SlideShare a Scribd company logo
Primitive Variable Formulation
MEL 807
Computational Heat Transfer (2-0-4)
Dr. Prabal Talukdar
Assistant Professor
Department of Mechanical Engineering
IIT Delhi
N-S Equation
0
y
v
x
u
y
p
y
v
x
v
y
v
v
x
v
u
t
v
x
p
y
u
x
u
y
u
v
x
u
u
t
u
2
2
2
2
2
2
2
2
=
∂
∂
+
∂
∂
∂
∂
−
⎟
⎟
⎠
⎞
⎜
⎜
⎝
⎛
∂
∂
+
∂
∂
μ
=
∂
∂
ρ
+
∂
∂
ρ
+
∂
∂
ρ
∂
∂
−
⎟
⎟
⎠
⎞
⎜
⎜
⎝
⎛
∂
∂
+
∂
∂
μ
=
∂
∂
ρ
+
∂
∂
ρ
+
∂
∂
ρ
⎟
⎟
⎠
⎞
⎜
⎜
⎝
⎛
∂
ζ
∂
+
∂
ζ
∂
μ
=
∂
ζ
∂
ρ
+
∂
ζ
∂
ρ
+
∂
ζ
∂
ρ 2
2
2
2
y
x
y
v
x
u
t
2-D, Navier-Stokes equation,
We discussed—Stream-function-vorticity approach
ζ
−
=
∂
ψ
∂
+
∂
ψ
∂
2
2
2
2
y
x
Sequential Solution
• In a sequential approach, we solve set of coupled PDEs
in a sequential manner
• We identify
• u-momentum equation for u velocity component
• v-momentum for v velocity component
• Continuity as PDE for pressure
• Possible solution loop
• Discretize and solve u over domain, assuming v, p known
from current iterates
• Similarly solve v from discrete v-mom equations
• Similarly solve p from discrete continuity
• Continue until convergence
Pressure into Continuity
• Look into the problems of introducing pressure into
continuity equation for incompressible flows
• We have 3 equations (u, v momentum for 2D and
continuity) and 3 unknowns (u, v and p)
• But how can pressure be linked to continuity?
• For compressible flow- not a problem
• But for incompressible flow density is constant
– No obvious way to introduce pressure into continuity.
RT
/
P
0
)
.(
t
=
ρ
=
ρ
∇
+
∂
ρ
∂
V
Sequential vs. Direct Schemes
• Important to recognize that the problem of pressure
computation for incompressible flows is an artifact of
sequential solution procedure
• Only for sequential procedures it is necessary to identify
a PDE for each unknown variable
– U-mom for u velocity; v-mom for v velocity, continuity equation
for pressure; energy equation for temperature etc.
• For direct solution procedures, no such one-on-one
identification is necessary
Direct Scheme
• Discretize u, v momentum equations and
continuity at all cells
• Assemble a big matrix: Ax = b
• Solution vector is:
x = [u1,u2,----uN,v1,v2,---vN,p1,p2,---pN]T
• No need to identify which unknown gets solved
using which PDE
• Unfortunately too expensive for practical use as
of this writing
Pressure Based Methods
• Sequential methods are preferred for
incompressible flows
– Low storage and reasonable convergence rate
• For incompressible flows, how to obtain
pressure from continuity?
– Density unrelated to pressure for incompressible
flows
• Pressure based methods are techniques to
introduce pressure into continuity equation
Other Methods
• Density based methods
– Popular in compressible flow community
– (u,v,w,ρ,T) used as unknowns
• Artificial compressibility/preconditioning
schemes
– Effectively add some artificial compressibility
