KATHMANDU UNIVERSITY

SCHOOL OF ENGINEERING

DEPARTMENT OF MECHANICAL ENGINEERING

REPORT ON

CFD ANALYSIS OF INTERNAL PIPE FLOW USING ANSYS SOFTWARE

SUBMITTED BY : SUBMITTED TO :
Prithivi Gurung [41160] Sailesh chitrkar, Phd
Samman Singh Pradhan [41170] DOME, TTL
Sujan Adhikari [41174]

Date of submission: 5/1/2021

Table of Contents
1. Background ................................................................................................................................... 3
2. Objectives....................................................................................................................................... 3
3. Literature Review ......................................................................................................................... 4
3.1 Meshing ........................................................................................................................................ 4
3.2 Mesh generation using ICEM .................................................................................................... 5
3.3 Formation of boundary layer and velocity profile ................................................................... 5
3.3.1 Case: non-circular pipe (Rectangular pipe) ...................................................................... 5
4. Geometry of Pipes ......................................................................................................................... 7
5. MESH GENERATION................................................................................................................. 9
6. Results and discussion ................................................................................................................ 13
6.1 Flow at 6 different locations including the inlet and a fully developed region .................... 13
6.2 At 6 meters from the inlet, comparison between result between the coarse and the
medium sized mesh generated in ICEM ....................................................................................... 14
6.3 Discussion of the deviation in the results when V increases from 1 m/s to 3 m/s keeping
other parameters the same. ............................................................................................................ 16
6.4 Discussion of the deviation in the results when D increases from 0.2 m to 0.5 m keeping
other parameters the same ............................................................................................................. 17
6.5 Discussion on the deviation in the results when the circular profile is replaced with a
rectangular profile (0.1 m x 0.2 m). ............................................................................................... 18
Boundary conditions:...................................................................................................................... 18
6.6 Discussion of deviation in the results when the straight pipe is replaced with a bend of 90
degree and the velocity profiles at different locations of the bend section................................. 20
7. Conclusion ................................................................................................................................... 23
1. Background
In this report a geometric description of a pipe is obtained in discretized form, following it a
mesh is generated around the object, it is developed through a structured approach on the pipe.
Meshing provides a set of points on which the flow field solution is calculated by solving the
appropriate governing equations of fluid dynamics. The flow field solution data for pressure,
temperature and velocity variables. The CFD simulation in this report is specifically done for
internal flows: pipes. Internal flows are constrained laterally by bounding walls. The basics of
internal flow in pipes specifically circular and rectangular cross sections with varying
dimensions and boundary conditions are studied using computational fluid dynamics in
ANSYS 20 R1 software. An internal flow is constrained by the bounding walls, and the viscous
effects will grow and meet and permeate the entire flow. There is an entrance region where a
nearly inviscid upstream flow converges and enters the tube. Viscous boundary layers grow
downstream, retarding the axial flow u (r, x) at the wall and thereby accelerating the center
core flow to maintain the incompressible continuity requirement. At a finite distance from the
entrance, the boundary layers merge and the inviscid core disappears as shown in the figure
below. The various points in between entrance and outlet regions are studied for understanding
the concept of pipe flow as studied in fluid dynamics before and apply the theoretical
knowledge to study internal flow in pipes.

Figure 1: Development of velocity

profile
2. Objectives
Based on the first CFD activity on pipe flow the velocity profiles of the flow at 6 different
locations including the inlet and a fully developed region is studied; the results are analyzed to
compare similarities with the theoretical results studied in the 'Fluid Mechanics' course
(entrance region). From the simulations performed, the results with different boundary
conditions and different meshing approaches are also studied. They are as follows:
 • The difference in the successive size of the mesh at 6 meters from the inlet. The result
between the coarse and the medium sized mesh (in terms of the length) will not exceed
2x. The two results with the analytical result obtained for the fully developed flow will
be compared graphically.
• Discussion of the deviation in the results when V increases from 1 m/s to 3 m/s
keeping other parameters the same.
• Discussion of the deviation in the results when D increases from 0.2 m to 0.5 m keeping
other parameters the same.
• Discussion the deviation in the results when the circular profile is replaced with a
rectangular profile (0.1 m x 0.2 m).
• Discussion of deviation in the results when the straight pipe is replaced with a bend of
90 degree and the velocity profiles at different locations of the bend section will be
compared.

3. Literature Review
3.1 Meshing
A mesh partitions space into elements (or cells or zones) over which the equations can be
solved, which then approximates the solution over the larger domain. The mesh influences the
accuracy, convergence and speed of the solution.

