CFD Simulation Project
LAMINAR FLOW BETWEEN TWO PARALLEL PLATES WITH
A UNIFORM HEAT FLUX (UHF)
Vipul P. Deshpande | Computational Fluid Dynamics 10th
April 2026
DISCLAIMER
The ideas, analysis and technical content of this project are original work by
me(Vipul Deshpande). An AI-based assistant was used only to improve the
readability, language and presentation of the report; it was not used to generate or
verify the underlying technical results.
ABSTRACT
Laminar flow between two parallel plates subjected to a uniform heat flux at their
walls is a classical benchmark problem in fluid mechanics and heat transfer and a
useful analogue for coolant channels in research reactor fuel assemblies. In this
work, a fully developed, incompressible, laminar flow of water between two
parallel plates separated by a distance 2𝐻 is analysed using both analytical
methods and numerical simulation in ANSYS Fluent.
The continuity, momentum and energy equations are simplified using standard
assumptions and solved to obtain closed-form expressions for the velocity profile,
the pressure gradient and the temperature distribution, from which the Nusselt
number is derived.
A two-dimensional CFD model is then built in ANSYS Workbench, including the
construction of the channel geometry, mesh generation, specification of material
properties, boundary conditions, and solver settings with the energy equation
enabled.
Velocity, pressure, and temperature contours and profiles from Fluent are
compared against the analytical solution and the agreement in the parabolic
velocity profile and the small pressure loss and temperature rise between inlet and
outlet demonstrates that Fluent accurately reproduces the theoretical solution for
this laminar benchmark.
INTRODUCTION
Internal flows between parallel plates occur in many engineering systems such as
compact heat exchangers, micro-channels and the coolant channels between fuel
plates in research reactors. When the flow is laminar and fully developed and the
walls are subjected to a uniform heat flux, the problem admits a well-known
analytical solution for both the velocity and temperature fields, which makes it an
ideal validation case for CFD codes.
In nuclear engineering, accurate prediction of coolant velocity and temperature
distributions in fuel plate channels is critical for ensuring that cladding and fuel
temperatures remain below safety limits. Before performing detailed simulations of
full reactor geometries with complex physics, it is good practice to verify the CFD
setup using simplified canonical problems where the correct solution is known.
In this study, a laminar, fully developed water flow between two parallel plates
with uniform wall heat flux is analysed first using Analytical methods and then
using ANSYS Fluent. The main objectives are:
• To derive the velocity and temperature distributions between parallel plates
from the governing equations under appropriate assumptions.
• To set up and solve the same problem numerically in ANSYS Fluent,
including both momentum and energy equations.
• To compare the analytical and numerical results (velocity profile, pressure
gradient, temperature field, and Nusselt number) and discuss sources of any
discrepancies.
PROBLEM DESCRIPTION AND MODELING
ASSUMPTIONS
2.1 Geometry and Co-ordinate system
The physical domain consists of two infinite, parallel, rigid plates separated by a
total distance 2𝐻, with a fluid flowing in the 𝑥-direction between them. The plates
are assumed to be wide in the spanwise 𝑧-direction and long in the streamwise 𝑥-
direction, such that edge effects are negligible and the flow can be treated as
two-dimensional in the 𝑥–𝑦 plane.
A Cartesian coordinate system is adopted where:
• 𝑥 is the streamwise direction along the channel.
• 𝑦 is the transverse direction, measured from the mid-plane, so that the plates
are located at 𝑦 = ±𝐻.
• 𝑧 is the spanwise direction, along which the flow is assumed uniform and
derivatives vanish.
In the numerical model, a finite channel length of 60 mm and a plate separation
corresponding to a channel width of 14 mm are used to approximate the theoretical
infinite-plate case.
2.2 Fluid and thermal conditions
The working fluid is liquid water at
approximately 27∘ C, with thermophysical
properties taken as constant:-
Density ρ=996.5 kg/m3
Viscosity 𝜇= 8.8 10-4 kg/m·s
Thermal conductivity k = 0.612 W/m·K,
Prandtl number 𝑃𝑟 = 5.72
A uniform inlet velocity of 𝑢in = 0.005 m/s
Inlet temperature 𝑇in = 300 K
Both plates are subjected to a uniform heat
flux 𝑞 ′′ = 20W/m2 ,
which heats the fluid as it flows downstream. The outlet is maintained at a static
pressure of approximately 5 Pa, and the resulting pressure distribution along the
channel is computed by Fluent and compared qualitatively with the analytical
pressure gradient.
