MECH4620 Computational Fluid Dynamics - Lecture 1 Notes
Course Information
Course Code: MECH4620
Lecturers:
Prof Guan Heng Yeoh (J17, Room 401B, g.yeoh@unsw.edu.au)
A/Prof Victoria Timchenko (Teams v.timchenko@unsw.edu.au)
CFD Demonstrators:
Ivan (i.decachinhocordeiro@unsw.edu.au)
Rachel (jiayi.li2@unsw.edu.au)
Lingchen (lingcheng.kong@student.unsw.edu.au)
Jingang (jingang.li@student.unsw.edu.au)
Frank (ao.li@unsw.edu.au)
Course Details: Lectures and Labs are located in Moodle.
Weekly Schedule
Week 1 (GHY): Introduction to CFD and examples.
Lab: Backward facing step exercise, problem setup.
Week 2 (GHY): Introduction to ANSYS CFX and Fluent.
Lab: Creating geometry and meshing, defining a CFD problem, creating/importing geometry in Design Modeler.
Due: Release group allocation.
Week 3 (GHY): Mass and momentum conservation, Navier-Stokes equations.
Lab: Creating geometry and meshing, heat exchanger exercise (meshes), discussions of group project topics.
Week 4 (VT): Energy conservation and dynamic similarity.
Lab: T1 work, group project work.
Due: T1: Conservation Laws.
Week 5 (GHY): Turbulence: Basics and introduction.
Lab: Backward facing step exercise; Convergence and Discretisation, Turbulence Models, T2 work, group project work.
Feedback: T1: Conservation Laws.
Week 7 (GHY): Turbulence: applications of models.
Lab: T2 work, group project work.
Due: T2: Turbulence.
Week 8 (VT): Initial and boundary conditions: practical guidelines, post-processing (analysis, validation, and verification).
Lab: Characterization of boundary conditions, Heat exchanger exercise, group project work.
Feedback: T2: Turbulence.
Week 9 (VT): Computational methods – discretisation.
Lab: Computational method online tutorial, T3 work, group project work.
Due: Group project report.
Week 10 (VT): Solution Procedures.
Lab: Revision/consultation, consultation for exams.
Due: T3: Discretisation.
Study Week (GHY, VT): Exam revision.
Feedback: Group project report and T3 discretisation.
Main Texts
J. Tu, G.Yeoh, C. Liu "Computational Fluid dynamics: A Practical Approach”, Butterworth-Heinemann/Elsevier, 3rd Edition, 2018; 4rd Edition, 2023
H.K. Versteeg and W. Malalasekera, An introduction to Computational Fluid Dynamics. The Finite Volume Method, 2nd Edition
Additional Reading
J.D. Anderson, Computational Fluid Dynamics
P.J. Roache, Fundamentals of Computational Fluid Dynamics
P.J. Roache, Verification and Validation in Computational Science and Engineering
J.C. Tannehill, D.A. Anderson and R.H. Pletcher, Computational Fluid Mechanics and Heat Transfer
S.V. Patankar, Numerical Heat Transfer and Fluid Flow
D.C. Wilcox, Turbulence modelling for CFD
Assessment
Tutorial style problems:
Length: 2-3 pages
Weight: 15% (3 x 5% each)
Due: 4 pm Friday, Week 4, Week 7 and Week 10 via Moodle
Deadline absolute fail: Same as assignment deadline.
Marks returned: 1 week after due date.
Group Project:
Length: 15 pages
Weight: 35%
Due: 4 pm Friday, Week 9 via Moodle
Deadline absolute fail: 4 pm Monday, Week 10
Marks returned: 1 week after due date.
Final exam:
Length: 2 hours
Weight: 50%
Due: Exam period, date TBC
Deadline absolute fail: N/A
Marks returned: During exam period.
Group Project Details
Solve a problem using ANSYS, a commercial CFD code.
CFD analysis process: CAD, meshing, pre-processing, solving, and post-processing.
Groups announced in Week 2, project topics announced in Week 3.
Focus on analysis and report writing.
ANSYS help in tutorials, but self-learning is expected via tutorial problems and examples.
Set project problems will be offered.
Use the preliminary template to write your report.
Group Project Report Structure
Short literature review (~ ½ page):
Find 2-3 relevant papers with numerical work descriptions (mesh, turbulence model).
Synthesize info and comment on the difference of your approach.
Project Description (~1 page):
Describe original project proposal and how the final project differs.
Outline what you wanted to achieve and why.
Grid Description and Refinement (~2 pages):
Discuss features of the grids and design rationale, comparing solutions on two or three grid sizes.
Use <2 million cells.
Suggest whether you have reached grid convergence.
Numerical Model (~8 pages, total):
Discuss validation and verification, assumptions, and limitations.
Boundary Conditions: Examine the effect of changing the boundary conditions, both type and specification of values. Are your boundaries placed at the correct location?
Discretization Schemes: Obtain solutions using two different discretization schemes. Explain the scheme. Are the results as you expect? Which scheme do you recommend for this problem?
Convergence: What convergence level do you suggest for your problem? Why?
CFD Models (3 pages): Discuss extra models (turbulence, energy, multiphase). Which model do you recommend and why?
Results (~3 pages): Discuss major flow features, their significance, and justification.
The project is 35% of the total mark.
Thinking about the CFD problem early saves time in the long run.
What is CFD?
