5.

PRESSURE AND VELOCITY SPRING 2024

5.1 The momentum equation

5.2 Pressure-velocity coupling
5.3 Pressure-correction methods
Summary
References
Examples

5.1 The Momentum Equation

Each component of momentum satisfies its own scalar-transport equation. For one cell:
d ∂𝜙
(mass × 𝜙) + ∑( 𝐶𝜙 −Γ 𝐴 ) = 𝑆
d𝑡 ∂𝑛 (1)
faces
rate of change advection diffusion source
where 𝐶 is the mass flux through a cell face. For the momentum-equation components:
concentration, 𝜙 ← velocity component (𝜙 = 𝑢, 𝑣, 𝑤)
diffusivity, Γ ← viscosity, 𝜇
source ← non-viscous forces

However, the equations for the momentum components differ from those for passive scalars
(those not affecting the flow) because they are:
• non-linear;
• coupled;
• required also to be mass-consistent.

For non-linearity, note that the x-momentum flux through an 𝑥-directed face is
𝐶𝑢 = (𝜌𝑢𝐴)𝑢 v mass flux
u
The mass flux 𝐶 is not constant but changes with 𝑢. The momentum uA
equation is therefore non-linear and must be solved iteratively.

For coupling, note that the 𝑦-momentum flux through an 𝑥-directed face is
𝐶𝑣 = (𝜌𝑢𝐴)𝑣
Again, via the mass flux, the 𝑣 equation depends on the solution of the 𝑢 equation, and vice
versa. Hence, the momentum equations are coupled and must be solved together.

Pressure also appears in the momentum equations. The need to determine it, together with the
fact that velocity components 𝑢, 𝑣, 𝑤 must satisfy mass conservation as well as momentum,
further couples the equations. This distinguishes compressible and incompressible CFD.
• In compressible flow, continuity provides a transport equation for density 𝜌. Pressure
is obtained by solving an energy equation to find temperature T and then using an
equation of state (e.g. 𝑝 = 𝜌𝑅𝑇).
• In incompressible flow, density variations are not determined by pressure. A pressure
equation arises from the requirement that the solutions of the momentum equations are
also mass-consistent; i.e. mass conservation leads to a pressure equation.

CFD 5–1 David Apsley

In most incompressible-flow solvers mass and momentum equations are solved sequentially
and iteratively. This is called a segregated approach, as in the following pseudocode:

DO WHILE (not_converged)
CALL SCALAR_TRANSPORT( 𝑢 )
CALL SCALAR_TRANSPORT( 𝑣 )
CALL SCALAR_TRANSPORT( 𝑤 )

CALL MASS_CONSERVATION
END DO

By contrast, in many compressible-flow codes the main fluid variables are assembled and
solved as a vector (𝜌, 𝜌𝑢, 𝜌𝑣, 𝜌𝑤, 𝜌𝑒). This is called a coupled approach.

5.2 Pressure-Velocity Coupling

Question 1. How are velocity and pressure linked?

Question 2. How does a pressure equation arise?

Question 3. Should velocity and pressure be co-located (stored at the same locations)?

area A

5.2.1 Link Between Pressure and Velocity uP

p p
w e

The discretised momentum equation for one cell is of the form

𝑎𝑃 𝑢𝑃 − ∑ 𝑎𝐹 𝑢𝐹 = 𝐴(𝑝
⏟ 𝑤 − 𝑝𝑒 ) + other forces
⏟ 𝐹 pressure force
(2)
net momentum flux

Hence,
𝐴
𝑢𝑃 = 𝑑𝑃 (𝑝𝑤 − 𝑝𝑒 ) + ⋯ where 𝑑𝑃 = (3)
𝑎𝑃

Answer 1

(a) The force terms in the momentum equation provide a link between velocity and pressure.

(b) Velocity depends on the pressure gradient or, when discretised, on the difference between
pressure values ½ cell either side. Symbolically:
𝑢 = −𝑑Δ𝑝 + ⋯
where Δ indicates a centred difference (“right minus left”).

CFD 5–2 David Apsley

Substituting for velocity in the continuity equation, N
0 = (𝜌𝑢𝐴)𝑒 − (𝜌𝑢𝐴)𝑤 + ⋯ n

W P E
= (𝜌𝐴𝑑)𝑒 (𝑝𝑃 − 𝑝𝐸 ) − (𝜌𝐴𝑑)𝑤 (𝑝𝑊 − 𝑝𝑃 ) + ⋯ w e

= −𝑎𝑊 𝑝𝑊 + 𝑎𝑝 𝑝𝑝 − 𝑎𝐸 𝑝𝐸 + ⋯ s
S
This has the same algebraic form as the scalar-transport equation.

