled as:
T T l
Where vT and l are characteristic velocity and length scale of turbulence respectively. The
problem is to find suitable means of evaluating them.
Algebraic turbulence models invariably utilize boussinesq assumption. One of the most
successful of this type of model was suggested by Prandtl and is known as "mixing length
hypothesis".
T l 2
u
y
Where l a mixing length can be thought of as a transverse distance over which particles
maintain their original momentum, some what on the order of a mean free path for the
collision or mixing of globules of fluid. The product l * u/y can be interpreted as the
characteristic velocity of turbulence, VT. In the above equation, u is the component of
velocity in the primary flow direction, and y is the coordinate transverse to the primary flow
direction.
There are other models, which use one partial differential equation for the transport of
turbulent kinetic energy (TKE) from which velocity scales are obtained. The length scale is
prescribed by an algebraic formulation. The most common turbulence model generally used
is the two-equation turbulence model or k- model. There are so many variants of this
model. In these models the length scale is also obtained from solving a partial differential
equation.
47
The most commonly used variable for obtaining the length scale is dissipation rate of
turbulent kinetic energy denoted by E. Generally the turbulent kinetic energy is expressed as
turbulent intensity as defined below.
1/ 2
T C k /
2
Dk
Dt x j
T // Prk
u
k ui u j 2
i
k ij
T
x j x j xi 3
at the discrete points only. In this sense partial differential equations have been discretized.
This method of discretization is called Finite difference method. Most common finitedifference representations of derivatives are based on Taylors series expansion.
ui , j 1 ui , j
x
u
O ( x )
Forward difference
u i , j u i , j 1
i, j
O ( x )
Backward difference
T
dR S q.nds 0
t
The first term in the above equation is an integral over the control volume, represents the time
rate of increase in the energy stored in the volume. The second term, an integral over the
surface of the volume, represents the net rate at which energy is conducted out through the
surface of the volume. This is the integral or control-volume form of conservation law. The
integral approach includes the Finite volume method and Finite element method. The FVM
method has an obvious advantage over a FDM. If the physical domain is highly irregular
and complicated since arbitrary volumes can be utilized to subdivide the physical domain.
49
Also since the integral equations are solved directly in the physical domain, no co-ordinate
transformations required. Another advantage of FVM is that mass, momentum and energy
are automatically conserved
6. ANSYS CFX
6.1 Introduction to ANSYS CFX
ANSYS CFX is a high-performance, general purpose CFD program that has been
50
applied to solve wide-ranging fluid flow problems for over 20 years. At the heart of ANSYS
CFX is its advanced solver technology, the key to achieving reliable and accurate solutions
quickly and robustly. The modern, highly parallelized solver is the foundation for an
abundant choice of physical models to capture virtually any type of phenomena related to
fluid flow: laminar to turbulent (including transition), incompressible to fully compressible,
subsonic to trans- and supersonic, isothermal or with heat transfer by convection and/or
radiation, non-reacting to combusting, stationary and/or rotating devices, single fluids and
mixtures of fluids in one or more phases (incl. free surfaces), and much, much more. The
solver and its many physical models are wrapped in a modern, intuitive, and flexible GUI
and user environment, with extensive capabilities for customization and automation using
session files, scripting, and a powerful expression language.
52
1. Vortex structures in a four-stroke engine just after injection of fuel and intake valve
opening.
2.
3.
Exchanger.
53
4.
conditions for
7.METHODOLOGY
54
upstream of the blade leading edge, and like-wise also extends downstream of the trailing
edge. The sample data for line elements are given in the, this data is arranged in order to
obtain hub and shroud blade profiles. This process requires programming file in
TURBOGRID, which can transfer large amount line data instantly.
CUTTING THE TRAILING AND DRIVING SURFACES
Hub.curve
X
Y
0
40
80
120
160
195
275
345
665
0
0
0
0
0
0
0
0
0
Z
-195
-155
-115
-75
-35
0
0
0
0
Shroud.curve
X
249.9640
250.4691
252.0511
254.964
259.964
271.0792
665
Y
0
0
0
0
0
0
0
Z
-129.6601
-124.6601
-119.6601
-114.6601
-109.6601
-104.9638
-43
shroud curve runs upstream to downstream and must extend upstream of the blade leading
edge and downstream of the blade trailing edge the points are listed, line by line in free
format style in order from upstream to downstream. These data points are used to place the
nodes on the shroud surface, which is defined as the surface of revolution of a curve joined
by these points.
