Turbulent Flow Simulation on HPC-Systems

An awarded course:

The lab course has been awarded with the Ernst-Otto-Fischer-Lehrpreis 2012/2013 by the faculty of Mechanical Engineering. The award has been given, because the lab course trains interdisciplinary team competency which is needed to solve nowadays complex technical problems. Therefore, an interdisciplinary team of developers is simulated by the students of different faculties that need to combine their knowledge in order to solve a specific problem – the efficient numerical simulation of a technical configuration which is hard to investigate experimentally.

Description:

Simulation of Turbulent Flows on HPC-System is a multidisciplinary course for students of the master program of the faculties Mechanical Engineering, Computer Science and Computational Science and Engineering. The participants ought to show a great interest in working in independent, interdisciplinary teams on developing a working CFD code. Thereby, the fundamentals taught in previous classes within the own faculty are elaborated and basic concepts of the other faculty (Mechanical Engineering ↔ Computer Science) are learned by means of group projects. Interdisciplinary communication skills are thus trained. The lab course is held in the winter term, because prerequisite courses are held in the summer semester.

The course language is English.

Course outline:

In teams of two to four students, the participants develop a three-dimensional RANS flow code from a C++ flow solver with basic functionalities. A suitable turbulence model is implemented and the RANS code is parallelized for distributed systems (i.e. LRZ). Generic test cases are simulated and compared to reference data to validate the new features. For the final project, both a numerical simulation with the own code and experimental measurements of the equivalent configuration will be conducted and compared. The considered test case is the flow over a backward facing step, commonly used for turbulence model performance evaluation.

Prerequisites:

You should be familiar with the operating system Linux. Fundamental concepts of fluid mechanics and computational fluid mechanics need to be understood. The courses 'Turbulent flows', 'Numerical Methods for Engineers' or ' Fundamentals of Computational Fluid Mechanics' are recommended.

Student evaluation:

The students are evaluated on basis of oral exams, student presentations and evaluation of code development. Details on the contributions are given in the first lecture.

Where and When:

The lab course is held in the Wintersemester. Further information about dates and the course schedule are available on TUMonline.

Class dates: >>TUMonline

Registration:

For participating, register under TUMonline.

For registering to the exam, also enlist yourself here .

Course materials:

Course materials ( lecture notes, student assignments, reference data, etc.) will be available to registered useres on >>TUM - Moodle

Responsible for the course:

Mechanical Engineering:

Jakob Kaiser, M. Sc.

Raum: MW 1616

E-Mail: jakob.kaiser@tum.de

Nico Fleischmann, M. Sc.

Raum: MW 1677

E-Mail: nico.fleischmann@aer.mw.tum.de

Informatik:

Ionut-Gabriel Farcas M.Sc.

Raum: MI 02.05.053

E-Mail: farcasi@in.tum.de