Answer 2

The momentum equation provides a link between velocity and pressure which, when
substituted into the continuity equation, gives an equation for pressure.

Hence, a pressure equation arises from the requirement that solutions of the momentum
equation be mass-consistent.

5.2.2 Co-located Storage of Variables

Suppose pressure and velocity are co-located (stored at the same positions) and that advective
velocities (cell-face velocities in the mass fluxes) are calculated by linear interpolation.

In the momentum equation the net pressure force involves

1 1
𝑝𝑤 − 𝑝𝑒 = (𝑝𝑖−1 + 𝑝𝑖 ) − (𝑝𝑖 + 𝑝𝑖+1 ) i-1 i i+1
2 2 w e
1
= (𝑝𝑖−1 − 𝑝𝑖+1 )
2
Hence, the discretised momentum equation has the form:
1
𝑢𝑖 = 𝑑𝑖 (𝑝𝑖−1 − 𝑝𝑖+1 ) + ⋯
2

In the continuity equation the net outward mass flux depends on

1 1
𝑢𝑒 − 𝑢𝑤 = (𝑢𝑖 + 𝑢𝑖+1 ) − (𝑢𝑖−1 + 𝑢𝑖 )
2 2
i-2 i-1 i i+1 i+2
1 w e
= (𝑢𝑖+1 − 𝑢𝑖−1 )
2
1
= [𝑑𝑖+1 (𝑝𝑖 − 𝑝𝑖+2 ) − 𝑑𝑖−1 (𝑝𝑖−2 − 𝑝𝑖 )] + ⋯
4

So, both mass and momentum equations only link pressures at alternate nodes.
p
Thus, the combination of:
• co-located 𝑢, 𝑝;
• linear interpolation for advective velocities;
leads to decoupling of odd nodal values 𝑝1 , 𝑝3 , 𝑝5 , … from even
nodal values 𝑝2 , 𝑝4 , 𝑝6 , … . This odd-even decoupling or
checkerboard effect leads to oscillations in pressure. x
1 2 3 4 5 6 7 8 9 10

CFD 5–3 David Apsley

There are two common remedies:
(1) use a staggered grid (velocity and pressure stored at different locations); or
(2) use a co-located grid but Rhie-Chow interpolation for the advective velocities.
Both provide a link between adjacent pressure nodes, preventing odd-even decoupling.

5.2.3 Staggered Grid (Harlow and Welch, 1965)

In the staggered-grid arrangement, velocity components are u p

stored half-way between the pressure nodes that drive them.
v

This leads to different sets of control volumes: one

pressure/scalar set for pressure (and other scalars) and others for the
control volume
up pp different velocity components.
u control volume
One common indexing convention (used here) gives
each velocity node the same index (P or 𝑖𝑗𝑘) as the
vp v control volume
pressure node to which it points. An alternative
convention refers to, e.g., the 𝑢𝑖−1/2 velocity.

On a cartesian mesh …
pi-1
• In the momentum equation, pressure is stored at precisely ui pi
the points required to compute the pressure force.
𝑢𝑖 = 𝑑𝑖 (𝑝𝑖−1 − 𝑝𝑖 ) + ⋯
pi-1 ui pi ui+1 pi+1
• In the continuity equation velocity is stored at precisely the
points required to compute mass fluxes. For net mass flux:
𝑢𝑖+1 − 𝑢𝑖 + ⋯ = 𝑑𝑖+1 (𝑝𝑖 − 𝑝𝑖+1 ) − 𝑑𝑖 (𝑝𝑖−1 − 𝑝𝑖 ) + ⋯
= −𝑑𝑖 𝑝𝑖−1 + (𝑑𝑖 + 𝑑𝑖+1 )𝑝𝑖 − 𝑑𝑖+1 𝑝𝑖+1
No interpolation is required for cell-face values and there is a strong linkage between
successive pressure nodes, avoiding odd-even decoupling.

In shallow-water codes, depth ℎ replaces pressure 𝑝 and can also be staggered from velocity.

Advantages
• No interpolation required; on cartesian meshes, variables are stored exactly where
needed.
• No odd-even pressure decoupling.

Disadvantages
• Added geometrical complexity from multiple sets of nodes
and control volumes.
• If the mesh is not cartesian then the velocity nodes may
cease to lie between the pressure nodes that drive them (see
right).
• Very difficult to implement on unstructured meshes.

CFD 5–4 David Apsley

5.2.4 Rhie-Chow Velocity Interpolation (Rhie and Chow, 1983)

The alternative approach uses co-located pressure and velocity W P E

w e
but employs a different, pressure-dependent, interpolation for
advective velocities (cell-face velocities in the mass fluxes).