Example: Considering XZ Plane with Z as Axis of Rotation
Fig: Hub Curve and Shroud Curve Profile curve Data File:
The profile data file contains the blade profile curves in Cartesian or cylindrical form.
The profile points are listed, line-by-line, in a closed loop surrounding the blade. The blade
profiles should lie on a surface of revolution to facilitate transformation to m-prime, theta
conformal space.
A minimum of two blade profiles are required, one which lies exactly on the hub surface
and one which lies exactly on the shroud surface. The profiles must be listed in the file in
order from hub to shroud. Multi bladed geometries are handled by placing multiple blade
profile definitions in the same profile.
57
Profile. Curve:
# Profile 1
X
275
275.899
276.689
277.6571
300.2754
323.6533
341.0639
353.9198
363.2347
369.43
372.7376
373.1085
371.1548
366.3716
362.0027
366.9361
369.2318
368.8591
365.7523
359.6106
350.7593
338.2689
321.3768
298.8742
277.4731
276.5684
275.8225
275
Y
0
-0.1864
0
0.9718
53.4761
119.0809
178.6069
235.2993
290.3198
344.3058
397.6074
450.4228
502.8609
554.9747
557.8342
505.9475
453.7542
401.2081
348.21
294.5772
239.9852
183.8454
125.1295
60.8211
10.1544
8.1688
6.5400
0
#Profile 2
58
Z
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
X
275
275.899
276.689
277.6571
300.2754
323.6533
341.0639
353.9198
363.2347
369.43
372.7376
373.1085
371.1548
366.3716
362.0027
366.9361
369.2318
368.8591
365.7523
359.6106
350.7593
338.2689
321.3768
298.8742
277.4731
276.5684
275.8225
275
Y
0
-0.1864
0
0.9718
53.4761
119.0809
178.6069
235.2993
290.3198
344.3058
397.6074
450.4228
502.8609
554.9747
557.8342
505.9475
453.7542
401.2081
348.21
294.5772
239.9852
183.8454
125.1295
60.8211
10.1544
8.1688
6.5400
0
Z
104.000
103.859
103.736
103.584
99.308
93.051
86.795
80.538
74.282
68.026
61.769
55.513
49.256
43.000
43.000
49.256
55.513
61.769
68.026
74.282
80.538
86.795
93.051
99.308
103.584
103.736
103.859
104.000
The first step is to check whether the blade profile data obtained from solid model is
intersecting hub and shroud curves or not. We use CFX-Turbogrid intersect option for this
purpose. Using this option, we have to see that blade profile must lie on the surface of
revolution of hub and shroud as shown in fig Turbo grid intersecting capability can convert
an existing set of blade profiles that does not necessarily lie on the surface of revolution into
one that can be used in a CFX-Turbogrid template.
Next step is generating grid. Among the various templates available in turbogrid, Multi
Block Grid template as shown in fig is used. By the way of adjusting control points in fig a
good quality hexahedral grid can be generated. Flip topology is used to correct negative grid
volume due to left-handed system. The mesh command creates mesh grid but also calculates
and displays the minimum and maximum skew angle in the grid and the node at which it
occurs. The View command in the GUI window can be used to see the different views of
the grid like Cartesian view, Meridional view and blade-to-blade view as shown in the
figure.
59
Adjusting the control points at the Leading Edge & Trailing Edge
60
61
The mesh generated by adjusting the control points as shown in Fig and
correspondingly Circumferential view of 3d Impeller surfaces & Periodical
arrangement of blades through out the circumference are shown in Figs
The following parameters were considered to check the quality of the grids:
Skew angle: It is defined as the internal angle of the octahedron. Ideally, all the
angles should be equal to 90 degrees to get a perfect orthogonal grid. However, for
practical purposes, the grid is considered to be of high quality if the minimum
skew angle is not lower than 15 degrees and the maximum skew angle is not
greater than 165 degrees.
Grid volume: Negative volume meant overlapping of adjacent grids, which would
lead to errors in solver. Care was taken to ensure that there was no negative
volume in the grids.