2.3 Assumptions and justification
To simplify the governing equations and obtain closed-form solutions, the
following assumptions are made:-
1. Steady flow: All time derivatives are zero, so the flow has reached a steady
regime and does not change with time.
2. Incompressible, constant-property fluid: Density, viscosity, specific heat
and thermal conductivity are treated as constants.
3. Newtonian fluid: The fluid obeys a linear relationship between shear stress
and strain rate, so the standard Navier–Stokes equations apply.
4. Laminar regime: The Reynolds number based on hydraulic diameter and
mean velocity is sufficiently low for laminar flow, so no turbulence model is
needed in Fluent.
5. Fully developed flow: The velocity profile no longer changes in the
streamwise direction, so ∂𝑢/ ∂𝑥 = 0.
6. Two-dimensional flow: Variations in the spanwise 𝑧-direction are neglected
(∂/ ∂𝑧 = 0), and the velocity components in 𝑧 are zero.
7. Negligible body forces: Gravity effects are neglected in the momentum
equation because the channel is horizontal and pressure gradients due to
weight are small compared to those induced by the imposed pressure
difference.
8. Uniform wall heat flux: Both plates are subjected to the same uniform heat
flux 𝑞 ′′ along their length, which is a standard idealization for analytical
solutions and aids comparison with CFD.
These assumptions are consistent with the analytical development and with the
CFD setup implemented in ANSYS Fluent.
THEORETICAL BACKGROUND
3.1 Governing equations
Under the stated assumptions, the governing equations for incompressible flow and
heat transfer are:
• Continuity (mass conservation):
∂𝑢 ∂𝑣 ∂𝑤
⃗ =
∇⋅𝑉 + + =0
∂𝑥 ∂𝑦 ∂𝑧
With fully developed, two-dimensional flow and 𝑣 = 𝑤 = 0, this reduces to
∂𝑢
=0
∂𝑥
indicating that the axial velocity depends only on 𝑦.
• Momentum (Navier–Stokes) equation for incompressible Newtonian
fluid:
⃗ ⋅ ∇𝑉
𝜌(𝑉 ⃗ ) = −∇𝑝 + 𝜇∇2 𝑉
⃗
For steady, fully developed, two-dimensional flow between plates, the
x-momentum equation simplifies to
∂2 𝑢 ∂𝑝
𝜇 =
∂𝑦 2 ∂𝑥
while the y-momentum equation yields ∂𝑝/ ∂𝑦 = 0, so pressure varies only in the
streamwise direction.
• Energy equation (with negligible radiation and viscous dissipation):
∂𝑇 ∂𝑇 ∂2 𝑇 ∂2 𝑇
𝜌𝑐𝑝 (𝑢 + 𝑣 ) = 𝑘 ( 2 + 2)
∂𝑥 ∂𝑦 ∂𝑥 ∂𝑦
With 𝑣 = 0 and axial conduction neglected, this becomes
∂𝑇 ∂2 𝑇
𝑢 =𝛼 2
∂𝑥 ∂𝑦
where 𝛼 = 𝑘/(𝜌𝑐𝑝 ) is the thermal diffusivity.
3.2 Relevant dimensionless numbers
To characterize the flow and heat transfer, the Reynolds number, Prandtl number,
and Nusselt number are defined:
• Reynolds number:
𝜌𝑢𝑚 𝐷ℎ
𝑅𝑒 =
𝜇
where 𝑢𝑚 is the mean velocity and 𝐷ℎ is the hydraulic diameter.
• Prandtl number:
𝜇𝑐𝑝
𝑃𝑟 =
𝑘
For water at 27∘ C, 𝑃𝑟 ≈ 5.72.
• Nusselt number:
ℎ𝐷ℎ
𝑁𝑢 =
𝑘
where ℎ is the convective heat transfer coefficient at the wall.
For fully developed laminar flow with uniform heat flux between parallel plates,
the analytical Nusselt number is a constant that depends only on geometry and
flow conditions, not on Reynolds number.