Structured mesh: Grids that have a network of curvilinear gridlines which conform to the
boundaries, surface or both. The structured grid generation techniques are based on algebraic
interpolation methods and partial differential equations involving elliptic systems. Grid
adaption for structured grids prefer redistribution approach and is treated as a two-step process:
(1) an adaptation to the geometry of the body

(2) adaptation to the requirements of the flow.

Unstructured mesh: Grids that are composed of triangles (2-D) and tetrahedrons (3-D). The
grid information is provided by a set of nodes (coordinates) and the connectivity between the
nodes. This technique enables the generation of elements of variable size with desired
stretching. The unstructured grids offer a substantially greater geometric flexibility compared
with that of structured grids.
3.2 Mesh generation using ICEM
ICEM allows interactive editing of the mesh it is actually manipulating the mesh interactively
(or with automatic routines) to improve quality or physics capture. Common operations include
moving and merging nodes, splitting edges or nodes, adjusting element normals, creating
elements, refining or coarsening mesh, etc. There are also more automated operations such as
merging or smoothing meshes. An advantage of AIE (now just called ANSYS ICEM CFD) is
that the node projection is maintained (keeps the geometry shape much better). Curve projected
nodes stay on curves, surface projected nodes stay on the surface, etc.

3.3 Formation of boundary layer and velocity profile

An internal flow is constrained by the bounding walls, and the viscous effects will grow and
meet and permeate the entire flow. There is an entrance region where a nearly inviscid upstream
flow converges and enters the tube. Viscous boundary layers grow downstream, retarding the
axial flow u (r, x) at the wall and thereby accelerating the center core flow to maintain the
incompressible continuity requirement. The velocity profile formed by a laminar flow inside a
straight pipe is defined by the equation:

……………………………(1)

3.3.1 Case: non-circular pipe (Rectangular pipe)

For laminar flow in rectangles and triangles, the wall friction varies greatly, being largest near
the midpoints of the sides and zero in the corners. In turbulent flow through the same sections,
the shear is nearly constant along the sides, dropping off sharply to zero in the corners. This is
because of the phenomenon of turbulent secondary flow, in which there are non-zero mean
velocities (v and w) in the plane of the cross section. Some measurements of axial velocity and
secondary flow patterns are shown in the figure 2
 Figure 2: secondary flow in non-circular ducts

Due to the turbulence caused by the secondary flow we can use the concept of hydraulic
diameter which is defined by the equation

4×𝐴
𝐷ℎ = ………………………..(2)
𝑃
Where,
A= cross section area
P = perimeter
4. Geometry of Pipes
The first step to CFD analysis of internal flow in pipes is to create the geometry of the pipes for fluid
domain.

Figure 3: The pipe geometry with 0.2m internal diameter, 8 m length.

Figure 4: The pipe geometry with 0.5 m internal diameter and 8m length
Figure 5: The rectangular pipe or duct geometry with dimension (0.1m *0.2m) and length 8 m .

Figure 6: The 90 degree bend pipe with 0.5m internal diameter, 4m length on each sides and bend with radius of
0.5m.
5. MESH GENERATION
For the circular pipes and rectangular pipes structured meshing was used. For circular pipes the
number of nodes to form the O-grid mesh in lines of both cross-section (inlet and outlet) was set to
20 for coarse mesh generation and 30 for the medium mesh generation. Similarly, along the length
number of nodes was set to 50 for coarse mesh generation and 80 for medium mesh generation. For
rectangular pipes no O-grid mesh were used, the number of nodes was set to 40 for 0.2m edges and
20 for 0.1m edges. Similarly, the number of nodes was set to 100 along the length. For 90 degree bend
pipe simply ANSYS meshing was used. The element size was kept 0.05 m. The total number of nodes
formed was 406809 and element was 391300.

Figure 7: Structured Coarse mesh generation in the cross-section of 0.5m internal diameter circular-pipe.
Figure 8: Structured medium mesh generation for 0.5m internal diameter circular pipe.

Figure 9: Structured coarse mesh generation along the length for circular pipe of 0.5m diameter and 8m length.
Figure 10: Structured medium mesh generation for circular pipe of 0.5m diameter and 8m length

Figure 11: Structured mesh generation along the cross section (inlet and outlet) for rectangular pipe (0.1m
*0.2m).
Figure 12: structured mesh generation along the length of the rectangular pipe. The mesh shown here is only a
portion along the length.

Figure 13: ANSYS mesh generation for 90 degree bend pipe.

