Department
of Mechanical and Materials Engineering
Florida International University is a community of
faculty, staff and students dedicated to generating and imparting knowledge
through 1) excellent teaching and research, 2) the rigorous and respectful
exchange of ideas, and 3) community service. All students should respect the
right of others to have an equitable opportunity to learn and honestly
demonstrate the quality of their learning. Therefore, all students are expected
to adhere to a standard of academic conduct, which demonstrates respect for
themselves, their fellow students, and the educational mission of the
University. All students are deemed by the University to understand that if
they are found responsible for academic misconduct, they will be subject to the
Academic Misconduct procedures and sanctions, as outlined in the Student
Handbook.
This is the EGM 6422 Advanced Computational Engineering Analysis Spring 2018
Here is the (12/26/2017)
updated syllabus
for the course.
My office is in EC3442, and email address is levyez@fiu.edu
My tel. no. is 305-348-3643. My fax no. is the department fax no. 305-348-1932
Office Hours: Monday from 10am-12pm
and Wednesday from 300-430pm
TA
information: will be announced. If none,
please see me in my office EC3442
The following reviews will be discussed initially
1. Review of Matrix Equations and their
solutions: example of where used‑‑
Modelling of dynamic systems or
curve fitting. Generation of systems of
equations from the discretization process.
Defining matrix equations.
a.
Review of
matrix notation, matrix norms, conditioning of matrices. What does the ill‑conditioning of
matrices mean from an engineering point of view.
This
material and all the linked materials provided, except where stated
specifically, are copyrighted © Cesar Levy 2018 and is provided to the students
of this course only. Use by any other
individual without written consent of the author is forbidden.
b. Solution
of Matrix equations using Gauss elimination, Gauss‑ Jordan and other
methods (LU) decomposition.
c. Iterative
techniques, convergence of iterative techniques, Jacobi and Gauss-Seidel
iterative methods
d. Iterative
techniques and relation to fixed point iteration; definition of the
pseudoinverse; show the pseudoinverse for Jacobi and G-S method
Here
is the video of the first
lecture that is related to the first part of this review.
Here
is the video of the second
lecture that is related to this review.
Here
is the video of the third
lecture that is related to this review.
Here
is the video of the fourth
lecture that is related to this review and also the beginning of finite
difference solution of differential equations
2. Implicit versus explicit methods for
ODEs
Here
is the video of the fifth
lecture that is related to how to find the difference formulas for
derivatives given the error we wish to achieve, how these formulas in governing
equations give rise to the implicit method of solution, how the solution to the
implicit method relates to the previous lectures, and how higher order
differential equations can be solved by several first order differential
equations that can be solved by the shooting method that incorporate the
methods discussed in previous lectures.
3. Nonlinear Algebraic Sets of
Equations-used to solve nonlinear PDEs
4. Ordinary Differential Equations
Single step methods available
Multistep methods available
Here is the video of the sixth
lecture that is related to single step methods for solution of ODE’s both
linear and nonlinear (Euler, Modified Euler, Runge Kutta, Runge Kutta
Fehlberg) leading to Multistep Methods such as (Adams, Adams- Moulton, Milne’)
and the solution of difference equations in comparison to real solution to the
ODE’s, where applicable.
This
material and all the linked materials provided, except where stated
specifically, are copyrighted © Cesar Levy 2018 and is provided to the students
of this course only. Use by any other
individual without written consent of the author is forbidden.
Here is the video of the sixth
lecture that is related to single step methods for solution of ODE’s both
linear and nonlinear (Euler, Modified Euler, Runge Kutta, Runge Kutta
Fehlberg) leading to Multistep Methods such as (Adams, Adams- Moulton, Milne’)
and the solution of difference equations in comparison to real solution to the
ODE’s, where applicable.
Here is the video of the seventh
lecture that is related to multistep methods for solution of ODE’s both
linear and nonlinear (such as Adams Method, Adams-Moulton, Milne’s methods) and
a discussion of proving how the solutions of the difference equations generated
by these numerical methods converge to the solutions of the differential
equations. Also a discussion is given as
to why Milne’s method converges some of the times and how the choice of the
difference equation representation of the derivative can lead to conditional or
unconditional stability of the numerical method.
Topics to be covered in the next lecture
5. Introduction to PDE's
What PDE's characterize
Classification of PDE's
Finite Difference Notation
6. Parabolic Differential Equations
Explicit and Implicit Methods
Here is the video of the eighth
lecture that discusses the instability of Milne’s method, discusses the
characterization of PDEs and begins the discussion of numerical solution of
PDEs. We begin by discussing the
parabolic PDEs, such as the heat equation using explicit numerical methods.
This
material and all the linked materials provided, except where stated
specifically, are copyrighted © Cesar Levy 2018 and is provided to the students
of this course only. Use by any other
individual without written consent of the author is forbidden.
