Andre Weiner, weiner@mma.tu-darmstadt.de
First supervisor: Prof. Dr. rer. nat. Dieter Bothe
Second supervisor: Prof. Dr.-Ing. Peter Stephan
micro reactor
size: millimeter
source: SPP 1740
prediction of
bubble column reactor
size: meter
source: R. M. Raimundo, ENI
U. D. Kück, M. Schlüter, N. Räbiger:
Analyse des grenzschichtnahen Stofftransports an frei aufsteigenden Gasblasen (2009)
$d_b=1~mm$ water/oxygen at room temperature
$\delta_h/\delta_c$ typically 10 ... 100
feasible simulations up to $Pe\approx 1000$ (3D, HPC)
Idea: Leverage coherent structures in convection-dominated species boundary layers.
→ saves one refinement level (half cell-width)
$Sh = \frac{k_L d_b}{D}$ with $k_L = \frac{\dot{N}}{A_{eff}\Delta c}$
Idea: Replace analytical solution with machine learning (ML) model.
$s$ instances of feature - label pairs
$s$ | $x_{1,s}$ | $x_{2,s}$ | $x_{3,s}$ | $y_s$ |
---|---|---|---|---|
1 | 0.1 | 0.4 | 0.6 | 0.5 |
2 | ... | ... | ... | ... |
numerical solution
data extraction
Multilayer Perceptron
model training
implementation
Aim:
Create reference data for complex shapes and flow scenarios to assess model generalization.
Idea:
Decoupling of two-phase flow and species transport.
1. Two-phase flow simulation (Volume-of-Fluid)
2. Parametrization of shape and interfacial velocity
3. Geometry generation and export (STL format)
4. Single phase mesh
5. Flow solution
6. Species transport
Prismatic cell layers around a spherical-cap bubble with increasing mesh resolution.
Special thanks to David Merker, Jens Timmermann, Chiara Pesci and Dennis Hillenbrand
Thank you for your attention!
Get in touch: weiner@mma.tu-darmstadt.de
Solution based on Git/Github, Docker/Dockerhub and TUDatalib.
Feature density for the source terms of a parallel-consecutive reaction of type $A+B\rightarrow P$ and $A+P\rightarrow S$.
Video by courtesy of David Merker.
Image by courtesy of David Merker. Measurement apparatus by dataphysics-instruments.
Idea: ranking of feature importance.
algorithms
example (left)
$f_m = f_m(\bar{c},Da,\tilde{r})$
Histogram (left) and heatmap (right) of the label error for the transfer species (A). Reminder: the height of a bar in a histogram depicts the number of instances in the range the bar covers. The index $01$ indicates a min-max-scaling.