The momentum equation connects velocity and pressure:

∑ 𝑎𝐹 𝑢𝐹
𝑢𝑃 = − 𝑑𝑃 (𝑝𝑒 − 𝑝𝑤 ) + ⋯ (4)
𝑎𝑃
Symbolically,
𝑢 = 𝑢̂ − 𝑑Δ𝑝 (5)
𝑢̂ is called the pseudovelocity. It represents everything on the RHS except pressure. In the Rhie-
Chow algorithm this symbolic relation is applied at both nodes and faces.

(i) Invert (5) to work out 𝑢̂ at nodes:

𝑢̂ = 𝑢 + 𝑑Δ𝑝face with centred difference Δ𝑝 from interpolated face values

(ii) Then linearly interpolate 𝑢̂ and 𝑑 to cell faces:

𝑢face = 𝑢̂face − 𝑑face Δ𝑝 with centred difference Δ𝑝 taken from adjacent nodes

This amounts to adding and subtracting centred pressure differences worked out at different
places. For example, on the east face, with an overbar denoting linear interpolation to a face:
𝑢𝑒 = (𝑢 + 𝑑Δ𝑝)𝑒 − 𝑑𝑒 (𝑝𝐸 − 𝑝𝑃 ) (6)

Using this interpolative technique, mass conservation gives:

0 = (𝜌𝐴𝑢)𝑒 − (𝜌𝐴𝑢)𝑤 + ⋯
= (𝜌𝐴𝑢̂)𝑒 − (𝜌𝐴𝑢̂)𝑤 − (𝜌𝐴𝑑)𝑒 (𝑝𝐸 − 𝑝𝑃 ) + (𝜌𝐴𝑑)𝑤 (𝑝𝑃 − 𝑝𝑊 ) + ⋯ (7)
= (𝜌𝐴𝑢̂)𝑒 − (𝜌𝐴𝑢̂)𝑤 − 𝑎𝑊 𝑝𝑊 + 𝑎𝑃 𝑝𝑃 − 𝑎𝐸 𝑝𝐸 + ⋯

Notes.
• The central pressure value does not cancel; there is no odd-even decoupling.
• This is a pressure equation. Moreover, it has the same algebraic form as a scalar-
transport equation:
𝑎𝑃 𝑝𝑃 − ∑ 𝑎𝐹 𝑝𝐹 = 𝑏𝑃 , where 𝑎𝐹 ≥ 0 , 𝑎𝑃 = ∑ 𝑎𝐹
• In practice it is common to solve iteratively for pressure corrections rather than pressure
itself (Section 5.3). Equation (7) then becomes
0 = (𝜌𝐴𝑢∗ )𝑒 − (𝜌𝐴𝑢∗ )𝑤 − 𝑎𝑊 𝑝′𝑊 + 𝑎𝑃 𝑝′𝑃 − 𝑎𝐸 𝑝′𝐸 + ⋯
where ∗ here denotes “current value of”. This can be rearranged as:
−𝑎𝑊 𝑝′𝑊 + 𝑎𝑃 𝑝′𝑃 − 𝑎𝐸 𝑝′𝐸 + ⋯ = − current mass outflow (8)

CFD 5–5 David Apsley

Example

For the uniform cartesian mesh shown below the momentum equation gives a velocity/pressure
relationship
𝑢 = −4Δ𝑝 + ⋯
for each cell, where Δ denotes a centred difference.

u= 1 2 3 3
p= pA 0.8 0.7 0.6

For the 𝑢 and 𝑝 values given, calculate the advective velocity on the cell face marked f:
(a) by linear interpolation;
(b) by Rhie-Chow interpolation, if 𝑝𝐴 = 0.6 (local pressure maximum to the left of face);
(c) by Rhie-Chow interpolation, if 𝑝𝐴 = 0.9 (constant pressure gradient).

Answer

(a) By linear interpolation,

1
𝑢𝑓 = 2 (2 + 3) = 2.5

(b) The working for Rhie-Chow may be set out as follows.

│ • │ • │ • │ • │
𝑢 1 2 3 3
𝑝 0.6 0.8 0.7 0.6

𝑝face 0.7 0.75 0.65 by interpolation of 𝑝

𝑢̂ 2.2 2.6 𝑢̂ = 𝑢 + 4(𝑝𝑒 − 𝑝𝑤 )

𝑢̂face 2.4 by interpolation of 𝑢̂
𝑢face 2.8 𝑢face = 𝑢̂face − 4(𝑝𝐸 − 𝑝𝑊 )

Answer: 𝑢𝑓 = 2.8
This is higher than that obtained from linear interpolation, a result of the velocity field trying
to alleviate the local pressure peak just to the left of the face.