6. Results and discussion
6.1 Flow at 6 different locations including the inlet and a fully developed region

The boundary conditions and dimension of pipe are as follows

Velocity of fluid = 1 m/s

Diameter of pipe = 0.2 m
Length of pipe = 8 m

Hydrodynamically Fully developed flow

2.5

1.5

0.5

0
-0.15 -0.1 -0.05 0 0.05 0.1 0.15

Graph 1: The graph shows the analytical result for the velocity profile for a fully developed flow. Y-axis
represents velocity and X-axis represents the points along the diameter of pipe.

We can see how a curved profile is developed at edges of the wall in both figures due to the no
slip condition . This no slip condition creates a entrance region where the velocity
increses as the boundary layers converge towards the middle. In figure 3 and 4 we can
see velocity is increasing as the flow continues through the pipe and the boundary layer
converges Thought We can see a huge difference between the analytical solution and
the simulated results of hydrodynamically fully developed flows at the core region of
the velocity profile. Figure 3 shows that the analytical solution has parabolic shaped
profile which is how it is idealized and figure 4 shows a flat profile even after the flow
is fully developed. This shows that the flow is not laminar like in the ideal case and
turbulence induced to various factors like high velocity , large diameter of pipe.
 Velocity profile curves for 0.2 m dia(1m/s)

1.20

1.10

1.00
INLET
Velocity

0.1M FROM INLET

0.90
0.2 FROM INLET
2 FROM INLET
0.80
4 FROM INLET
6 FROM INLET
0.70

0.60
0.00 5.00 10.00 15.00 20.00

Position

Graph 2: Velocity profile at different section of the circular pipe

6.2 At 6 meters from the inlet, comparison between result between the coarse and the
medium sized mesh generated in ICEM
Boundary conditions:
Velocity of fluid = 3 m/s
Diameter of pipe = 0.5 m
Length of pipe = 8 m
It is noticed that there is not much difference between the velocity profiles at 6 m from the
inlet, simulated using coarse and medium sized mesh. Comparing both profiles with the one
obtained theoretically we observe that the boundary layer in the medium sized mess slightly
more well defined then in the coarse mesh. Though meshing is vital to the accuracy of results,
figure shows that the flat prolife at the core of the flow is not because of the difference in
meshing and because of other physical phenomena. So, we can use the coarse mesh to simulate
the pipe flow at these boundary conditions without detailing the mesh further saving time and
cost.
 Hydrodynamically developed velocity profile
7.00
6.00
5.00
4.00
velocity 3.00
2.00
1.00
0.00
-0.30 -0.20 -0.10 0.00 0.10 0.20 0.30
Positon

Graph 3: Hydrodynamically developed velocity profile

Velocity profile at 6m from inlet (Coarse vs Fine Meshed)

3.50

3.30

3.10

2.90

2.70
Velocity

2.50
6 from inlet (fine)
2.30 6 from inlet (coarse)
2.10

1.90

1.70

1.50
0.00 5.00 10.00 15.00 20.00

Position

Graph 4: Velocity profiles of the circular pipe at 6 m from inlet. The small difference can be seen between the
fine mesh and coarse mesh result
6.3 Discussion of the deviation in the results when V increases from 1 m/s to 3 m/s keeping
other parameters the same.
The boundary condition are as follows:

Velocity of fluid = 1 m/s and 3m/s

Diameter = 0.5 m

Length of pipe = 8m

Graph 5: The above two graphs show the velocity profile at inlet and 6m from inlet for circular pipe (1m/s and
3m/s).

Graph 6: The below two graphs show the velocity profile contour for circular pipe with flow velocity of 1 m/s
and 3m/s.
 Velocity profile at 6m form inlet (1m/s Vs 3m/s)
3.50

3.00

2.50

2.00
Velocity

1.50 6 from inlet (3m/s)

6 from inlet (1m/s)
1.00

0.50

0.00
0.00 5.00 10.00 15.00 20.00

Position

Graph 7: the velocity profiles for flow velocity of 1m/s and 3 m/s keeping other parameters
same.

These simulations show that the flow having the higher velocity when the diameter is equal.
has more well-defined parabolic velocity profile than one with the lower velocity when
observed at the same length. As both velocity profiles are observed at 6 m from the inlet, it can
be concluded that higher the inlet velocity shorter the entrance region as the flow with higher
velocity is fully developed before the lower velocity one.
6.4 Discussion of the deviation in the results when D increases from 0.2 m to 0.5 m
keeping other parameters the same
Boundary conditions:
Velocity of fluid = 3 m/s
Diameter = 0.5 m and 0.2 m

Figure 14: The velocity contour image for outlet of 0.5m (left) and 0.2m (right) internal diameter circular pipes.
 Velocity profile at 6 m from inlet (0.2 Vs 0.5) at 3m/s

4.00

3.50
Velocity

3.00 6 from inlet (Fine)

2.50
6 from inlet
(Coarse)
2.00

1.50
0.00 5.00 10.00 15.00 20.00
Position

Graph 8: Velocity profile for 0.2m (blue curve) and 0.5m (orange curve) at 6m from inlet.