to enable use of density-based schemes for
incompressible flows
– We will focus on pressure based methods as
they are widely used
Representation of the pressure
gradient
N-S equation
)
(
y
3
2
y
u
x
y
v
y
f
S
)
(
x
3
2
x
v
y
x
u
x
f
S
S
j
p
)
v
(
)
v
(
S
i
p
)
u
(
)
u
(
v
v
u
u
v
u
V
V
V
V
⋅
∇
μ
∂
∂
−
⎟
⎟
⎠
⎞
⎜
⎜
⎝
⎛
∂
∂
μ
∂
∂
+
⎟
⎟
⎠
⎞
⎜
⎜
⎝
⎛
∂
∂
μ
∂
∂
+
=
⋅
∇
μ
∂
∂
−
⎟
⎠
⎞
⎜
⎝
⎛
∂
∂
μ
∂
∂
+
⎟
⎠
⎞
⎜
⎝
⎛
∂
∂
μ
∂
∂
+
=
+
⋅
∇
−
∇
μ
⋅
∇
=
ρ
⋅
∇
+
⋅
∇
−
∇
μ
⋅
∇
=
ρ
⋅
∇
Body force
Source term
Representation of the pressure
gradient (cont’d)
u
S
i
p
)
u
(
)
u
( +
⋅
∇
−
∇
μ
⋅
∇
=
ρ
⋅
∇ V
dx
dp
−
In 1D,
Integrating this term over
a control volume, we get
2
p
p
2
p
p
2
p
p
p
p
dx
dx
dp
E
W
E
P
P
W
e
w
−
=
+
−
+
=
−
=
−
∫
w is assumed midway between P and W,
But not a restriction
Implications
2
p
p
dx
dp E
W −
=
−
Implies that the momentum equation will contain a term
of pressure difference of alternate grid points instead of
adjacent grid points
Is that a problem??
Implication: The pressure is calculated from a coarser
grid than the one actually employed -reduces accuracy
But there is another serious problem-follows!!!
Checkerboard pressure field
• Serious Problem:
– See the zig-zag field which is unrealistic. The proposed scheme
will give PW-PE zero at every grid point
– Although it represents a wavy pressure field, proposed scheme
will result a uniform pressure field in the momentum equation
– Patterns are even more unrealistic in 2D and 3D
0
0
Representation of the Continuity
Equation
• Continuity equation
• 1D steady,
incompressible flow
• Integrating over a CV
• Using a piecewise
linear profile for u
0
u
u
0
2
u
u
2
u
u
0
u
u
0
dx
du
0
)
V
.(
W
E
P
W
E
P
w
e
=
−
=
+
−
+
=
−
=
=
ρ
∇
Checkerboard Velocity Field
• So, the discretized continuity equation demands the
equality of velocities at alternate grid points and not at
adjacent ones
• Although the following wavy velocity field satisfies the
continuity equation, it is unrealistic
• Conclusions: first order derivative always creates problem
0
0
U =
Remedy
• The staggered grid
– Store pressure at main cell centroid
– Store velocities at staggered control volumes
Staggered grid: Momentum Eq.
v
u
S
j
p
)
v
(
)
v
(
S
i
p
)
u
(
)
u
(
+
⋅
∇
−
∇
μ
⋅
∇
=
ρ
⋅
∇
+
⋅
∇
−
∇
μ
⋅
∇
=
ρ
⋅
∇
V
V
u-momentum
v-momentum
x
)
p
p
(
j
p
y
)
p
p
(
i
p
N
P
E
P
Δ
−
=
⋅
∇
−
Δ
−
=
⋅
−∇
Advantages:
• No need to
Interpolate pressure
• Pressure Checker
boarding avoided
Discretization of Momentum Eq.
• Discretization of other terms are similar to the scalar
transport equation
e
E
P
nb
nb
nb
e
e b
y
)
p
p
(
u
a
u
a +
Δ
−
+
= ∑
n
N
P
nb
nb
nb
n
n b
x
)
p
p
(
v
a
v
a +
Δ
−
+
= ∑
Discretization of Continuity Eq.
• Discretize continuity on main control volume
0
x
)
v
(
x
)
v
(
y
)
u
(
y
)
u
( s
n
w
e =
Δ
ρ
−
Δ
ρ
+
Δ
ρ
−
Δ
ρ
• No need to interpolate velocities
-No checkerboarding