(c) If 𝑝𝐴 = 0.9 then Rhie-Chow interpolation produces 𝑢𝑓 = 2.5, as for linear interpolation.

CFD 5–6 David Apsley

Analysis of Rhie-Chow Interpolation

Parts (a) and (c) of the example above show that Rhie-Chow interpolation gives the same result
as linear interpolation if there is a constant pressure gradient, whereas if (as in part(b)) there is
a local pressure peak to one side of the face then the advective velocity increases to try to
reduce it.

For general values of 𝑢𝑖 and 𝑝𝑖 (and a uniform mesh and a constant coefficient 𝑑), Rhie-Chow
interpolation gives (exercise):
1 𝑑
𝑢face = (𝑢2 + 𝑢3 ) + (−𝑝1 + 3𝑝2 − 3𝑝3 + 𝑝4 )
2 4
Thus, Rhie-Chow interpolation adds a “pressure-smoothing” term to the part obtained simply
by interpolation.

Pressure-velocity coupling is the dominant feature of the Navier-Stokes equations. Staggered

grids are an effective way of handling it on cartesian meshes. However, for non-cartesian
meshes, co-located storage is the norm; the Rhie-Chow algorithm is employed in most general-
purpose CFD codes.

A generalised form of the Rhie-Chow algorithm (based on local pressure gradients) is used for
unstructured meshes.

CFD 5–7 David Apsley

5.3 Pressure-Correction Methods

Consider how changing pressure could be used to enforce mass conservation.

Net mass flux in. Net mass flux out.

Increase cell pressure to drive mass out. Decrease cell pressure to suck mass in.

What are pressure-correction methods?

• Iterative numerical schemes for pressure-linked equations.

• Seek to produce velocity and pressure satisfying both mass and momentum equations.
• Consist of alternating updates of velocity and pressure:
– solve the momentum equation for velocity with the current pressure;
– use velocity-pressure link to rephrase continuity as a pressure-correction equation;
– solve for pressure corrections to “nudge” velocity towards mass conservation.
• There are two common schemes: SIMPLE and PISO.

Velocity and Pressure Corrections

The momentum equation connects velocity and pressure:

𝑢 = 𝑑(𝑝−1/2 − 𝑝+1/2 ) + ⋯
where –½ and +½ indicate the relative locations as multiples of the grid spacing.

One must correct velocity to satisfy continuity:

𝑢 → 𝑢∗ + 𝑢′
but simultaneously correct pressure so as to retain a solution of the momentum equation:
′ ′
𝑢′ = 𝑑(𝑝−1/2 − 𝑝+1/2 )+⋯

The velocity-correction formula is, therefore,

′ ′
𝑢 → 𝑢∗ + 𝑑(𝑝−1/2 − 𝑝+1/2 )+⋯ (9)

CFD 5–8 David Apsley

Classroom Example 1 (Patankar, 1980)

A 1 B 2 C 3
In the steady, one-dimensional, constant-density situation shown, the pressure p is stored at
locations 1, 2 and 3, whilst velocity 𝑢 is stored at the staggered locations A, B and C. The
velocity-correction formula is
′
𝑢 = 𝑢∗ + 𝑢′ , where 𝑢′ = 𝑑(𝑝𝑖−1 − 𝑝𝑖′ )
where pressure nodes 𝑖 − 1 and 𝑖 lie on either side of the location for 𝑢. The value of 𝑑 is 2
everywhere. The inflow boundary condition is 𝑢𝐴 = 10. If, at a given stage in the iteration
process, the momentum equations give 𝑢𝐵∗ = 8 and 𝑢𝐶∗ = 11, calculate the values of 𝑝1′, 𝑝2′ , 𝑝3′
and the resulting corrected velocities.

Classroom Example 2

In the 2-d staggered-grid arrangement shown below, 𝑢 and 𝑣 (the 𝑥 and 𝑦 components of
velocity), are stored at nodes indicated by arrows, whilst pressure 𝑝 is stored at the intermediate
nodes A – D. The grid spacing is uniform and the same in both directions. The velocity is fixed
on the boundaries as shown. The velocity components at the interior nodes (𝑢𝐵 , 𝑢𝐷 , 𝑣𝐶 and 𝑣𝐷 )
are to be found.

At an intermediate stage of calculation the internal velocity values are found to be

𝑢𝐵 = 11, 𝑢𝐷 = 14, 𝑣𝐶 = 8, 𝑣𝐷 = 5
whilst correction formulae derived from the momentum equation are
′
𝑢′ = 2(𝑝𝑤 − 𝑝𝑒′ ) , 𝑣′ = 3(𝑝𝑠′ − 𝑝𝑛′ )
with geographical (w,e,s,n) notation indicating the relative location of pressure nodes.