Aspect ratio: It is defined as the ratio of the longest side to the shortest side. Its
minimum value is 1. For good quality grid creation, the maximum aspect ratio
63
S.No
Component
Number of Number of
Minimum
Nodes
3DIMPELLER 19380
Elements
16352
Maximum
18
163
Importing the mesh assembly and region definition: The mesh file (.grd) file
was imported separately for the 3DImpeller. The grid file (with extension .grd) is
the file necessary to generate the grid.
The .gcf file contains the topographical details of the file, while the boundary
conditions file (.bci) specifies the inlet, exit, periodic and the blade regions for
each assembly.
II.
blade are defined. The hub, shroud and the blades of both assemblies
were treated as walls. The interface between the periodic1 and periodic2 is defined
as rotational periodicity. The boundary conditions were applied at inlet and outlet.
III.
Initial conditions: - The initial condition for the pressure field should be the
average of the highest value of pressure specified on any of the Outlet boundaries
and the lowest value of pressure specified on any of the Inlet boundaries. This
reduces the likelihood of spurious inflow at Outlets, or outflow at Inlets, during the
course of the solution. A sensible initial guess for the temperature field is an
64
In the pre processing the following fluid domains and boundary conditions are
specified.
1. Simulation
: Steady State
2. Domains
: Fluid
R1
: Impeller (Rotating)
3. Boundary Conditions:
Inlet
: Impeller inlet
Outlet
: Impeller exit
: 1.0132 bar
Wall
: smooth
Mass flow
: 3.672 kg/s
4. Fluid Properties:
Working Fluid
: air at 25C
Density
: 1.2 kg/m3
65
5. Rotation Axis
: Z
6. Turbulence Model:
Turbulence Model
: k-Epsilon
: None
7. Interface:
Type
: Fluid -Fluid
Interface models
: Rotational periodicity
The solver solves the mass, momentum and energy equations and calculates
pressure, velocity, enthalpy etc in the flow domain in each control volumes. The
inlet relative pressure and reference pressure plays a vital role to avoid round-off
errors. Reference pressure is the absolute pressure datum from which all other
pressure values are taken. It is a property of the entire simulation. So all domains
must use the same reference pressure value. The reference pressure will affect the
value of every other pressure set in simulation. It is used to avoid problems with
round-off errors which can occur when the dynamic pressure change in a fluid, that
drives the flow are small compared to the absolute pressure level. The relative
pressure specification set is measured relative to the reference pressure value. The
solver parameters are
1. Basic Settings: Steady State Simulations Advection Scheme is carried out using a
Numerical Advection Correction Scheme (Specify Blend). This selection allows
setting a Blend Factor between 0.0 and 1.0 for the advection scheme. A value of
0.0 is equivalent to using the First Order Advection Scheme and is the most robust
option. A value of 1.0 uses Second Order differencing for the advection terms; this
is not the same as the High Resolution advection scheme. This setting is more
66
accurate but less robust. Values between 0.0 and 1.0 blend First and Second Order
differencing, with increased accuracy and reduced robustness as you approach 1.0.
At the higher values overshoots and undershoots can appear, at lower values
excessive diffusivity can occur. It is therefore recommended to use a value of 0.75
for good accuracy of CFD results.
2. Timescale Control for a steady state simulation: The selection of an appropriate
time step size is essential in order to obtain good convergence rates for simulation.
In general there are two situations in which we use a physical time step:
Advection scheme
67
Maximum Iterations
: 200
: RMS
68
Meridional view
Once the solution is converged, the solver writes all the data related to grid,
boundary conditions and flow parameters are stored in the result file. It is a binary
file, which can be opened by loading result file in CFX-Post, and the results are
analyzed. The performance of compressor stage is studied by using suitable
macros. The various plots are drawn and listed in results. Using the function
calculator option parameters like Mass flow rate, Velocity, Pressure, Enthalpy,
Entropy etc can be calculated. Plots are also available for various parameters like
Velocity, Pressure and Mach number etc, which show the variation of parameters
through out the domain. The efficiency, torque and power are obtained using
softwares macro.
The similar type of stage analysis is carried out for different mass flows i.e.
70%,80%,90%,110%,120%,130%etc.
