Fundamentals of Numerical Thermo-Fluid Dynamics
| Credits | 4 SWS, 4 ECTS |
| Contact | Surendran, Aswathy |
| Sessions | Tuesdays , 9 a.m. until approx.4 p.m. |
| Room | MW 1701, afterwards computer lab MW 0704 |
| Term | Sommersemester 2022 |
|---|---|
| Language of instruction | English |
Admission information
See TUMonline
Note: If you want to get on the waiting list, please eMail to Aswathy Surendran surendran@tfd.mw.tum.de and show up for the first meeting. Experience shows that lab courses are heavily overbooked, so chances are good that you can participate if you are on the waiting list.
Note: If you want to get on the waiting list, please eMail to Aswathy Surendran surendran@tfd.mw.tum.de and show up for the first meeting. Experience shows that lab courses are heavily overbooked, so chances are good that you can participate if you are on the waiting list.
Objectives
– After successful completion of the course, the participants will be able to implement basic numerical algorithms in the MATLAB programming language, or to use functions and tools already implemented in MATLAB. Furthermore, they will understand the basic concepts of deep learning and will be able to create and train neural networks using the Deep Learning Toolbox provided by MATLAB.
– The students will be able to understand the solution approaches to selected problems dealing with thermo-fluid dynamics and apply what they have learned to analogous problems from all areas of engineering.
– The students will be able to understand the solution approaches to selected problems dealing with thermo-fluid dynamics and apply what they have learned to analogous problems from all areas of engineering.
Description
– The practical course ”Numerical Thermo-Fluids - From Differential Equations to Deep Learning” teaches knowledge of the most important numerical algorithms as well as the principles of good programming.
– Besides classical numerical approaches, the participants will get to know the basic principles of machine learning and deep learning.
– Advantages and disadvantages of different methods will be discussed, enabling a differentiated interpretation and critical evaluation of numerical results. The algorithms will be implemented by the participants in the programming environment MATLAB, dealing with exemplary problems in the field of thermo-fluid dynamics and heat transfer.
– Besides classical numerical approaches, the participants will get to know the basic principles of machine learning and deep learning.
– Advantages and disadvantages of different methods will be discussed, enabling a differentiated interpretation and critical evaluation of numerical results. The algorithms will be implemented by the participants in the programming environment MATLAB, dealing with exemplary problems in the field of thermo-fluid dynamics and heat transfer.
Prerequisites
Successful completion of the lectures Advanced Mathematics 1,2,3, Thermodynamics I, Heat Transport Phenomena and Mathematical Tools.
Teaching and learning methods
– Lectures, supervised completion of numerical problems in groups of two.
– Each course chapter will be presented in an introductory lecture, which will cover the theoretical background and give tips for practical implementation. Afterwards students will work in groups of two on exemplary problems under the supervision of a TFD associate.
– After completing the last chapter, students will define a final project (in groups of two). Each group will work independently on their respective problem for approximately one month. A TFD associate will assist with any upcoming problems and questions. Results of
the final projects will be presented to all participants of the practical course (written report and presentation).
– Each course chapter will be presented in an introductory lecture, which will cover the theoretical background and give tips for practical implementation. Afterwards students will work in groups of two on exemplary problems under the supervision of a TFD associate.
– After completing the last chapter, students will define a final project (in groups of two). Each group will work independently on their respective problem for approximately one month. A TFD associate will assist with any upcoming problems and questions. Results of
the final projects will be presented to all participants of the practical course (written report and presentation).
Examination
The MATLAB programs created during the eight sessions account for 2/3s of the grade. Evaluation criteria are functionality, structure, universality, utilization of MATLAB syntax, commenting, and quality of program output, i.e. numerical results.
Recommended literature
Wolfgang Polifke und Jan Kopitz. Wärmeübertragung: Grundlagen, analytische und numerische Methoden. Pearson, 2009.
Brian P Flannery, William H Press, Saul A Teukolsky und William Vetterling. Numerical recipes in C, 1992, URL: http://www.nr.com
"Deep Learning" by Ian Goodfellow, Yoshua Bengio and Aaron Courville, MIT Press, 2016, URL: https://www.deeplearningbook.org
Brian P Flannery, William H Press, Saul A Teukolsky und William Vetterling. Numerical recipes in C, 1992, URL: http://www.nr.com
"Deep Learning" by Ian Goodfellow, Yoshua Bengio and Aaron Courville, MIT Press, 2016, URL: https://www.deeplearningbook.org