(a) Show that applying mass conservation to control volumes centred on pressure nodes
leads to simultaneous equations for the pressure corrections. Solve for the pressure
corrections and use them to generate a mass-consistent flow field.

5 10

10 C uD D 20

vC vD

5 A uB B 5

y
15 10
x

(b) Explain why, in practice, it is necessary to solve for the pressure correction and not just
the velocity corrections in order to satisfy mass conservation.

CFD 5–9 David Apsley

5.3.1 SIMPLE: Semi-Implicit Method for Pressure-Linked Equations
(Patankar and Spalding, 1972)

Stage 1. Solve the momentum equation with current pressure. START

∗
𝑎𝑃 𝑢𝑝 − ∑ 𝑎𝐹 𝑢𝐹 = 𝐴(𝑝
⏟ 𝑤 − 𝑝𝑒∗ ) + ⏟𝑏𝑃
pressure force other forces
The resulting velocity generally won’t be mass-consistent. solve momentum
equations
Stage 2. Formulate the pressure-correction equation.

(i) Relate changes in 𝑢 to changes in 𝑝: solve pressure-

correction equation
∑ 𝑎𝐹 𝑢𝐹′ 𝐴
𝑢𝑃′ = ′
+ 𝑑𝑃 (𝑝𝑤 − 𝑝𝑒′ ), 𝑑𝑃 =
𝑎𝑃 𝑎𝑃
correct velocity
and pressure
(ii) Make the SIMPLE approximation: neglect ∑ 𝑎𝐹 𝑢′𝐹 .
𝑢𝑃′ ≈ 𝑑𝑃 (𝑝𝑤
′
− 𝑝𝑒′ )
no
(Legitimate, since corrections vanish in the final solution.) converged?

yes
(iii) Apply mass conservation to control volumes centred on the
pressure nodes. The net mass flux results from current (u*) plus END
correction (u) velocity fields:

net mass flow out = ∑ 𝜌𝑢𝑛∗ 𝐴 + ∑ 𝜌𝑢𝑛′ 𝐴 = 0

faces faces

Hence,
(𝜌𝑢′𝐴)𝑒 − (𝜌𝑢′𝐴)𝑤 + ⋯ = −𝑚̇∗ (minus the current net mass flux)
or, writing in terms of the pressure correction (staggered or non-staggered mesh):
(𝜌𝐴𝑑)𝑒 (𝑝𝑃′ − 𝑝𝐸′ ) − (𝜌𝐴𝑑)𝑤 (𝑝𝑊
′
− 𝑝𝑃′ ) + ⋯ = −𝑚̇∗
This results in a pressure-correction equation of the form:
𝑎𝑃 𝑝𝑃′ − ∑𝐹 𝑎𝐹 𝑝𝐹′ = −𝑚̇∗ (10)

Stage 3. Solve the pressure-correction equation

The discretised pressure-correction equation (10) has the same form as the discretised
scalar-transport equation, and hence the same solver may be used.

Stage 4. Correct pressure and velocity:

𝑝𝑃 → 𝑝𝑃∗ + 𝑝𝑃′
(11)
𝑢𝑃 → 𝑢𝑃∗ + 𝑑𝑃 (𝑝𝑤
′
− 𝑝𝑒′ )

Iterate.
Repeat stages 1 to 4 until mass and momentum equations are simultaneously satisfied.

CFD 5 – 10 David Apsley

Notes.
• The source term for the pressure-correction equation is minus the current net mass flux
(−𝑚̇∗ ), as expected. If at any stage there is net mass flow into a control volume then
the pressure there must rise in order to “push” mass back out.

• In practice, substantial under-relaxation of the pressure update is needed to prevent

divergence. In the correction step the pressure (but not the velocity) update is relaxed:
𝑝 → 𝑝∗ + 𝛼𝑝 𝑝′
Typical values of 𝛼𝑝 are 0.1 − 0.3. Velocity is under-relaxed in the momentum
transport equations, but the under-relaxation is generally less severe: 𝛼𝑢 ≈ 0.6 − 0.8.

• As they involve the velocities being computed, matrix elements change at each
iteration. There is little to be gained by solving matrix equations exactly at each stage,
but only doing enough iterations of the matrix solver to reduce the residuals by a
sufficient amount. Alternative strategies at each SIMPLE iteration are:
(1) 𝑚 iterations of each 𝑢, 𝑣, 𝑤 equation, followed by 𝑛 iterations of the 𝑝′ equation
(typically 𝑚 = 1, 𝑛 = 4); or
(2) do enough iterations of each equation to reduce the residual error to a small
fraction of the original (say 10%).

