Go back to overview

Modeling Seminar 2022

After a long wait­ing period caused by the Coro­na pan­dem­ic, the con­clu­sion of the mod­eling semi­nar could take place out­side the uni­versi­ty for the first time. In the heart of Fran­coni­an Swit­zer­land, near Gößwein­stein, the par­tici­pants of the semi­nar pre­sented their re­sults to their su­pervi­sors and com­pany repre­senta­tives.

Spring comparison 4.0?

This pro­ject was a con­tinua­tion from the 3rd mod­eling semi­nar of the Elite Grad­uate Pro­gram “Sci­en­tific Com­puting”. A pro­duced spring was to be auto­mati­cally com­pared with the previ­ously pro­duced CAD de­sign. A 3D scan of the pro­duced com­po­nent served as the data basis.
The work in this pro­ject con­sisted of re-ex­amining the algo­rithms de­veloped in the previ­ous year and adapt­ing them better to the ge­ometry of a spring. Ul­ti­mate­ly, this re­sulted in a high­er-qual­ity over­all result. By incor­porat­ing more effi­cient data struc­tures, the com­pari­son was also accel­er­ated.

Reinforcement Learning in Power Electronics

The focus is on the opti­miza­tion of oper­ating and com­po­nent pa­rame­ters in elec­tronic cir­cuits. Since the be­havior of these cir­cuits is very com­pli­cated, math­emat­ical and pro­grammatic meth­ods are used for opti­miza­tion. For these pur­poses, the cir­cuits must first be math­emat­ically mod­eled and then simu­lated.
The pro­ject in­volved op­timiz­ing the com­po­nent pa­rame­ters using a re­in­forcement learn­ing algo­rithm. The opti­miza­tion should ad­di­tion­ally be im­proved by an exten­sion of the used algo­rithm for dif­ferent input and output varia­bles. 

Dynamics of elastic ropes

The back­ground to this sub­ject is the fas­tening of a ship to the quay wall or the tow­ing of a ship. In both cases, an elas­tic con­nec­tion is used, for ex­ample in the form of an elas­tic rope. The aim of the pro­ject was the math­emat­ical, time-depen­dent mod­eling, nu­meri­cal simu­lation and visu­aliza­tion of an elas­tic rope using suita­ble nu­meri­cal meth­ods. Using a wide varie­ty of case sce­nari­os, it was possi­ble to demonstrate how the de­veloped model works.

History of Franconian Switzerland

After the aca­demic ef­forts, the region and nature were ex­plored through a short hike. With a visit to the Dev­il's Cave in near­by Pot­ten­stein, the par­tici­pants gained an in­sight into the geol­ogy and early histo­ry of Fran­coni­an Swit­zer­land.

Text: Maximilian Bauer, Coordinator of the Elite Graduate Program "Scientific Computing”