Fluid mechanics: study of fluids in motion (dynamic) or at rest (stationary).
CFD focuses on fluids in motion and how fluid flow influences processes like heat transfer and chemical reactions.
Physical characteristics of fluid motion are described by fundamental mathematical equations, usually partial differential equations (governing equations).
Art in Meshing
A good mesh minimizes errors.
The grid should be fine enough not to influence results.
Assess grid convergence via Validation and Verification lecture.
Use at least 2 meshes to compare (3 is better).
Meshing Types
Structured Meshing:
Better results for a given mesh density.
Lower residual errors and faster solver.
Can take a long time to create.
Sometimes not possible.
Unstructured Meshing:
Tetrahedral meshes combined with hexahedral meshes.
Polyhedral meshes becoming more popular.
Cut-cell meshing (ANSYS).
Flexibility of meshing:
Sweep meshing.
Automatic (Tetrahedral & Sweep).
Assembly.
CutCell.
Geometry containing small details:
Patch Conforming: All geometric detail is captured
Patch Independent: Can ignore and defeature geometry
Meshing Summary
Use structured mesh (inflation layer) where a boundary layer will form.
Put more mesh where flow gradients are strongest (use a coarse mesh to check).
Mesh should look “smooth” with no sudden changes in volume.
Different mesh types should blend nicely.
Check mesh for skew and negative volumes.
CFD Analysis Framework
Pre-processor:
Creation of geometry
Mesh generation
Material properties
Boundary conditions
Solver:
Governing equations solve on a mesh
Transport Equations
Mass
Momentum
Energy
Other transport variables
Equation of state
Supporting physical models
Physical Models
Turbulence
Combustion
Radiation
Other processes
Solver Settings
Initialization
Solution control
Monitoring solution
Convergence criteria
Post-processor:
X-Y graphs
Contour
Velocity vectors
Others
Importance of Governing Equations
Represent mathematical statements of conservation laws.
Conservation of Mass: Mass is conserved for the fluid.
Conservation of Momentum: Newton’s second law, the rate of change of momentum equals the sum of forces acting on the fluid.
Conservation of Energy: First law of thermodynamics, the rate of change of energy equals the sum of the rate of heat addition to and the rate of work done on the fluid.
Importance of Boundary Conditions
Physical aspects and mathematical statements of boundary conditions must be developed.
Numerical form depends on mathematical form of governing equations and numerical algorithm.
"Garbage In, Garbage Out" (GIGO) - Late Prof Eddie Leonardi (UNSW)
Importance of Turbulence
Many engineering flows are turbulent.
Turbulence effects must be captured in solving everyday problems.
Laminar flows are described by the governing equations.
More complex flows require numerical solutions with turbulence modeling.
What is Discretisation?
CFD is based on approximate forms of governing equations solved iteratively.
Overview of Solution Procedure
Solution control
Initialization
Monitoring solution
CFD calculation
Check for convergence
If no, modify solution parameters or mesh and return to step 4.
If yes, stop.
Analysis of Results
Qualitative analysis: Visual representation of flow features.
Quantitative analysis: X-Y plots.
Verification and Validation
Verification: Assessing numerical simulation uncertainty and estimating error.
Validation: Assessing simulation model uncertainty using benchmark experimental data and estimating modeling error.
Is the solution correct? How correct do we need?
Application of CFD
Automotive engineering:
Shorten design cycles, optimize components for energy efficiency, improve in-car environment, and study external aerodynamics.
Biomedical science and engineering:
Particle formation/dispersion from nasal sprayers and particle transport/deposition in the nasal cavity
Bronchial tree geometry created from CT data for CFD simulation and CFD prediction of pressure coefficient (C_p) values for specific patient cases with and without asthma
Flow behavior in a carotid bifurcation of artery.
Nanoparticle flow in arteries and arterioles (swirling flow vs. viscosity-dominated Stokes flow).
Abrasive Jet:
Mesh of the computational domain is 75,817 polyhedral elements
Nozzle Diameter 0.36 mm
Inlet Pressure 0.43 MPa
Aggregation of nanoparticles:
Self-organization of nanoparticles (50 nm) in a dense suspension medium
Aggregation of microparticles:
Numerical simulation is performed by examining the aggregation behavior of micro aerosols advected in a circular pipe flow
Magnetic particles in a quiescent fluid:
Nanoparticles uniform particle diameter of 100 nm in a computational domain of 5 μm × 5 μm × 0.25 μm
Magnetic particles in a shear flow:
Nanoparticles of 100 nm with a shear velocity of 0.005 m/s with periodic boundaries, B = 1 Tesla
Bubble motion in a quiescent fluid:
Cases presented are 1mm, 2.5mm, 5mm and 10mm bubbles.
Fire engineering:
Buoyant fire: Temperature, Density, Air Entrainment, Convective Heat Transfer, Radiative Heat Transfer, Combustion, Soot Production & Oxidation
Compartment fire: Layout a full scale compartment fire
Swirling fire: Presentation of Magnitude and Temperature results
Power generation:
Optimize turbine blades for constant power under varying wind conditions.
Model wind farm resource distribution, even for complex terrain.
PV-integrated facade.
Turbulent structures enhance heat transfer and lower surface temperature at the heated wall.
Key Framework Reminder
Physical interpretation of computational results is critical.
Reiterate the CFD analysis framework (Pre-processor, Solver, Post-processor).
This information is found in the "Art in Meshing" section.