5.3.2 Variants of SIMPLE

The SIMPLE scheme can be inefficient and requires considerable pressure under-relaxation.
This is because the corrected fields are good for updating velocity (since a mass-consistent
flow field is produced) but not pressure (because of the inaccuracy of the approximation
connecting velocity and pressure corrections). To remedy this, a number of variants of SIMPLE
have been produced.

SIMPLER (Patankar, 1980)

This variant acknowledges that the correction equation is good for updating velocity but not
pressure, and precedes the momentum and pressure-correction equations with the equation for
the pressure itself (equation (7)).

SIMPLEC (Van Doormaal and Raithby, 1984)

This scheme seeks a more accurate relationship between velocity and pressure changes. From
the momentum equations, the velocity and pressure equations are related by
1
𝑢𝑃′ = ∑ 𝑎𝐹 𝑢𝐹′ + 𝑑𝑃 (𝑝𝑤
′
− 𝑝𝑒′ )
𝑎
⏟𝑃 (12)
(∗)

The SIMPLE approximation is to neglect the term (*). However, this is actually of comparable
size to the LHS. In the SIMPLEC scheme, (12) is rewritten by subtracting (1/𝑎𝑃 ) ∑ 𝑎𝐹 𝑢𝑃′ from
both sides:
1 1
(1 − ∑ 𝑎𝐹 )𝑢𝑃′ = ∑ 𝑎𝐹 (𝑢𝐹′ − 𝑢𝑃′ ) + 𝑑𝑃 (𝑝𝑤
′
− 𝑝𝑒′ )
𝑎𝑃 𝑎
⏟𝑃
(∗∗)

Assuming that |𝑢𝐹′ − 𝑢𝑃′ | ≪ |𝑢𝑃′ |, it is more accurate to neglect the term (**), thus producing

CFD 5 – 11 David Apsley

an alternative formula connecting velocity and pressure changes:
𝑑𝑃
𝑢𝑃′ ≈ (𝑝′ − 𝑝𝑒′ ) (13)
1 − ∑ 𝑎𝐹 /𝑎𝑃 𝑤
Compare:
𝑢𝑃′ ≈ 𝑑𝑃 (𝑝𝑤
′
− 𝑝𝑒′ ) (SIMPLE)
𝑑𝑃 (14)
𝑢𝑃′ ≈ (𝑝′ − 𝑝𝑒′ ) (SIMPLEC)
1 − ∑ 𝑎𝐹 /𝑎𝑃 𝑤

Although conceptually appealing, there is a difficulty. The “sum-of-the-neighbouring-

coefficients” constraint on 𝑎𝑃 means that, in steady-state calculations, ∑ 𝑎𝐹 /𝑎𝑃 = 1, and hence
the denominator of (13) vanishes. This problem doesn’t arise in time-varying calculations,
where a time-dependent part is added to 𝑎𝑃 , removing the singularity.

SIMPLEX
This scheme assumes that velocity and pressure corrections are linked by some general
relationship
𝑢𝑃′ ≈ 𝛿𝑃 (𝑝𝑤
′
− 𝑝𝑒′ ) (15)
This includes both SIMPLE and SIMPLEC as special cases (see equation (14)). However, the
SIMPLEX scheme attempts to improve on this by actually solving equations for the 𝑃 . These
equations are derived from (12) but we shall not go into the details here.

This author’s experience is that SIMPLER and SIMPLEX offer substantial performance
improvements over SIMPLE on staggered grids, but that the “advanced” schemes are difficult
(impossible?) to formulate on co-located grids and offer little advantage there.

5.3.3 PISO

PISO – Pressure Implicit with Splitting of Operators (Issa, 1986).

This was originally proposed as a time-dependent, non-iterative pressure-correction method.

Each timestep (𝑡 old → 𝑡 new ) consists of a sequence of three stages:
(i) Solution of the time-dependent momentum equation with the 𝑡 old pressure.
(ii) A pressure-correction equation and pressure/velocity update à la SIMPLE to produce a
mass-consistent flow field.
(iii) A second corrector step to produce a second mass-consistent flow field but with time-
advanced pressure.

Apart from time-dependence, steps (i) and (ii) are essentially the same as SIMPLE. However,
step (iii) is designed to eliminate the need for outer iteration at each time step as would be the
case with SIMPLE. At first sight this sounds great; however, the pressure-correction equations
(ii) and (iii) must be solved to a fine tolerance on each pass, so requiring many more inner
iterations of the matrix equations.