From the graph 8, we observe that when we have same velocity the small diameter pipe has a
more well-defined parabolic velocity profile. The pipe with the bigger diameter has flatter
velocity profile because the turbulence increases as the diameter increases. The bigger diameter
has a longer entrance region than the one with the smaller diameter. The velocity of the fluid
at the outlet of the small diameter pipe is higher than that of the bigger diameter one as seen in
velocity contour image at outlet (fig.14).

6.5 Discussion on the deviation in the results when the circular profile is replaced with a
rectangular profile (0.1 m x 0.2 m).
Boundary conditions:
Velocity of fluid = 3 m/s
Cross section = 0.1 m x 0.2
Hydraulic diameter = 0.104 m (from equation .2)

.
 Figure 15: Velocity contour plot for outlet of rectangular pipe.

velocity profile at 6m from inlet

4

3.5

2.5

1.5
0 5 10 15 20

6m FI-circular pipe 6m FI-rectangular pipe

Graph 9: Velocity profile difference between 0.2m diameter circular pipe and rectangular pipe at 6m distance
from inlet.

Figure 16: Velocity profile vector image at inlet and 6 m distance from inlet for rectangular pipe
The first thing that can be observed here is the parabolic shape of the velocity profile at 6 m
from inlet in the rectangular pipe as seen in graph 9 and fig.16 and when compared to the
velocity profile of the circular pipe the parabolic profile is little incomplete though the avg
velocity is same of the circular pipe. The reason may be because when fluid through the same
sections, the shear is nearly constant along the sides, dropping off sharply to zero in the corners.
This is because of the phenomenon of turbulent secondary flow, in which there are nonzero
mean velocities v and w, in the plane of the cross section as shown in figure 15.
6.6 Discussion of deviation in the results when the straight pipe is replaced with a bend
of 90 degree and the velocity profiles at different locations of the bend section
Boundary condition:

Diameter = 0.5 m
Each side length = 4 m
Bend radius = 0.5 m
Flow velocity = 3 m/s

Inlet velocity profiles

3.05

2.95

2.9

2.85

2.8

2.75
0 5 10 15 20

inlet-pipe inlet-bendpipe

Graph 10: Inlet velocity profile of 0.5m diameter circular pipe and 0.5m 90 degree bend pipe.
 mid-section velocity profile
5

4.5

3.5

2.5

1.5

0.5

0
0 5 10 15 20

mid-section-circular pipe mid-section-bend pipe

Graph 12: velocity profile at 0.5 m diameter circular pipe mid-section and bend part
at 90 degree bend-pipe.

Near Outlet veocity profile

3.5

2.5

1.5

0.5

0
0 5 10 15 20

6m-from-inlet 2m-back-from-outlet

Graph 11: velocity profile at 6m from inlet for circular pipe and 90 degree bend pipe.
Figure 17: The velocity profile at different section of the 90 degree bend pipe.

Figure 18: Streamline image for 90 degree bend pipe.

In Figure 17 and 18 the velocity profile does not have enough pipe length to fully develop in
the pipe as the pipe length before or after the bend is relatively shorter. Velocity magnitude of
streamlines and the plot of velocity profile at the bend of the pipe show spike in the velocity of
the fluid and the flow velocity at the lower bend is lower as the fluid is forced to turn to follow
the ongoing normal flow. The high velocity flow at the base portion follows through the lover
circumference of the pipe so flow has a higher velocity at the lower part of the pipe as seen in
the graph.

7. Conclusion
In this report we have a taken a numerical approach to see how fluid flows through pipes in
various cases. CFD is a tool Governed by the Navier’s stokes equations which is always applied
to predict fluid motion through a pipe. Discretization of geometrical figures using meshing
should be done according to the boundary conditions given to us to get the best results. We can
conclude from the results that a ideal velocity profile cannot be observed in the pipe flow as it
is dependent on various factors like diameter, velocity, surface condition of pipe which all lead
to turbulence and as turbulence increase it disturbs the development of the velocity profile.
THANK YOU !!!