Andre Weiner
TU Dresden, Institute of fluid mechanics, PSM
These slides and most
of the linked resources are licensed under a
Creative Commons Attribution 4.0
International License.
Refer to OPAL for a detailed schedule of all lecture and exercises.
learn to communicate professionally
Note: a Github account is not required but recommended.
How to open a new issue:
How to write an issue report:
Optional tips for writing issues:
See also: 45 Github Issues Dos and Don't
the study of fluid mechanics ...
ML ...
primary (physical) data
secondary data (the rest)
Progression of snapshot size in direct numerical simulations (DNS).
Simulations become more and more sophisticated:
$\rightarrow$ closure models and decision making required
Examples: surrogate modeling, flow analysis
Examples: boundary conditions, material models
Examples: scale/complexity-reduced modeling
Examples: flow control, parameter optimization
matching given inputs and continuous outputs
matching given inputs and discrete outputs
finding low-dimensional representations
grouping similar data points
learning optimal control in dynamic environments
Creating a transport model $\mu (T)$ based on experimental data ...
Predicting the impact behavior of a droplet on a surface ...
Switching between two turbulence models in a RANS simulation ...
Finding coherent structures in turbulent flows ...
Closed-loop active flow control of the flow past a cylinder
hints to create sensible ML applications
Lectures 2 & 3: write a CFD solver from scratch
Lectures 4 & 5: learn approximate velocity profiles
Lecture 6: predict the stability regime of rising bubbles
Lectures 7: compute mass transfer at high Schmidt numbers
Lectures 8 & 9: find coherent structures in turbulent flows
Lecture 10: create a reduced-order flow model
Lecture 11: optimal open-loop flow control
Lectures 12 & 13: learn an agnostic control law from scratch