Evidence suggests that PISO can be more efficient in some time-dependent calculations, but
SIMPLE and its variants are better in direct iteration to steady state.

CFD 5 – 12 David Apsley

Summary

• Each component of momentum satisfies its own scalar-transport equation:

concentration, 𝜙 ← velocity component (𝑢, 𝑣, 𝑤)
diffusivity, Γ ← viscosity, 𝜇
source, 𝑆 ← non-viscous forces

• However, the momentum equations are:

non-linear;
coupled;
required also to be mass-consistent.
and, as a consequence, have to be solved:
iteratively;
together;
in conjunction with the mass equation.

• For incompressible flow, the requirement that solutions of the momentum equation be
mass-consistent generates a pressure equation.

• Pressure-gradient source terms lead to odd-even decoupling when all variables are co-
located (stored at the same nodes). This may be remedied by using either:
– a staggered velocity grid;
– a non-staggered grid, but Rhie-Chow interpolation for advective velocities.

• Pressure-correction methods make small corrections to pressure in order to “nudge” the

velocity field towards mass conservation whilst still preserving a solution of the
momentum equation.

• Widely-used pressure-correction algorithms are SIMPLE (and its variants) and PISO.
The first is an iterative scheme; the second is a non-iterative, time-dependent scheme.

References

Harlow, F.H. and Welch, J.E., 1965, Numerical calculation of time-dependent viscous
incompressible flow of fluid with a free surface, Physics of Fluids, 8, 2182-2189.
Issa, R.I., 1986, Solution of the implicitly discretised fluid flow equations by operator splitting,
J. Comput. Phys., 62, 40-65.
Patankar, S.V., 1980, Numerical Heat Transfer and Fluid Flow, McGraw-Hill.
Patankar, S.V., 1988, Elliptic systems: finite difference method I, Chapter 6 in Handbook of
Numerical Heat Transfer, Minkowycz, W.J., Sparrow, E.M., Schneider, G.E. and
Pletcher, R.H. (eds.), Wiley.
Raithby, G.D. and Schneider, G.E., 1988, Elliptic systems: finite difference method II, Chapter
7 in Handbook of Numerical Heat Transfer, Minkowycz, W.J., Sparrow, E.M.,
Schneider, G.E. and Pletcher, R.H. (eds.), Wiley.
Rhie, C.M. and Chow, W.L., 1983, A numerical study of the turbulent flow past an isolated
airfoil with trailing edge separation, AIAA Journal, 21, 1525-1532.
Van Doormaal, J.P. and Raithby, G.D., 1984, Enhancements of the SIMPLE method for
predicting incompressible fluid flows, Numerical Heat Transfer, 7, 147-163.

CFD 5 – 13 David Apsley

Examples

Q1.
For the rectangular control volume with surface pressures shown, what is:
A
(a) the net force in the 𝑥 direction? pw pe
(b) the net force in the 𝑥 direction, per unit volume?
(c) the average pressure gradient in the 𝑥 direction?
x

Q2. (1,4)
p=4
The figure right shows a quadrilateral cell in a 2-d mesh,
p=10 (5,3)
together with the coordinates of its vertices and the average
pressures on the cell faces. Calculate the 𝑥 and 𝑦 components
of the net pressure force on the cell (per unit depth). p=2
(-1,1)

p=1
(4,0)

Q3.
(3,5)
The figure shows a triangular cell in a 2-d mesh for an inviscid
CFD calculation. The vertex coordinates and the average
pressures on the cell edges are shown in the figure. The density p=2 p=5
𝜌 = 1.0 everywhere. Find the 𝑥 and 𝑦 components of:
y
(i) the net pressure force on the cell;
(ii) the fluid acceleration. x
(0,0) p=3 (4,0)

Q4.
The figure shows part of a cartesian mesh with the f
velocity u and pressure p at the centre of 4 control
volumes. If the momentum equation leads to a u= 5 4 3 2
pressure-velocity linkage of the form p = 0.6 0.7 1.1 1.6
𝑢 = −3Δ𝑝 + ⋯,
(where Δ represents a centred difference) use the Rhie-Chow procedure to find the advective
velocity on the cell face marked 𝑓.

CFD 5 – 14 David Apsley

Q5.
(a) By considering an infinitesimal cuboid cell with edges aligned with the coordinate axes,
show that the pressure force per unit volume in the 𝑥 direction is −𝜕𝑝/𝜕𝑥.

(b) For what purpose is the Rhie-Chow algorithm used in pressure-based finite-volume
simulations of incompressible fluid flow?

