3.

APPROXIMATIONS AND SIMPLIFIED EQUATIONS SPRING 2019

3.1 Steady-state vs time-dependent flow

3.2 Two-dimensional vs three-dimensional flow
3.3 Incompressible vs compressible flow
3.4 Inviscid vs viscous flow
3.5 Hydrostatic vs non-hydrostatic flow
3.6 Boussinesq approximation for density
3.7 Depth-averaged (shallow-water) equations
3.8 Reynolds-averaged equations (turbulent flow)
Examples

Fluid dynamics is governed by equations for mass, momentum and energy. The momentum
equation for a viscous fluid is called the Navier-Stokes equation; it is based upon:
 continuum mechanics;
 the momentum principle;
 shear stress proportional to velocity gradient.

A fluid for which the last is true is called a Newtonian fluid; this is the case for most fluids in
engineering. However, there are important non-Newtonian fluids; e.g. mud, cement, blood,
paint, polymer solutions. CFD is very useful for these, as their governing equations are
usually impossible to solve analytically.

The full equations are time-dependent, 3-dimensional, viscous, compressible, non-linear and
highly coupled. However, in most cases it is possible to simplify analysis by adopting a
reduced equation set. Some common approximations are listed below.

Reduction of dimension:
 steady-state;
 two-dimensional.

Neglect of some fluid property:

 incompressible;
 inviscid.

Simplified forces:
 hydrostatic;
 Boussinesq approximation for density.

Approximations based upon averaging:

 depth-averaging (shallow-water equations);
 Reynolds-averaging (turbulent flows).

The consequences of these approximations are examined in the following sections.

CFD 3–1 David Apsley

3.1 Steady-State vs Time-Dependent Flow

Many flows are naturally time-dependent. Examples include waves,

tides, turbines, reciprocating pumps and internal combustion engines.

Other flows have stationary boundaries but become time-

dependent because of an instability. An important example
is vortex shedding around cylindrical objects. Depending on
the Reynolds number the instability may or may not
progress to fully-developed turbulence.

Some computational solution procedures rely on a time-stepping method to march to steady

state; examples are transonic flow and open-channel flow (where the mathematical nature of
the governing equations is different for Mach or Froude numbers less than or greater than 1).

Thus, there are three major reasons for using the full time-dependent equations:
 time-dependent problem;
 time-dependent instability;
 time-marching to steady state.

Time-dependent methods will be addressed in Section 6.

3.2 Two- Dimensional vs Three-Dimensional Flow

Geometry and boundary conditions may dictate that the flow is two-dimensional. Two-
dimensional calculations require considerably less computer resources.

“Two-dimensional” may include “axisymmetric”. This is easier to achieve in the laboratory

than Cartesian 2-dimensionality, because duct flows have side-wall boundary layers.

Even if the geometry is two-dimensional, instabilities may lead to three-dimensional flow.

Turbulence is always three-dimensional, even if the mean flow is two-dimensional.

3.3 Incompressible vs Compressible Flow

A flow (not a fluid) is said to be incompressible if flow-induced pressure or temperature

changes do not cause significant density changes. Compressibility is important in high-speed
flow or where there is significant heat input.

Liquid flows are usually treated as incompressible, but gas flows can also be regarded as
incompressible at speeds much less than the speed of sound; (a common rule of thumb being
Mach number < 0.3).

Density variations within fluids can occur for other reasons, notably from salinity (oceans)
and temperature (atmosphere). These lead to buoyancy forces. Because the density variations
are not flow-induced these flows can still be treated as incompressible; i.e. “incompressible”
does not necessarily mean “uniform density”.

CFD 3–2 David Apsley

The key differences in CFD between compressible and incompressible flow concern:
(1) whether there is a need to solve a separate energy equation;
(2) how pressure is determined.

Compressible Flow

First Law of Thermodynamics:

A transport equation has to be solved for an energy-related variable (e.g. internal energy e or
enthalpy ) in order to obtain the absolute temperature T. For an ideal gas,
or (1)
cv and cp are specific heat capacities at constant volume and constant pressure respectively.

Mass conservation provides a transport equation for ρ, whilst pressure is derived from an
equation of state; e.g. the ideal-gas law:
(2)

For compressible flow it is necessary to solve an energy equation.

In density-based methods for compressible CFD:

 mass equation → ρ;
 energy equation → T;
 equation of state → p.

