JOHN GATSIS
gatsis AT utias DOT utoronto DOT ca
University of Toronto Institute for Aerospace Studies
4925 Dufferin Street
Downsview, ON
M3H5T6
Profile
I am a Ph.D. candidate at the University of Toronto Institute
for Aerospace Studies. Thank you for visiting my website.
Curriculum Vitae
Available upon request.
Important Links & People
Supervisor: Professor David Zingg
Doctoral Examination Committee Chair: Professor Clinton Groth
Doctoral Examination Committee Member: Professor Hugh Liu
University of Toronto ROSI Student Service
Colleagues: UTIAS CFD Lab
Ph.D. Candidate: Mohammad Tabesh
UTIAS Webmail
Facebook: Me!
Toronto Math Tutor.
Toronto Tutor John Gatsis - Math, Physics and Chemistry tutoring services within the GTA.
Papers / Publications
A Fully-Coupled Newton-Krylov Algorithm for Aerodynamic Design
Optimization
Co-authored by Professor D.W. Zingg
16th AIAA Computational Fluid Dynamics Conference from June 23-26,
2003
Orlando, USA
Conference Paper:Download PDF file
Conference Presentation:Download PDF file
A Fully-Coupled Algorithm For Aerodynamic Design Optimization
Co-authored by Professor D.W. Zingg
48th Annual CASI Conference on May 1, 2001
Toronto, Canada
Conference Paper:Download PDF file
Conference Presentation:Download PDF file
Results
This is a Mach number contour of a supercritical flow about a RAE2822
airfoil. This is a fully-turbulent flowfield, at a Mach number of
0.75 and an angle of attack of 3 degrees. Notice the stagnation
region at the leading edge, the supersonic region, the lambda shock, and the
shock-boundary-layer interaction including shock-induced flow
separation. Finally, note the recirculation region on the upper
surface of the airfoil at the trailing edge.
This simulation was conducted using a Newton-GMRES algorithm,
which solved the discretized, compressible, steady Navier-Stokes
equations with the Spalart-Allmaras turbulence model. On my
desktop computer, this simulation took roughly 2 minutes.
This is a 3D contour plot of CPU time versus angle of attack and
freestream Mach number for flow about a symmetric NACA0012
airfoil. The Euler equations are used to model the flow. A fully Newton-GMRES algorithm is used to solve the discretized equations. Local
preconditioning can be used to improve the algorithm's performance for
low-Mach number flows and for flows with a Mach number in the
neighbourhood of unity. Simulations for large angles of attack are
not feasible since viscous effects are not described by
the Euler equations.
AER307F Aerodynamics Course
Division of Engineering Science, Faculty of
Applied Science and Engineering, Univeristy of Toronto
Fall 2003: Course instructor: John Gatsis
Fall 2004: Course instructor: John Gatsis
Fall 2005: Course instructor: Professor David Zingg
Fall 2006: Course instructor: John Gatsis
Fall 2007: Course instructor: John Gatsis
Current course instructor: Philippe Lavoie
The course website is on the UofT PORTAL. To access this site you must first login to your UTORID account. This course should appear under your courses.
Ph.D. Thesis
Preconditioning Techniques for a
Newton-Krylov Algorithm for the Compressible Navier-Stokes Equations
Funded by the Ontario Graduate Scholarship Program (OGS & OGSST)
Supervisor: Professor D. W. Zingg
One of the key goals in computational fluid dynamics (CFD) is to
create efficient flow solvers. The solution of analysis problems in
industry is important in predicting the performance and in the design
optimization of aircraft. In fact, the optimization process requires
a large amount of flow solves to complete its task accurately. The
compressible Navier-Stokes equations that describe the fluid flow are
extremely difficult to solve numerically in an efficient manner.
Furthermore, aerodynamic flows are typically turbulent and have high
Reynolds numbers. This means that the flow solver must operate on a
very fine computational grid to produce accurate results. However,
with the larger system size, comes larger computational cost. A
promising approach to solving these equations is to use Newton-Krylov
methods. In Professor Zingg's research group at UTIAS, promising
results using the GMRES (generalized minimum residual) Krylov subspace
method have been found. However, the challenge with GMRES is choosing
the best possible preconditioner for the system. Various techniques
in preconditioning will be explored in this research. Some areas of
exploration include algebraic multigrid, sparse approximate inverses,
orderings, and Schwarz domain decomposition.
Course Research
Matrix Calculations / Numerical Linear Algebra
Professor Christara, University of Toronto, Department of Computer Science
Course Home Page
Numerical Solutions for Optimization Problems
Professor Wolkowicz, University of Waterloo, Department of Combinatorics and Optimization
Course Home Page
Air-Breathing Propulsion
Professor Sislian, University of Toronto, Institute for Aerospace Studies
The Finite Element Method
Professor Hansen, University of Toronto, Institute for Aerospace Studies
M.A.Sc. Thesis
A Fully-Coupled Algorithm for Aerodynamic Design Optimization
Funded by the Natural Science and Engineering Research Council of
Canada (NSERC)
Supervisor: Professor D. W. Zingg
A fully-coupled algorithm is presented to solve aerodynamic design
optimization problems. The discretized flow, adjoint, and optimality
condition equations are solved as a single system of nonlinear
equations using an inexact Newton method with linesearching.