70
Velocity magnitude
2)
Pressure ratio
3)
4)
Head coefficient
5)
Shaft power
6)
Overall efficiency
The above mentioned parameters are plotted with respect to radius, mass flow rate
and speed. The following plots are generated from the present CFD analysis for
better understanding of the following phenomenon to centrifugal fan:
71
1)
Vector Plots
2)
Path Lines
3)
Contour Plots
8.2.3Variation of Static Pressure and Total Pressure with flow rate and
effect of Speed
Static Pressure: It is found that static pressure decreases with flow rate and static
pressure ratio is found to increase with speed for respective mass flow rates. Static
pressure values are tabulated in the table. Variation of static pressure ratio is also
found to be similar to static pressure variation.
Total Pressure: It is found that the total pressure decrease with flow rate. Total
pressure ratio is found to increase with speed for respective mass flow. The total
pressure values are tabulated in the table. Variation of total pressure ratio is also found
to be similar to total pressure variation
8.2.4Variation of static pressure along the pressure and suction side of the
impeller vane
It is clear that static pressure increases up to a certain radius, but reduces there after
on the pressure side of the blade. This behaviour is because the blade extends only
upto a particular radius.
73
8.3RESULTS
8.3.1Values obtained from CFX for 980 rpm
flow
per
% flow passag
%
e, cu.m
Flow coeff
exit total
pressure
(pa)
Pr
rise(pa)
pre
.static
head coeff
Cu2
total efficiency
flow angle
alpha
70
0.25
0.009
104924
97.8786
75.1928
80
0.29
0.0103
104795
98.1630
72.5490
90
0.33
0.0116
104673
98.2193
69.8385
100
0.36
0.0129
104558
98.0767
67.1102
110
0.40
0.0142
104428
97.6966
64.3117
120
0.44
0.0155
104295
96.8224
61.4162
130
0.47
0.0167
104138
95.2642
58.4596
Flow coeff
exit total
pressure
(pa)
Pr
rise(pa)
pre
.static
head coeff
Cu2
total
efficiency
flow angle
alpha
70
0.25
0.0061
109831
-71.51
96.42
63.94
80
0.29
0.007
109512
-68.13
97.17
57.76
90
0.33
0.0078
109475
-69.86
97.41
77.44
100
0.36
0.0087
109286
-67.93
97.82
75.82
110
0.40
0.0096
109092
-66.12
98.11
74.05
120
0.44
0.0104
108905
-64.51
98.25
72.26
130
0.47
0.0113
108727
-63.05
98.30
70.44
74
Flow coeff
exit total
pressure
(pa)
Pr
rise(pa)
pre
.static
head coeff
Cu2
total
efficiency
flow angle
alpha
70
0.25
.0044
118556
17240.2 111481
0.188
-106.23
93.78
37.64
80
0.29
.005
118163
16846.7 111397
.0.183
-103.5
94.9
37.04
90
0.33
.0057
117753
16437.4 111285
0.179
-100.4
95.88
64.70
100
0.36
0.0063
117372
16057.0 111149
0.175
-97.5
96.7
63.53
110
0.40
0.0069
116913
15600 110958
0.17
-94
97.21
56.53
120
0.44
0.0076
116860
15547 111042
0.169
-96.6
97.33
74.57
130
0.47
0.0082
116889
15378 111002
0.167
-95.2
97.62
76.82
75
PRESSURE
Total pressure:-
76
Relative pressure:-
77
Static pressure:-
78
79
VELOCITY
Velocity vectors at 50% span:-
80
MERIDONIAL PLOTS
81
82
BLADE LOADING
At 1450 RPM and 70% flow :-
83
84
85
86
8.5 GRAPHS
87
88
89
90
91
9.CONCLUSIONS
A low specific speed centrifugal fan was designed for the given flow and
head conditions. The fan impeller was modelled using ANSYS Turbo Grid and was
analysed using CFX package.
The fan performance was evaluated and studied for different flow
conditions covering design and off-design points of operation and also for different
speeds.
The performance is seen to be following the normal trend for a low specific
speed fan and the flow and head curve shifts upwards with increasing speed.
The impeller efficiency seen to be maximum at the design point and decreasing
at off-design conditions. The efficiency is found to be above 90%, this is because
the windage losses, frictional losses have not been accounted.
The different contour and vector plots as well as the blade loading curve are
included for typical cases of design and off-design conditions.
The pressure rise is seen to increase uniformly along the impeller passage.
92
11.REFERENCES
b) Wolfgang Scheer
c) Balje, O.D.
d) Pfleiderer
e) Wikipedia.org
93