Incompressible Flow

In incompressible flow, pressure changes (by definition) cause negligible density changes.
Temperature is not involved and so a separate energy equation is not necessary. The
Mechanical Energy Principle:
change of kinetic energy = work done on fluid
is equivalent to, and readily derived from, the momentum equation. In the inviscid case it is
often expressed as Bernoulli’s equation (see the Examples).

Incompressibility implies that density is constant along a streamline:

(3)

but may vary between streamlines (e.g. due to salinity differences). Conservation of mass is
then replaced by conservation of volume:

or (4)

Pressure is not derived from an equation of state but from the requirement that solutions of
the momentum equation be mass-consistent (Section 5).

CFD 3–3 David Apsley

In incompressible flow it is not necessary to solve a separate energy equation.

In pressure-based methods for incompressible CFD:

 incompressibility → volume is conserved as well as mass;
 requiring solutions of momentum equation to be mass-consistent → equation for p.

3.4 Inviscid vs Viscous Flow

If viscosity is neglected, the Navier-Stokes equations become the Euler equations.

Consider streamwise momentum in a developing 2-d boundary layer:

mass × acceleration = pressure force + viscous force

Dropping the viscous term reduces the order of the highest derivative from 2 to 1 and hence
y
one less boundary condition is required.
U

 Viscous (real) flows require a no-slip (zero-relative-velocity) viscous

condition at rigid walls – the dynamic boundary condition.

y
U
 Inviscid (ideal) flows require only the velocity component normal
to the wall to be zero – the kinematic boundary condition. The inviscid
wall shear stress is zero.

Although its magnitude is small, and consequently its direct influence via the shear stress is
tiny, viscosity can have a global influence out of all proportion to its size. The most important
effect is flow separation, where the viscous boundary layer required to satisfy the non-slip
condition is first slowed and then reversed by an adverse pressure gradient. Boundary-layer
separation has two important consequences:
 major disturbance to the flow;
 a large increase in pressure drag.
separation

rec
irc
u
flo latin
w g

CFD 3–4 David Apsley

3.4.1 Inviscid Approximation: Potential Flow

Velocity Potential, 

In inviscid flow it may be shown1 that the velocity components can be written as the gradient
of a single scalar variable, the velocity potential :

(concisely written: ) (5)

Substituting these into the continuity equation for incompressible flow:

(concisely written: )

gives

which is often written

(L ’s q ) (6)

Stream Function, ψ

In 2-d incompressible flow there exists (Section 9) a stream function ψ such that:

(7)

If the flow is also inviscid then it may be shown that the fluid is irrotational and

Substituting the expressions for u and v gives an equation for ψ:

(L ’s q ) (8)

In both cases above the entire flow is completely determined by a single scalar field ( or ψ)
satisfying Laplace’s equation. Since Laplace’s equation occurs in many branches of physics
(electrostatics, heat conduction, gravitation, optics, ...) many good solvers already exist.

Velocity components u, v and w are obtained by differentiating  or ψ. Pressure is then

recovered from Bernoulli’s equation:

(9)

where U is the magnitude of velocity.

1
Since pressure acts perpendicularly to a surface and cannot impart rotation, an inviscid fluid can be regarded as
irrotational ( ), and so the velocity field can be written as the gradient of a scalar function.

CFD 3–5 David Apsley

The potential-flow assumption often gives an adequate description of velocity and pressure
fields for real fluids, except very close to solid surfaces where viscous forces are significant.
It is useful, for example, in calculating the lift force on aerofoils and in wave theory
(Hydraulics 3). However, in ignoring viscosity it implies that there are neither tangential
stresses on boundaries nor flow separation, which leads to the erroneous conclusion
(D’A mb ’s P x) that an object moving through a fluid experiences no drag.

3.5 Hydrostatic vs Non-Hydrostatic Flow

The equation for the vertical component of momentum can be written:

(10)

For large horizontal scales the vertical acceleration Dw/Dt is much less than g and the viscous
forces are small. The balance of terms is then the same as in a stationary fluid:

pressure forces balance weight

With this hydrostatic approximation, in constant-density flows with a free surface the
pressure is determined everywhere by the height of the free surface h(x,y):
(11)
This results in huge computational savings because the position of the surface automatically
determines the pressure field without the need to solve a separate pressure equation.

patm

h-z

h(x) p  p atm  ρg (h  z )

The hydrostatic approximation is widely used in conjunction with the depth-averaged

shallow-water equations (see Section 3.7 below and the Shallow-Flows part of the course).
However, it is not used in general-purpose flow solvers because it limits the type of flow that
can be computed.