Quasi-1D, inverse, nozzle design and inviscid, 2D, inverse, airfoil
shape design cases are solved, however the method is applicable to
more general design cases.
While the system is approximately twice as large as the system
describing the flow solution, fewer iterations are required to find
the optimum. The most important advantage of this method is that
only a single flow solve is needed to perform the optimization, as
opposed to the discrete adjoint method which requires several. Many
improvements have been made to the algorithm, both in the system
formation and solution method. The result is an algorithm that is
reliable, robust, accurate, and fast in providing optimum solutions
for the cases considered.
Course Research
Graduate Course: Unsteady Gasdynamics
Professor Gottlieb, University of Toronto, Institute for Aerospace Studies
Graduate Course: Fundamentals of Computational Fluid Dynamics
Professor Zingg, University of Toronto, Institute for Aerospace Studies
Graduate Course: Topics in Computational Fluid Dynamics
Professor Zingg, University of Toronto, Institute for Aerospace Studies
Graduate Course: Viscous Flows and Boundary Layers
Professor Walsh, University of Toronto, Institute for Aerospace Studies
Results
ARC1D Flow Solver
ARC1D Flow Solver Results:Download PDF file
1D Halving Linesearch Results:Download PDF file
1D Backtracking Linesearch Results:Download PDF file
A typical flow solution in 2D:Download PDF file
Final Report
Download PDF file
Download PS file
Visit the Canadian National Archives to get a copy at
http://amicus.nlc-bnc.ca
B.A.Sc. Thesis
Feedback Control Systems
for Automotive Aerodynamic Devices
Supervisor: Professor P. A. Sullivan
The objective of this thesis is to implement a feedback control system for automotive aerodynamic devices such as spoilers and stabilizers in order to achieve a higher degree of driver comfort, road handling, and traction.
I'd like to especially thank Dr. Phil Sullivan for his exceptional guidance. He has an important place in history for many reasons, including for his significant contribution to the safe return of the Apollo 13 astronauts.
Course Work
Year 4
Atmospheric Flight
Gasdynamics
Spacecraft Dynamics and Control I
Thesis
Complex Analysis
Aircraft Design
Advanced Mechanics of Structures
Aerospace Propulsion
Confronting Global Change
Robotics
Year 3
Dynamics
Fluid Mechanics
Mechanics of Solids and Structures
Partial Differential Equations
Control Systems
Aerodynamics
Scientific Computing
Economic Analysis and Decision Making
Materials Design and Engineering
Probability and Statistics
Year 2
Engineering Design
Chemical Processes
History of North American Technology
Calculus III - Vector Calculus
Physics IV - Quantum Mechanics
Differential Equations
Electric Circuits
Digital and Computer Systems
Varieties of Fiction
Physics II/III - Vibrations and Waves, and Thermal Physics
Year 1
Engineering, Society, and the Environment
Algorithms and Data Structures
Electricity and Magnetism
Linear Algebra
Calculus II
Chemistry
Structures and Materials
Computer Programming
Calculus I
Physics I - Mechanics
Course-Related Work and Projects
Engineering Design -- M.J. Malone
Mars Rescue Rover (MRR):Website
Fundamentals of Computational Fluid Dynamics -- D.W. Zingg
Convergence Acceleration by use of Multigrid in MATLAB:Download PDF file
Unsteady Gasdynamics -- J.J. Gottlieb
"New" Riemann Solver in C++:Header File and Code
Topics in Computational Fluid Dynamics -- D.W. Zingg
Hirsch's Nozzle and Shock Tube Exact Solutions:Download PDF file
ARC1D Flow Solver Results:Download PDF file
FLOMG Flow Solver Results:Download PDF file
Applied Upwinding Dissipation Schemes Results:Download PDF file
Applied Upwinding Dissipation Schemes Presentation:Download PDF file
Viscous Flows and Boundary Layers -- C.B. Walsh
Investigating the Spalart-Allmaras Turbulence Model Report:Download PDF file
Investigating the Spalart-Allmaras Turbulence Model Presentation:Download PDF file
Matrix Calculations / Numerical Linear Algebra -- C.C. Christara
Final Course Project:Download PDF file
Tutorial Service
***FREE*** Advanced Functions, Introductory Calculus, and Advanced
Placement (AP) Calculus test preparation questions as well as Linear Algebra, and Science book
Toronto Tutoring Service
Toronto Math Tutor John Gatsis - Math, Physics and Chemistry tutoring services within the GTA.
Artwork
Click here
Miscellaneous and Interesting Items
Check out the V.P. of Pendulum Electromagnetics
Google Maps
Scientific Research at Fourmilab
Current Weather Conditions for Canada
Web-Related Items
Hotmail
Web Your Mind
Elite Music Academy
Photography by Katerina
CNN
Gmail
Google
Hotmail
yellow-net.com
Yahoo
Live
Wikipedia
Programming Items
C++ Homepage
C++
Reference
You are visitor #: 