Assignment 3

This homework assignment involves conducting a CFD simulation of turbulent flow over a backward facing step to analyze reattachment length and pressure drop. The document provides instructions on setting up the simulation in FDS including defining the geometry, mesh resolution, and boundary conditions based on a given Reynolds number. It also specifies requirements for running the base case model and sensitivity studies while varying grid resolution, turbulence model, and domain width. Results must be documented in a report analyzing reattachment length, pressure drop, and their sensitivity based on the model parameters.

Uploaded by

Suru Unii

Available Formats

Download as PDF, TXT or read online on Scribd

0% found this document useful (0 votes)

132 views

Assignment 3

Uploaded by

Suru Unii

Available Formats

Download as PDF, TXT or read online on Scribd

You are on page 1/ 4

Assignment 3 Flow Over a Backward Facing Step

Introduction

The backward facing step is a benchmark problem for validation of CFD modeling techniques. Flow over
the step undergoes separation, reattachment and recirculation. Turbulence models have been seen to
give widely varying results for this problem, and the results are also sensitive to the numerical method
and grid resolution.

Source: Computational Fluid Dynamics – A Practical Approach

In this homework assignment, you will conduct a CFD simulation of turbulent flow over a backward facing
step using the Fire Dynamics Simulator (FDS) software (Version 6.6 or 6.7 – current in-use versions). The
aims of this assignment are to:

1. Generate a 3D geometry and mesh in FDS (0.1 m to 0.2 m nominal cell size).
2. Set up monitors in the FDS model to record results and simulation diagnostics.
3. Simulate turbulent flow over the step and monitor results to demonstrate that a steady state
solution has been reached.
4. Post process the results (via Smokeview and Excel) and compare with experiment data.
5. Check sensitivity of results to the geometry (domain width), grid resolution and turbulence model.
6. Document results in a short report.

1
Assignment 3 Flow Over a Backward Facing Step

Data and Model Set Up Instructions

Geometry for the step is provided in the figure below. The Reynolds number for the flow is defined as:

Re = ρ* U * H / μ

ρ = density (kg/m3), nominally 1.2 kg/m 3

U = upstream bulk velocity (m/s)

H = step height (m)

μ = fluid viscosity (N/s/m2), nominally 1.8e-5 Ns/m2

The Reynolds number for this problem is 15,000. Expansion ratio (ER) for the geometry should be 2.0.
Width of the geometry should be at least 2h 2.

Download the following file from Courseworks and use it as a basis for building your model: CFD-03-10.fds

The CFD model inlet length prior to the step (Le), and outlet length downstream of the step, are up to the
individual to determine (you should state why you used a certain inlet and outlet length). The length
should be chosen such that boundary conditions are far enough away from the expansion to not influence
the result of interest. You should choose a grid resolution that balances computation time versus resolving
the flow adequately (hint: if your model takes more than about 60 minutes to run, the grid resolution is
much too fine, and a mesh of around 100,000 cells should run reasonably quickly, but also note that you
want to have at least five cells across the domain width and height).

Your simulation will be transient; you will need to choose a run time long enough to allow the flow to
develop and settle to a pseudo-steady state. FDS uses large eddy simulation turbulence modeling, which
resolves large scale turbulent motion, hence it is three-dimensional and transient (you should monitor the
velocity at a point the domain and total volume flow to confirm this steady condition is achieved).

2
Assignment 3 Flow Over a Backward Facing Step

Experiment data are for this problem are taken from published data1. For this assignment, the key piece
of experimental information to use is the reattachment length. In FDS this can be computed through
visualization of the time-averaged component of x velocity (the location where the velocity is 0 is the
reattachment length); visualization is accomplished with a contour slice of x velocity along the duct
centerline and in Smokeview the slice files can be averaged in time. Tick marks can be placed to help
visualize the exact length. Refer to file “Assignment 3 – FDS Basics.pptx” for a demo of how to do these
two things. Reattachment length for this problem, based on test data is around 8.5H downstream of the
step.

Submission Requirements and Instructions

1. Introduction and problem statement

a. Geometry (base case parameters) – physical dimensions including step height, inlet duct
height, domain length chosen downstream of the step, and domain width. Provide
reasoning for the domain length chosen downstream of the step.
b. Mesh resolution (show an image of the mesh / geometry, quote number of cells, note
whether any special approaches were needed to generate the mesh – single grid or
multiple grids).
c. Velocity boundary conditions for the CFD model (show working for computation of inlet
velocity).
d. Submit your base-case model (as a text file, for information only in case of a result not
making sense).
2. Run the model with FDS default parameters used (same as case CFD-03-10):
a. Submit a graph of velocity versus time at a point in the domain and total volume flow rate
to demonstrate that steady state is reached. Comment on what indicates steady state
and beyond what time.
b. Submit a mean (time-averaged) x velocity contour, label the recirculation zone and state
why a certain averaging time was used (such as visual observations, integral flow time,
etc.). Mark the recirculation point and approximate length (report the length in terms of
step heights).
c. Report the pressure loss through the geometry, compute K factor based on velocity
before the step (ignore friction loss at the wall – make it part of the K factor you compute,
show working).
3. Sensitivity analysis (summarize how each model change affects reattachment length, in step
height, and pressure loss in a tabular form):
a. Adjust grid resolution (coarser grid) – keep everything the same as above but make the
grid coarser by a factor of 2 (not that each duct must have at least 5 cells across the height
or width), report the reattachment length and pressure loss, and compare with the result
from 3c.
b. Adjust turbulence model – keep everything the same as 3a, but make the turbulence
model “DYNAMIC SMAGORINSKY”, report the reattachment length and pressure loss, and
compare with the result from 3a.

1
Flow separation over the inclined step, Ruck and Makiola, Physics of Separated Flows—Numerical, Experimental,
and Theoretical Aspects, Vieweg+Teubner Verlag (1993): 47-55.

3
Assignment 3 Flow Over a Backward Facing Step

c. Adjust domain width – keep everything as per 3b and make the domain wider by a factor
of 2, report the reattachment length and pressure loss, and compare with the result from
3b.
d. Compare the results with respect to reattachment length sensitivity.
4. Document results in a short report/summary that responds to the questions 1 through 3 given
above.

Assessment

Total value: 10% of final grade (10 points – see below for breakdown)

Assessment:

1. Introduction and problem statement 3.0

2. Base case runs 3.0
3. Sensitivity analysis 3.5
4. Presentation/report 0.5

___________________________________

Total 10

Work is to be conducted individually. Be descriptive as possible in your answers. Highlight results.

Organize text clearly and neatly. Clearly state all assumptions. Label and properly refer to figures, tables
and equations presented (0.5 marks (out of 10) of the assignment grade is assigned for presentation).

Due date: October 21, 2019 at 6 pm (0.5 mark deduction for each day of delayed submission)