CFD 3–6 David Apsley

3.6 Boussinesq Approximation for Density2

Density variations may arise at low speeds because of changes in temperature or humidity
(atmosphere), or salinity (oceans) which give rise to buoyancy forces. The effect of these
density changes can be significant even if the fractional change in density is small.

The Boussinesq approximation retains density variations in the gravitational term (giving
buoyancy forces) but disregards them in the inertial (mass  acceleration) term; i.e. in the
vertical momentum equation:

(12)

the Boussinesq approximation is simply to replace ρ on the LHS by the constant density ρ0.
The approximation is valid if relative density variations are not too large; i.e. .
This condition is usually satisfied in the atmosphere and oceans.

On the RHS of the momentum equation, the part of the weight resulting from the constant
reference density ρ0 is usually subsumed into a modified pressure , so that

(13)

The relative change in density is typically proportional to the change in some scalar θ (e.g.
temperature or salinity):

(14)

where α is the coefficient of expansion. (The sign adopted here is that for temperature, where
an increase in temperature leads to a reduction in density; the opposite sign would be used for
salinity-driven density changes.) The vertical momentum equation can then be written

where (15)

Temperature variations in the atmosphere, brought about by surface (or cloud-top) heating or
cooling, are responsible for significant changes in airflow and turbulence.
stable boundary layer
 On a cold night the atmosphere is stable. Cool, dense air
collects near the surface and vertical motions are  u
suppressed; the boundary-layer depth is 100 m or less. mixing depth

 On a warm day the atmosphere is unstable. Surface

convective boundary layer
heating produces warm; light air near the ground and
convection occurs; the boundary layer may be 2 km deep.
 u

Vertical dispersion of pollution is very different in the two cases.

2
Note that several other very-different approximations are also referred to as the Boussinesq approximation in
different contexts – e.g. shallow-water equations or eddy-viscosity turbulence models.

CFD 3–7 David Apsley

3.7 Depth-Averaged (Shallow-Water) Equations

This approximation is used for flow of a constant-density fluid with a free surface, where the
depth of fluid is small compared with typical horizontal scales.

In this “hydraulic” approximation, the fluid can be regarded z

as quasi-2d with: h(x,t)
 horizontal velocity components u, v;
u
 depth of water, h.

Note that the depth h may vary due to changes in the levels of
the free-surface, the bed, or both. x

By applying mass and momentum principles to a vertical column of constant-density fluid of

variable depth h, the depth-integrated equations governing the motion can be written (for the
one-dimensional case and in conservative form) as:

(16)

(17)

The term comes from (1/ρ times) the hydrostatic pressure force per unit width on a
water column of height h; i.e.
average hydrostatic pressure ( )  area (h  1)

The final term is the net effect of surface stress (due to wind) and bed shear stress (due to
friction). These equations are derived in the Examples and in the Shallow-Flows course.

The resulting shallow-water equations are mathematically similar to those for a compressible
gas. There are direct analogies between
 discontinuities: hydraulic jumps (shallow flow) and shocks (compressible flow);
 critical flow through a venturi (shallow) or gas flow through a throat (compressible).

In both cases there is a characteristic wave speed ( in the hydraulic case;

in compressible flow). Whether this is greater or smaller than the flow velocity determines
whether disturbances can propagate upstream and hence the nature of the flow. The ratio of
flow speed to wave speed is known as:

Froude number: in shallow flow

Mach number: in compressible flow

CFD 3–8 David Apsley

3.8 Reynolds-Averaged Equations (Turbulent Flow)

The majority of flows encountered in engineering are turbulent. Most, however, can be
regarded as time-dependent, three-dimensional fluctuations superimposed on a much simpler
(and often steady) mean flow. Usually, we are only interested in mean quantities, rather than
details of the time-varying flow.

The process of Reynolds-averaging (named after Osborne Reynolds3, first Professor of

Engineering at Owens College, later to become the Victoria University of Manchester) starts
by decomposing each flow variable into mean and turbulent parts:

m
mean fluctuation
The “mean” may be a time average (the usual case in the laboratory) or an ensemble average
(a probability mean over a hypothetical large number of identical experiments).