Topics to be covered in the next few lectures
6. Parabolic Differential Equations
Explicit and Implicit Methods
Initial and Boundary Conditions, Limiting
Conditions
Convergence, Stability and Consistency
Well Posedness and
Sufficiency
Lax-Wendroff
Conservation Equations
Nonlinear Parabolic Equations and Schemes to
Solve Them
Here is the video of the ninth
lecture that discusses the numerical solution of the parabolic PDE
characterized by the heat equation. We
show the Neumann stability analysis for the explicit Forward Difference in time
O(Dt, Dx2) scheme; the implicit Backward
Difference in time O(Dt, Dx2) scheme; and we started discussing
the Centered Difference in time O(Dt2,
Dx2) scheme. Here is a document related to the Neumann
Stability analysis
Here is the video of the tenth
lecture that discusses the numerical solution of the parabolic PDE
characterized by the heat equation. We
show the Neumann stability analysis for the Centered Difference in time O(Dt2, Dx2)
scheme and the problems related to that scheme; how the Crank-Nicholson Scheme
O(Dt2, Dx2)
helped fix those problems. We also
discuss the Lax Equivalence Theorem in the topic of Convergence, Stability and
Consistency.
The eleventh lecture discusses multi-level time schemes and solution of 2-D heat equation using the ADI (alternating direction implicit) scheme, schemes for polar coordinate systems, and schemes for nonlinear parabolic equations
Also a take-home midterm was provided and due in class next Wednesday. Please do your own work.
The twelfth lecture discusses more information on cylindrical and polar problems for solutions related to the origin. Extrapolation to the limit is also discussed for the explicit method and for the Crank-Nicholson Schemes. Finally, a discussion of the Newton Linearization scheme for nonlinear parabolic equations is given.
The twelfth lecture covers a numerical problem for the Newton Linearization scheme, and looks at the Richtmeyer Linearization Scheme. A handout was given regarding elliptic PDEs which will be our next topic and their numerical solution methods.
This
material and all the linked materials provided, except where stated
specifically, are copyrighted © Cesar Levy 2018 and is provided to the students
of this course only. Use by any other
individual without written consent of the author is forbidden.
7. Elliptic PDE's
Laplace Equation and Iterative Schemes to
Solve It
Poisson Equation
Dealing with Limiting Conditions of Boundary
Conditions
NonCartesian Meshes
and NonRegular Regions
The thirteenth lecture covers Elliptic PDEs. Some basic theoretical discussions are given such as the principle of the maximum/minimum and numerical solution of the Laplace Equation Ń2 u = 0 and the Poisson’s Equation Ń2 u = f such as the Liebmann, Southwell and Gauss-Seidel methods are discussed. Proof of convergence of the iterative techniques is discussed to include choice of initial approximations to the solution is covered. Lastly, solution of the problem using the implicit technique is also discussed.
The fourteenth and fifteenth lectures continue to discuss elliptic PDEs
The sixteenth lecture covers the Laplace Equation Ń2 u = 0 and the Poisson’s Equation Ń2 u = f solved using the implicit method. We also discuss the general topic of handling non-standard boundaries, where distances from the standard grid for the interior region to the boundary leaves a portion of a stepsize. We also discuss grids in equilateral triangles and solution of the equations in such regions. We lastly show how the ADI method discussed for the 2-D heat equation can be used to solve the Laplace equation if the value of C = is taken to be very large. Next time we will discuss the hyperbolic PDEs and their solutions.
The seventeenth and eighteenth lectures continue to deal with parabolic type PDEs
8. Hyperbolic PDE's
Implicit and Explicit Schemes
Problems of Stability in the Schemes
Courant-Friedrichs-Levy
Condition
Numerical Integration along
Characteristics-maintaining discontinuities
Second Order Equations
The nineteenth lecture begins discussion of solution of hyperbolic PDEs such as the wave equation. The general finite difference equation is derived and its stability is discussed in which the Courant Number C = is an important factor. The domain of dependence and range of influence is defined and the numerical characteristics and the analytical characteristics are defined and related to each other and related to the stability of the numerical method as well as the accuracy of the numerical method. The solution for time level 1 is discussed using one of the initial conditions and a more accurate result is given using the D’Alembert Solution.
This material and all the linked materials provided, except where stated specifically, are copyrighted © Cesar Levy 2018 and is provided to the students of this course only. Use by any other individual without written consent of the author is forbidden.
The twentieth and twenty-first lectures continue to discuss hyperbolic PDEs and their solutions.
The twenty-second lecture begins discussion of solution of hyperbolic PDEs using the method of characteristics. The equations to be solved are derived and two examples are discussed in detail. One example covers a situation in which initial conditions are discontinuous but the characteristics are constant. One example is where the characteristics are not constant but are only functions of the independent variables only, but the initial conditions are continuous.
The twenty-third lecture discusses the last example of the method of characteristics for a non-linear PDE in which the characteristics are a function of the dependent variable. We also discussed the one-dimensional wave equation and its solution.
The twenty-fourth lecture continues the discussion of the method of characteristics.
9. Optimization
Linear Programming
The twenty-fifth lecture and twenty-sixth lecture discuss the rudiments of linear programming, namely defining the feasible region, the constraint equations, solution of problems of two variables, solution of problems of multiple variables using the simplex method.
The twenty-seventh lecture will wrap up the course and cover the final exam.
This material and all the linked materials provided, except where stated specifically, are copyrighted © Cesar Levy 2018 and is provided to the students of this course only. Use by any other individual without written consent of the author is forbidden.