ANALYTICAL SOLUTION FOR FULLY DEVELOPED LAMINAR
FLOW
4.1 Velocity distribution between parallel plates
Starting from the simplified x-momentum equation:
∂2 𝑢 ∂𝑝
𝜇 2=
∂𝑦 ∂𝑥
The right-hand side is a constant (uniform pressure gradient) in the fully developed
region, so integrating twice with respect to 𝑦 gives:
1 𝑑𝑝 2
𝑢(𝑦) = 𝑦 + 𝐴1 𝑦 + 𝐴2
2𝜇 𝑑𝑥
Applying the no-slip boundary conditions at the walls, 𝑢(𝑦 = 𝐻) = 0 and 𝑢(𝑦 =
−𝐻) = 0, yields a system of equations to determine 𝐴1 and 𝐴2 . Solving these
1 𝑑𝑝
gives 𝐴1 = 0 and 𝐴2 = − 𝐻2 , leading to:
2𝜇 𝑑𝑥
1 𝑑𝑝 2
𝑢(𝑦) = (𝑦 − 𝐻2 )
2𝜇 𝑑𝑥
𝑑𝑝
Since the pressure decreases in the direction of flow, < 0, and the velocity
𝑑𝑥
profile becomes:
𝐻2 𝑑𝑝 𝑦2
𝑢(𝑦) = ( − ) (1− 2 )
2𝜇 𝑑𝑥 𝐻
This is a parabolic profile symmetric about the mid-plane.
4.2 Mean velocity, pressure gradient, and flow rate
The mean velocity is defined as the cross-sectional average:
1 𝐻
𝑢𝑚 = ∫ 𝑢(𝑦) 𝑑𝑦
2𝐻 −𝐻
Substituting the velocity distribution and integrating over the gap yields a relation
between mean velocity and pressure gradient:
𝑑𝑝 3𝜇𝑢𝑚
=−
𝑑𝑥 𝐻2
Thus, for a given mean velocity and channel height, the required pressure drop per
unit length can be computed directly. The maximum velocity occurs at the
centreline (𝑦 = 0):
3
𝑢max = 𝑢𝑚
2
In this study, the mean velocity is 𝑢𝑚 = 0.005 m/s, so the analytical maximum
velocity is
3
𝑢max = × 0.005 = 0.0075 m/s
2
which will be compared directly to the CFD results.
4.3 Temperature distribution for uniform heat flux
With the energy equation in the form:
∂𝑇 ∂2 𝑇
𝑢(𝑦) =𝛼 2
∂𝑥 ∂𝑦
and substituting the velocity profile 𝑢(𝑦), the equation is integrated with respect
to 𝑦 twice, subject to the boundary conditions of equal wall temperature at both
plates 𝑇(±𝐻) = 𝑇𝑠 (𝑥). The resulting expression shows that the local temperature
distribution can be expressed in terms of the mean and surface temperatures as:
𝑢𝑚 𝑑𝑇𝑚
𝑇(𝑦) − 𝑇𝑠 = 𝑓(𝑦, 𝐻)
𝛼 𝑑𝑥
where 𝑓(𝑦, 𝐻) is a polynomial in 𝑦/𝐻, indicating that both the velocity and
temperature profiles are smooth functions of the transverse coordinate. The
analysis further leads to expressions relating the mean temperature gradient,
surface temperature, and the applied heat flux.
4.4 Nusselt number for fully developed laminar flow
The local heat transfer coefficient is defined as:
𝑘 ∂𝑇
ℎ=− ∣
𝑇𝑠 − 𝑇𝑚 ∂𝑦 𝑦=𝐻
The Nusselt number is then computed as:
ℎ𝐷ℎ
𝑁𝑢 =
𝑘
Using the analytical temperature distribution for uniform heat flux between parallel
plates, the Nusselt number evaluates to:
140
𝑁𝑢 = ≈ 8.2353
17
This value serves as the analytical reference for comparison with empirical
correlations and CFD predictions.
CFD Modelling Using ANSYS Fluent
5.1 Geometry creation
The computational domain is a two-dimensional rectangle representing the fluid
region between the plates, with a length of 60 mm in the 𝑥-direction and a gap
width corresponding to 2𝐻 in the 𝑦-direction, here taken as 14 mm. In ANSYS
Workbench, the geometry is created in a pre-processor (e.g., DesignModeler or
SpaceClaim) by sketching a rectangle of the specified dimensions and designating
the two long edges as the heating walls and the short edges as inlet and outlet
boundaries.