More Related Content

PPTX
deep learning for fluid mechanics using PINNs
PPTX
ECE476_2016_Lect 12.pptx
PPT
Circuitanly
PPT
about power system operation and control13197214.ppt
PPTX
Numerical Solution of the navier stokes equations using the finite difference...
PPTX
IEEE APE
PDF
QX Simulator and quantum programming - 2020-04-28
PPTX
xcvbnmr5tyuio2324354675u86i7o864e75r86t7yo8
deep learning for fluid mechanics using PINNs
ECE476_2016_Lect 12.pptx
Circuitanly
about power system operation and control13197214.ppt
Numerical Solution of the navier stokes equations using the finite difference...
IEEE APE
QX Simulator and quantum programming - 2020-04-28
xcvbnmr5tyuio2324354675u86i7o864e75r86t7yo8

Similar to primitive variable for claculation of flow field (20)

PPTX
MATHS.pptxesrdytfyghjlk;essrudtifykglkdtfygl
PPTX
machine learning.pptx
PPTX
Av 738- Adaptive Filtering - Background Material
PDF
DAC15 Hao Zhuang poster
PDF
Thermodynamicsworkandheatdrkatte
DOCX
Chapter24rev1.pptPart 6Chapter 24Boundary-Valu.docx
PDF
BEF43303 - 201620171 W2 Power System Analysis and Protection.pdf
PDF
5_2021_02_16!07_56_39_PM.pdf
PPTX
EE-433-L2discrete_control_kfupm____.pptx
PPTX
Chapter7_Detailed_Reservoir_Simulation.pptx
PPTX
Load_Flow.pptx
PDF
9 pd es
PPT
This is related to numberical method, in engineering college
PPT
introduction to numerical analysis .ppt
PPT
Load Flow Analysis for electricity purpose
PDF
Newton raphson method
PPTX
Multicomponent system
PPT
2D CFD Code Based on MATLAB- As Good As FLUENT!
PDF
document(1).pdf
PPTX
Episode 50 : Simulation Problem Solution Approaches Convergence Techniques S...
MATHS.pptxesrdytfyghjlk;essrudtifykglkdtfygl
machine learning.pptx
Av 738- Adaptive Filtering - Background Material
DAC15 Hao Zhuang poster
Thermodynamicsworkandheatdrkatte
Chapter24rev1.pptPart 6Chapter 24Boundary-Valu.docx
BEF43303 - 201620171 W2 Power System Analysis and Protection.pdf
5_2021_02_16!07_56_39_PM.pdf
EE-433-L2discrete_control_kfupm____.pptx
Chapter7_Detailed_Reservoir_Simulation.pptx
Load_Flow.pptx
9 pd es
This is related to numberical method, in engineering college
introduction to numerical analysis .ppt
Load Flow Analysis for electricity purpose
Newton raphson method
Multicomponent system
2D CFD Code Based on MATLAB- As Good As FLUENT!
document(1).pdf
Episode 50 : Simulation Problem Solution Approaches Convergence Techniques S...
Ad

More from AjitParwani (7)

PPTX
compuatational fluid dynamics information
PPTX
computational fluid dynamics simulation of solar collector
PDF
mesh generation in computational fluid dynamics
PPTX
reynolds average navier stoke equation for turbulent flow modelling.pptx
PDF
computational fluid dynamics solution of NS equation.pdf
PPT
radiation fundamentals for heat and mass transfer course
PPTX
boiling and condensation lecture presentation
compuatational fluid dynamics information
computational fluid dynamics simulation of solar collector
mesh generation in computational fluid dynamics
reynolds average navier stoke equation for turbulent flow modelling.pptx
computational fluid dynamics solution of NS equation.pdf
radiation fundamentals for heat and mass transfer course
boiling and condensation lecture presentation
Ad

Recently uploaded (20)