When the Navier-Stokes equation is averaged, the result is (see Section 7):
 an equivalent equation for the mean flow,
except for
 turbulent fluxes, etc. (called the Reynolds stresses) which provide a net
transport of momentum.

For example, the viscous shear stress

is supplemented by an additional (and usually much larger) turbulent stress:

(18)

In order to solve the mean-flow equations, a turbulence model is required to supply these
turbulent stresses. Popular models exploit an analogy between viscous and turbulent transport
and employ an eddy viscosity μt to supplement the molecular viscosity. Thus,

(19)

This is readily incorporated into the mean momentum equation via a (position-dependent)
effective viscosity. However, actually specifying μt is by no means trivial – see the lectures on
turbulence modelling (Section 8).

3
Osborne Reynolds’ experimental apparatus – including that used in his famous pipe-flow experiments – is on
display in the basement of the George Begg building at the University of Manchester. A modern replica is in the
George Begg foyer.

CFD 3–9 David Apsley

Examples

Q1.
Discuss the circumstances under which a fluid flow can be approximated as:
(a) incompressible;
(b) inviscid.

Q2.
s
By resolving forces along a streamline, the steady-state momentum
equation for an inviscid fluid can be written U


where U is the velocity magnitude, s is the distance along a streamline and α is the angle
between local velocity and the horizontal. Assuming incompressible flow, derive Bernoulli’s
equation.

Q3.
A velocity field is given by the velocity potential .
(a) Calculate the velocity components u and v.
(b) Calculate the acceleration.
(c) Calculate the corresponding streamfunction, ψ.
(d) Sketch the streamlines and suggest a geometry in which one might expect this flow.

Q4.
For incompressible flow in a rotating reference frame the force per unit volume, f, is the sum
of pressure, gravitational, Coriolis and viscous forces:

where ez is a unit vector in the z direction and is the angular velocity of the rotating frame.

(a) If the density is uniform, show that pressure and gravitational forces can be combined
in a piezometric pressure (which should be defined).

(b) If the density varies, describe the “Boussinesq” approximation in this context and give
an application in which it is used.

(c) Show how the momentum equation (with Boussinesq approximation for density) can
be non-dimensionalised in terms of densimetric Froude number, Rossby number and
Reynolds number:

where ρ0, L0, U0 are characteristic density, length and velocity scales, respectively,
and Δρ is a typical magnitude of density variation.

CFD 3 – 10 David Apsley

Q5. (Exam 2015 – part)
The vector momentum equation for a viscous fluid of variable density is

(*)

where t is time, ρ is density, u = (u,v,w) is velocity, p is pressure, μ is dynamic viscosity, g is

the acceleration due to gravity and ez is a unit vector in the z direction.

(a) Define the operator D/Dt mathematically and explain its physical significance.

(b) Show that, for a constant density ρ0, the pressure and gravitational terms can be
combined as a single gradient term involving the piezometric pressure.

(c) If density variations occur as a result of temperature changes, then

where θ is temperature and α is the coefficient of thermal expansion. Describe the

Boussinesq approximation in this context, apply it in Equation (*), and state the
conditions under which it is valid.

(d) Show that, with the Boussinesq approximation, Equation (*) can be non-
dimensionalised as

where all variables are now non-dimensional, and Re and Fr are, respectively, the
Reynolds number and densimetric Froude number (both to be defined).

CFD 3 – 11 David Apsley

Q6.
(a) In flow with a free surface, by taking a control volume as a column of (time-varying)
height h(x,y,t) and horizontal cross-section Δx  Δy, assuming that density is constant,
τ13 and τ23 are the only significant stress components, the horizontal velocity field
may be replaced by the depth-averaged velocity (u, v) and the pressure is hydrostatic,
derive the shallow-water equations for continuity and x-momentum in the form

(b) Provide an alternative derivation by integrating the continuity and horizontal

momentum equations for incompressible flow:

over a depth h = zs – zb.

For part (b) you will need the boundary condition that the top and bottom surfaces
and are material surfaces:

or on z = zs

and similarly for zb, together with Leibniz’ Theorem for differentiating an integral:

Note: this is easily extended to consider additional forces such as Coriolis forces and other
stress-like terms (e.g. “horizontal diffusion”). This will be covered in the Shallow-Flows part
of the course.

CFD 3 – 12 David Apsley