5.2 Mesh generation
The domain is discretized using a structured or semi-structured mesh of
quadrilateral elements, refined in the transverse direction to resolve the velocity
and temperature gradients near the walls. A suitable first attempt is to bias the
mesh towards each wall with smaller cell heights so that near-wall gradients are
captured accurately without requiring an excessively fine mesh in the central
region. In a more advanced version, a mesh independence study can be performed
by comparing solutions for progressively refined meshes until the predicted
centreline velocity and wall heat transfer coefficients change negligibly.
5.3 Material properties and boundary conditions
In Fluent, water at 27∘ C is defined as the working fluid, with density, viscosity,
and thermal properties as given in Section 2.2. The following boundary conditions
are applied:
• Velocity inlet: uniform axial velocity 𝑢 = 0.005 m/s and temperature 𝑇 =
300 K.
• Pressure outlet: static pressure 𝑝 = 5 Pa, allowing Fluent to determine the
necessary pressure gradient within the channel.
• Walls: no-slip condition for velocity ( 𝑢 = 0 at the walls) and uniform heat
flux 𝑞 ′′ = 20 W/m2 specified as a thermal
boundary condition.
These conditions match the assumptions used to derive the analytical solution and
ensure a fair comparison between theory and simulation.
5.4 Solver settings and convergence
The flow is solved as a steady, laminar, incompressible problem with the energy
equation enabled. A pressure-based segregated solver is used, with appropriate
spatial discretization schemes (at least second-order for momentum and energy) to
minimize numerical diffusion.
Convergence is monitored using residuals for continuity, momentum, and energy,
as well as by observing the stabilization of key integral quantities such as mass
flow rate and area-averaged outlet temperature. In the reported simulation,
residuals decrease steadily and the final solution is considered converged based on
the iteration history shown in the residual plot.
RESULTS AND DISCUSSION
6.1 Hydrodynamic field
The velocity vector and contour plots from Fluent show a fully developed
parabolic velocity profile between the plates in the central region of the channel.
Near the inlet, the profile adjusts from the top-hat inlet profile (if specified)
towards the fully developed shape, but this entrance region is short compared to
the domain length. In the core region, the transverse variation of velocity matches
the analytical prediction of a symmetric parabola with zero wall velocity and
maximum centreline velocity.
The pressure field decreases monotonically from inlet to outlet, as expected for a
pressure-driven laminar flow in a straight channel. The pressure gradient is nearly
uniform in the fully developed region, consistent with the analytical relation
between pressure drop, viscosity, and mean velocity.
6.2 Thermal field
Temperature contours show a mild increase in fluid temperature along the flow
direction due to the imposed wall heat flux. The temperature rise is small (on the
order of 0.07 K from inlet to outlet), reflecting the relatively low heat flux and
limited residence time of the fluid in the channel.
The temperature distribution across the gap is consistent with expectations for
laminar forced convection with uniform heat flux, with higher temperatures near
the heated walls and a flatter profile near the centreline. Temperature gradient plots
emphasize that the strongest gradients occur near the walls where conduction
balances the imposed heat flux.
6.3 Quantitative CFD results
Fluent’s functional calculator yields the following numerical statistics in the fully
developed region:
• Velocity: minimum 0.00075 m/s, maximum 0.00750 m/s, mean 0.00500
m/s.
• Temperature: minimum 300.0000 K, maximum 300.0690 K, mean 300.0180
K.
• Temperature gradient: minimum 0.0 K/m, maximum 26.582 K/m, mean
11.534 K/m.
• Pressure gradient: minimum 0.110 Pa/m, maximum 8.295 Pa/m, mean 0.237
Pa/m.
The maximum velocity of 0.0075 m/s matches exactly the analytical
3
prediction 𝑢max = 𝑢𝑚 = 0.0075 m/s, demonstrating excellent agreement in the
2
hydrodynamic solution. The average temperature rise and heat flux are consistent
with the specified boundary conditions and the derived Nusselt number.
6.4 Comparison with analytical solution and correlations
The CFD velocity profile is in very close agreement with the analytical parabolic
profile, both in shape and magnitude. The centreline velocity and mean velocity
values match the theoretical relations, confirming that Fluent accurately solves the
laminar momentum equations for this geometry.