DOCX
ASol_English-Language-Literature-Set-1-27-02-2023-converted.docx
PDF
Digital Logic Computer Design lecture notes
PPTX
CH1 Production IntroductoryConcepts.pptx
PPTX
Internet of Things (IOT) - A guide to understanding
PPTX
M Tech Sem 1 Civil Engineering Environmental Sciences.pptx
PDF
TFEC-4-2020-Design-Guide-for-Timber-Roof-Trusses.pdf
PPT
Mechanical Engineering MATERIALS Selection
PDF
BMEC211 - INTRODUCTION TO MECHATRONICS-1.pdf
PPTX
Lecture Notes Electrical Wiring System Components
PPTX
MET 305 2019 SCHEME MODULE 2 COMPLETE.pptx
PPTX
UNIT-1 - COAL BASED THERMAL POWER PLANTS
PDF
PPT on Performance Review to get promotions
PPTX
CARTOGRAPHY AND GEOINFORMATION VISUALIZATION chapter1 NPTE (2).pptx
PPTX
bas. eng. economics group 4 presentation 1.pptx
PPTX
Engineering Ethics, Safety and Environment [Autosaved] (1).pptx
PPTX
OOP with Java - Java Introduction (Basics)
PDF
Enhancing Cyber Defense Against Zero-Day Attacks using Ensemble Neural Networks
PDF
keyrequirementskkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkk
PDF
The CXO Playbook 2025 – Future-Ready Strategies for C-Suite Leaders Cerebrai...
PDF
Well-logging-methods_new................
ASol_English-Language-Literature-Set-1-27-02-2023-converted.docx
Digital Logic Computer Design lecture notes
CH1 Production IntroductoryConcepts.pptx
Internet of Things (IOT) - A guide to understanding
M Tech Sem 1 Civil Engineering Environmental Sciences.pptx
TFEC-4-2020-Design-Guide-for-Timber-Roof-Trusses.pdf
Mechanical Engineering MATERIALS Selection
BMEC211 - INTRODUCTION TO MECHATRONICS-1.pdf
Lecture Notes Electrical Wiring System Components
MET 305 2019 SCHEME MODULE 2 COMPLETE.pptx
UNIT-1 - COAL BASED THERMAL POWER PLANTS
PPT on Performance Review to get promotions
CARTOGRAPHY AND GEOINFORMATION VISUALIZATION chapter1 NPTE (2).pptx
bas. eng. economics group 4 presentation 1.pptx
Engineering Ethics, Safety and Environment [Autosaved] (1).pptx
OOP with Java - Java Introduction (Basics)
Enhancing Cyber Defense Against Zero-Day Attacks using Ensemble Neural Networks
keyrequirementskkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkk
The CXO Playbook 2025 – Future-Ready Strategies for C-Suite Leaders Cerebrai...
Well-logging-methods_new................