Part of a uniform structured mesh is shown in the figure below, along with the cell-centred
values of the 𝑥-velocity component u and the pressure 𝑝 in consistent units. From the
discretised momentum equation, velocities and pressures are found to be linked by
1
𝑢 = − Δ𝑝 + ⋯
2
where Δ here denotes a centred difference.
w e
i-2 i-1 i i+1 i+2

p= 2 6 5 7 4
u= 1 2 3 4 5

(c) Find the velocity on the ‘w’ and ‘e’ faces of cell i (positions 𝑖 − ½ and 𝑖 + ½) using:
(i) linear interpolation;
(ii) Rhie-Chow interpolation.
Comment briefly on your results and the effect of the pressure field on advective
velocities.

(d) Assuming flow only in the 𝑥 direction, and using the face velocities found in part (c)(ii)
as the current velocity values, set up – but do not solve – the pressure-correction
equation for cell 𝑖.

Q6.
The figure defines the relative position of velocity (→) and pressure (•) nodes in a 1-d,
staggered-grid arrangement.
u1 u2 u3 u4
p1 p2 p3 p4

Velocity 𝑢1 = 4 is fixed as a boundary condition. After solving the momentum equation the
velocities at the other nodes are found to be
𝑢2 = 3, 𝑢3 = 5, 𝑢4 = 6.
The relationship between velocity and pressure is found (from the discretised momentum
equation) to be of the form
𝑢 = −4Δ𝑝 + ⋯
at each node, where Δ denotes a space-centred difference.

Apply mass conservation to cells centred on scalar nodes, calculate the pressure corrections
necessary to enforce continuity, and confirm that a mass-consistent velocity field is obtained.

CFD 5 – 15 David Apsley

Q7. (Exam 2020)
(a) Outline (without mathematical details) the main steps of the SIMPLE pressure-
correction algorithm.

(b) What particular problem can arise from storing pressure and velocity at the same nodal
points in a finite-volume CFD calculation? State two widely-used remedies for this.

In the 2-d staggered-grid arrangement shown below, 𝑢 and 𝑣 (the 𝑥 and 𝑦 components of
velocity) are stored at nodes indicated by arrows, whilst pressure 𝑝 is stored at intermediate
nodes A, B, C, D. Grid spacing is uniform and the same in both directions. Velocity is fixed at
inflow. Upper and lower boundaries are impermeable.

At an intermediate stage of calculation internal velocity components are

𝑢𝐶 = 4, 𝑢𝐷 = −4, 𝑣𝐴 = −2, 𝑣𝐶 = 1
whilst the outflow velocities are
𝑢𝐸 = 6, 𝑢𝐹 = 6,
Velocity-pressure correction formulae are
′
𝑢′ = 2(𝑝𝑤 − 𝑝𝑒′ ) , 𝑣′ = 3(𝑝𝑠′ − 𝑝𝑛′ )
with geographical (w,e,s,n) notation indicating the relative location of pressure nodes.

(c) Apply a uniform scale factor to the outflow velocities to enforce global mass
conservation, stating the scale factor and outflow velocities after its application.

(d) Show that applying mass conservation to control volumes centred on pressure nodes
leads to simultaneous equations for the pressure corrections. Solve for the pressure
corrections and use them to generate a mass-consistent flow field.

5 A uC C uE

vA vC
3 B uD D uF

inflow outflow

CFD 5 – 16 David Apsley

Q8. (Exam 2021)
(a) For what is the Rhie-Chow algorithm used in finite-volume CFD? State,
mathematically, the form of this algorithm for a structured mesh.

Part of a uniform structured mesh is shown in the figure below, along with the cell-centred
values of the 𝑥-velocity component u and the pressure 𝑝 in consistent units. From the
discretised momentum equation, velocities and pressures are found to be linked by
𝑢 = −2Δ𝑝 + ⋯
where Δ here denotes a centred difference.

w e
i-2 i-1 i i+1 i+2

p= 4 5 2 3 0
u= 10 10 10 10 10

(b) Find the velocity on the ‘w’ and ‘e’ faces of cell i (positions 𝑖 − ½ and 𝑖 + ½) using
Rhie-Chow interpolation. What anomaly in the pressure field is the Rhie-Chow
algorithm attempting to alleviate here?

(c) Assuming flow only in the 𝑥 direction, and using the cell-face velocities found in part
(b), set up the pressure-correction equation for cell 𝑖.
′ ′
(d) Assuming that 𝑝𝑖+1 = 𝑝𝑖−1 = 0, calculate 𝑝𝑖′ and use it to correct the pressure at node
𝑖 and the cell-face velocities found in part (b).

CFD 5 – 17 David Apsley