For heat transfer, the analytical Nusselt number of 8.2353 is compared with
empirical correlations suitable for laminar flow in parallel plate channels, which
yield approximate values of 6.8 and 7.8, as well as with the effective Nusselt
number implied by the CFD solution. The differences among these values are
small and primarily attributed to the treatment of hydraulic diameter and slight
differences in geometry idealization.
The paper notes that the slight mismatch between the analytical and “numerical”
Nusselt numbers can be ascribed to errors in the Reynolds number calculation,
particularly through the choice of hydraulic diameter for the parallel-plate
geometry. In practice, such small discrepancies are acceptable and expected when
comparing continuous analytical solutions, semi-empirical correlations, and
discrete numerical approximations.
6.5 Discussion of accuracy and sources of error
Overall, the CFD solution reproduces the analytical velocity profile and associated
pressure gradient with high fidelity, indicating that the mesh, boundary conditions,
and solver settings are appropriate for this laminar benchmark. The temperature
field also behaves as expected, with a slight but measurable increase from inlet to
outlet and higher temperatures near the heated walls.
Potential sources of error and uncertainty include:
• Finite domain length and entrance effects, whereas the analytical solution
assumes an infinitely long, fully developed region.
• Mesh resolution, especially near walls where gradients are steep; inadequate
refinement could slightly distort local gradients and heat transfer
coefficients.
• Use of averaged material properties for water, which may introduce minor
deviations if the actual fluid properties vary with temperature.
• Discretization error from the numerical schemes used in Fluent, although
this is small for laminar, well-resolved flows.
Despite these factors, the agreement between analytical predictions and CFD
results is strong, supporting the use of this case as a validation benchmark for
ANSYS Fluent.
Iteration Graph
Velocity vector contour
Temperature Gradient contour
Pressure Gradient contour
Pressure Distribution Velocity distribution
Temperature Distribution Velocity (u) Distribution
Wall adjacent Temperature Distribution Velocity (v) Distribution
Vipul Deshpande CFD Simulation Project
CONCLUSIONS
A detailed analytical and numerical study of fully developed laminar flow
between two parallel plates with uniform wall heat flux has been presented.
The analytical solution of the simplified Navier–Stokes and energy
equations yields a parabolic velocity profile and a constant Nusselt number
representative of laminar forced convection between parallel plates.
A two-dimensional ANSYS Fluent simulation using a laminar model and the
energy equation accurately reproduces the analytical velocity distribution,
maximum velocity, and qualitative pressure gradient, demonstrating correct
implementation of the governing equations in the CFD solver for this class
of flows. The predicted temperature rise and heat transfer behaviour are
consistent with the modest heat flux applied and with the theoretical Nusselt
number, with only slight discrepancies attributed to geometric idealizations
and hydraulic diameter definitions.
Page 1
Vipul Deshpande CFD Simulation Project
REFERENCES
1. ME 350 Heat Transfer lecture notes, California Polytechnic State
University, Example of laminar flow between parallel plates.
2. MIT Course Notes on fluid mechanics, Spring 2009.
3. Bonfanti F. et al., “Two Phase Pressure Drops in the Low Flow Rate
Region,” Energia Nucleare, 26(10), pp. 481–492, 1979.
4. Filip A. et al., “Comparison of Two Phase Pressure Drop Models for
Condensing Flows in Horizontal Tubes,” Mathematical Modeling in
Civil Engineering, 10(4), 2014.
5. Brada J., “Mathematical Model of Two Phase Flows,” WM’01
Conference, Tucson, AZ, 2001.
6. Chisholm D., “Pressure Gradients Due to Friction during the Flow of
Evaporating Two-Phase Mixtures in Smooth Tubes and
Channels,” International Journal of Heat and Mass Transfer, 16(2),
pp. 347–358, 1973.
7. Crowe C. T., Multiphase Flow Handbook, CRC Press, 2006.
8. Duda J. L. and Vrentas J. S., “Heat Transfer in a Cylindrical
Cavity,” Journal of Fluid Mechanics, 45, pp. 261–279, 1971.
9. Kandlikar S. et al., Heat Transfer and Fluid Flow in Minichannels
and Microchannels, Elsevier, 2007.
[Link] B. D., Fluid Dynamics and Heat Transfer: An Introduction to
the Fundamentals, Olin College, 2015.
Page 2