primitive variable for claculation of flow field

  • 1. Primitive Variable Formulation MEL 807 Computational Heat Transfer (2-0-4) Dr. Prabal Talukdar Assistant Professor Department of Mechanical Engineering IIT Delhi
  • 3. Sequential Solution • In a sequential approach, we solve set of coupled PDEs in a sequential manner • We identify • u-momentum equation for u velocity component • v-momentum for v velocity component • Continuity as PDE for pressure • Possible solution loop • Discretize and solve u over domain, assuming v, p known from current iterates • Similarly solve v from discrete v-mom equations • Similarly solve p from discrete continuity • Continue until convergence
  • 4. Pressure into Continuity • Look into the problems of introducing pressure into continuity equation for incompressible flows • We have 3 equations (u, v momentum for 2D and continuity) and 3 unknowns (u, v and p) • But how can pressure be linked to continuity? • For compressible flow- not a problem • But for incompressible flow density is constant – No obvious way to introduce pressure into continuity. RT / P 0 ) .( t = ρ = ρ ∇ + ∂ ρ ∂ V
  • 5. Sequential vs. Direct Schemes • Important to recognize that the problem of pressure computation for incompressible flows is an artifact of sequential solution procedure • Only for sequential procedures it is necessary to identify a PDE for each unknown variable – U-mom for u velocity; v-mom for v velocity, continuity equation for pressure; energy equation for temperature etc. • For direct solution procedures, no such one-on-one identification is necessary
  • 6. Direct Scheme • Discretize u, v momentum equations and continuity at all cells • Assemble a big matrix: Ax = b • Solution vector is: x = [u1,u2,----uN,v1,v2,---vN,p1,p2,---pN]T • No need to identify which unknown gets solved using which PDE • Unfortunately too expensive for practical use as of this writing
  • 7. Pressure Based Methods • Sequential methods are preferred for incompressible flows – Low storage and reasonable convergence rate • For incompressible flows, how to obtain pressure from continuity? – Density unrelated to pressure for incompressible flows • Pressure based methods are techniques to introduce pressure into continuity equation
  • 8. Other Methods • Density based methods – Popular in compressible flow community – (u,v,w,ρ,T) used as unknowns • Artificial compressibility/preconditioning schemes – Effectively add some artificial compressibility to enable use of density-based schemes for incompressible flows – We will focus on pressure based methods as they are widely used
  • 9. Representation of the pressure gradient N-S equation ) ( y 3 2 y u x y v y f S ) ( x 3 2 x v y x u x f S S j p ) v ( ) v ( S i p ) u ( ) u ( v v u u v u V V V V ⋅ ∇ μ ∂ ∂ − ⎟ ⎟ ⎠ ⎞ ⎜ ⎜ ⎝ ⎛ ∂ ∂ μ ∂ ∂ + ⎟ ⎟ ⎠ ⎞ ⎜ ⎜ ⎝ ⎛ ∂ ∂ μ ∂ ∂ + = ⋅ ∇ μ ∂ ∂ − ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ ∂ ∂ μ ∂ ∂ + ⎟ ⎠ ⎞ ⎜ ⎝ ⎛ ∂ ∂ μ ∂ ∂ + = + ⋅ ∇ − ∇ μ ⋅ ∇ = ρ ⋅ ∇ + ⋅ ∇ − ∇ μ ⋅ ∇ = ρ ⋅ ∇ Body force Source term
  • 10. Representation of the pressure gradient (cont’d) u S i p ) u ( ) u ( + ⋅ ∇ − ∇ μ ⋅ ∇ = ρ ⋅ ∇ V dx dp − In 1D, Integrating this term over a control volume, we get 2 p p 2 p p 2 p p p p dx dx dp E W E P P W e w − = + − + = − = − ∫ w is assumed midway between P and W, But not a restriction
  • 11. Implications 2 p p dx dp E W − = − Implies that the momentum equation will contain a term of pressure difference of alternate grid points instead of adjacent grid points Is that a problem?? Implication: The pressure is calculated from a coarser grid than the one actually employed -reduces accuracy But there is another serious problem-follows!!!
  • 12. Checkerboard pressure field • Serious Problem: – See the zig-zag field which is unrealistic. The proposed scheme will give PW-PE zero at every grid point – Although it represents a wavy pressure field, proposed scheme will result a uniform pressure field in the momentum equation – Patterns are even more unrealistic in 2D and 3D 0 0
  • 13. Representation of the Continuity Equation • Continuity equation • 1D steady, incompressible flow • Integrating over a CV • Using a piecewise linear profile for u 0 u u 0 2 u u 2 u u 0 u u 0 dx du 0 ) V .( W E P W E P w e = − = + − + = − = = ρ ∇
  • 14. Checkerboard Velocity Field • So, the discretized continuity equation demands the equality of velocities at alternate grid points and not at adjacent ones • Although the following wavy velocity field satisfies the continuity equation, it is unrealistic • Conclusions: first order derivative always creates problem 0 0 U =
  • 15. Remedy • The staggered grid – Store pressure at main cell centroid – Store velocities at staggered control volumes
  • 16. Staggered grid: Momentum Eq. v u S j p ) v ( ) v ( S i p ) u ( ) u ( + ⋅ ∇ − ∇ μ ⋅ ∇ = ρ ⋅ ∇ + ⋅ ∇ − ∇ μ ⋅ ∇ = ρ ⋅ ∇ V V u-momentum v-momentum x ) p p ( j p y ) p p ( i p N P E P Δ − = ⋅ ∇ − Δ − = ⋅ −∇ Advantages: • No need to Interpolate pressure • Pressure Checker boarding avoided
  • 17. Discretization of Momentum Eq. • Discretization of other terms are similar to the scalar transport equation e E P nb nb nb e e b y ) p p ( u a u a + Δ − + = ∑ n N P nb nb nb n n b x ) p p ( v a v a + Δ − + = ∑
  • 18. Discretization of Continuity Eq. • Discretize continuity on main control volume 0 x ) v ( x ) v ( y ) u ( y ) u ( s n w e = Δ ρ − Δ ρ + Δ ρ − Δ ρ • No need to interpolate velocities -No checkerboarding