Ege University Mechanical Engineering Department

Computational Fluid Dynamics Course Tutorials

Tutorial #1 :CFD Simulation of the 3D Viscometer Problem

1 Introduction
Understanding the flow of Newtonian fluids is fundamental in fluid mechanics and
engineering. In this tutorial, we delve into the flow simulations to explore the behavior
of a Newtonian fluid in a simple viscometer. By doing so, we will gain insights into
the principles that underlie fluid flow and viscosity measurement.
Viscometry is the science of measuring fluid viscosity, a crucial property for var-
ious industrial processes, from manufacturing to quality control. Accurate viscosity
measurement ensures the efficiency and quality of products in fields as diverse as food
production and oil refining. We will focus on Newtonian fluids, which exhibit a linear
relationship between shear stress and shear rate. Understanding Newtonian flow is
essential for accurately determining a fluid’s viscosity using a viscometer.
The viscometer problem in this tutorial can be modeled as two cylinders positioned
one inside the other. The gap between the two cylinders is filled with the fluid of
interest, creating a confined space for the fluid to flow through and be measured. One
of the cylinders is driven by a constant speed motor. Torque due to fluid shear is
measured and viscosity is calculated.

2 Problem Definition
Consider the simple viscometer setup as shown in Figure 1. The cavity formed by
the concentric cylinders is filled with liquid. For small clearances, we may assume
a linear velocity distribution in the liquid. Diameter of the inner cylinder is 75 mm
with a clearance gap of 1.5 mm. A torque of 0.00037 Nm is required to turn the inner
cylinder at 100 rpm. Outer cylinder is fixed. Height of the device is h = 0.15 m.
Determine the viscosity of the liquid. In simulations, take ρ = 1000 kg/m3 .

1
 Figure 1: Problem figure.

3 Solution By Hand
Inner cylinder rotates with a constant speed. The angular acceleration is zero. This
implies that the net torque on the inner cylinder is zero. So the torque imposed by
the electrical motor is balanced by the torque created by the fluid shear. The shear
force is,
F = τA (1)
where A = 2πRi h. For a Newtonian fluid, shear stress varies linearly with the velocity
gradient:
du
τ =µ (2)
dr
where r represents the radial direction. Since we assume a linear velocity profile,
slope is constant everywhere:
V
τ =µ (3)
d

2
Here d = 1.5 mm and V = ωRi . Since the hydrodynamic torque is T = Ri × F , we
write,
V ωRi
T = Ri µ A = Ri µ 2πRi h (4)
d d
or,
2πµRi3 ωh
T = (5)
d
Arrange for viscosity and plug in the values:
Td
µ= (6)
2πRi3 ωh
0.00037 × 0.0015
µ= (7)
2 × π × 0.03753 × 10.472 × 0.15
which yields µ = 0.001 Pa.s.

4 CFD Simulation
It is possible to simulate the flow in 2D, just as in the hand solution of the problem.
However the grid generator that we use, Fluent Meshing software, generates 3D grids
only. For this reason, we will consider the complete flow domain in 3D as a hollow
cylindrical shape. Our objective in simulations is to find the torque for a liquid with
a viscosity of µ = 0.001 Pa.s.
Follow the steps below to perform the analysis assuming steady, 3D, laminar flow.
Please note that detailed instructions for geometry creation and mesh generation will
be provided during the in-class lecture. This tutorial will focus on the CFD simulation
and analysis aspects.

1. Start ANSYS Fluent with the 3D, double-precision options.

2. Use the steady, pressure-based solver with the absolute velocity formulation.

3. Use the laminar model.

4. Change default air to liquid (ρ=1000 kg/m3 , µ=0.001 Pa.s).

5. Change unit for the the angular velocity to rev/min in the units section.

6. On the wall inner boundary condition, enter absolute, rotational wall motion
with a value of 100 rev/min. to impose wall motion.

7. Retain the default settings under methods.

3
 Figure 2: Mesh of viscometer.

8. Disable the residual based termination of solution procedure. We prefer to solve

for a prescribed number of iterations.

9. Add a report definition to monitor the moment on the wall inner.

10. Initialize the flow field using the hybrid initialization method.

11. Set the number of iterations to 200 and run calculation. Make sure the solver
has converged.

5 Results Comparison: What Did We Achieve?

Based on the flow assumptions and the grid resolution used, we obtain the moment
on wall inner of 0.00038 Nm. The relative error is only 2.70 percent, however we
must be careful since a grid sensitivity analysis has not been performed yet.
Figure 3 shows the expected linear variation of tangential velocity along the radial
direction.

6 Exercises
Answer the following:

• List the assumptions in your hand calculations versus in simulations. Should

we expect the same result from both? Why?

4
 (a) Vector plot
0.40

0.35

0.30

0.25
Vθ (m/s)

0.20

0.15

0.10

0.05

0.00
0.0376 0.0378 0.0380 0.0382 0.0384 0.0386 0.0388 0.0390
r (m)

(b) XY plot

Figure 3: Tangential velocity variation.

5
• Find the torque applied by the fluid on the inner cylinder and check with the
problem input.

• Double the cylinder spacing d. How does the torque change?

• Keep increasing d. When does the linearity of velocity profile break?

• Repeat simulations for 10, 25, 50, 75 rpm. Plot shear stress versus velocity
gradient. What is the slope?

• Update density as ρ = 100 kg/m3 . How does the torque change?