Data-driven modeling and validation of reactive mass transfer at rising bubbles

Andre Weiner, TU Braunschweig, Institute of Fluid Mechanics
These slides and most of the linked resources are licensed under a
Creative Commons Attribution 4.0 International License.

✉	Mail	☍	LinkedIn
✍	Blog	✪	Github

Outline

Data-driven modeling overview
Modeling of reactive mass transfer
Outlook
- Extracting coherent flow structures
- Closure modeling as a control problem

Combining ML and CFD
why and how

Why combine CFD and ML?

CFD

produces large amounts of complex data
requires data or representations thereof

finds patterns in data
creates useful representations of data

What is data?

primary data: scalar/vector fields, boundary fields, integral values


# log.rhoPimpleFoam
Courant Number mean: 0.020065182 max: 0.77497916
deltaT = 6.4813615e-07
Time = 1.22219e-06

PIMPLE: iteration 1
diagonal:  Solving for rho, Initial residual = 0, Final residual = 0, No Iterations 0
DILUPBiCGStab:  Solving for Ux, Initial residual = 0.0034181127, Final residual = 6.0056507e-05, No Iterations 1
DILUPBiCGStab:  Solving for Uy, Initial residual = 0.0052004883, Final residual = 0.00012352706, No Iterations 1
DILUPBiCGStab:  Solving for e, Initial residual = 0.06200185, Final residual = 0.0014223046, No Iterations 1
limitTemperature limitT Lower limited 0 (0%) of cells
limitTemperature limitT Upper limited 0 (0%) of cells
limitTemperature limitT Unlimited Tmax 329.54945
Unlimited Tmin 280.90821
					
					
						Checking geometry...
						...
						Mesh has 2 solution (non-empty) directions (1 1 0)
						All edges aligned with or perpendicular to non-empty directions.
						Boundary openness (1.4469362e-19 3.3639901e-21 -2.058499e-13) OK.
						Max cell openness = 2.4668495e-16 OK.
						Max aspect ratio = 3.0216602 OK.
						Minimum face area = 7.0705331e-08. Maximum face area = 0.00033983685.  Face area magnitudes OK.
						Min volume = 1.2975842e-10. Max volume = 6.2366859e-07.  Total volume = 0.0017254212.  Cell volumes OK.
						Mesh non-orthogonality Max: 60.489216 average: 4.0292071
						Non-orthogonality check OK.
						Face pyramids OK.
						Max skewness = 1.1453509 OK.
						Coupled point location match (average 0) OK.

secondary data: log files, input dictionaries, mesh quality metrics, ...

Examples for data-driven workflows

Example: creating a surrogate or reduced-order model based on numerical data.

Example: creating a space and time dependent boundary condition based on numerical or experimental data.

Example: creating closure models based on numerical data.

Example: active flow control or shape optimization.

But how exactly does it work?

ML is not a generic problem solver ...

Supervised learning

Creating a mapping from features to labels based on examples.

Unsupervised learning

Finding patterns in unlabeled data.

(Deep) Reinforcement learning

Create an intelligent agent that learns to map states to actions such that cumulative rewards are maximized.

What if my problem does not fit into these categories?

$\rightarrow$ mathematical, physical, numerical modeling

Reactive mass transfer at rising bubbles

A single-phase simulation approach to compute the mass transfer at rising bubbles

https://github.com/AndreWeiner/sgs_model_test_transient

Joint work with D. Bothe, CMY. Claassen, IR. Hierck, JAM. Kuipers, MW. Baltussen.

Gas-liquid reactors

micro reactor
size: millimeter
source: SPP 1740

prediction of

mass transfer
enhancement
mixing
conversion
selectivity
yield
...

bubble column reactor
size: meter
source: R. M. Raimundo, ENI

Specimen calculation

$d_b=1~mm$ water/oxygen at room temperature

$Pe = Sc\ Re = \nu_l / D_{O_2} \cdot U_b d_b/\nu_l \approx 10^5 $
$$ Re\approx 250;\quad \delta_h/d_b \propto Re^{-1/2};\quad\delta_h\approx 45~\mu m $$
$$ Sc\approx 500;\quad \delta_c/\delta_h \propto Sc^{-1/2};\quad\delta_c\approx 2.5~\mu m $$

$\delta_h/\delta_c$ typically 10 ... 100

feasible simulations up to $Pe\approx 10000$ (3D, HPC)

Why we might be interested in a simplified simulation approach:

perform parameter studies
validate boundary layer models
generate data for ML-based models
...

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

Comparison of two-phase and single-phase flow fields.

Influence of bubble size on local selectivity.

Two-phase velocity field (left half) versus single-phase velocity field (right half); speed-up of 20-40x with 120x finer mesh at surface.

Comparison of concentration fields: data-driven vs. free-slip.

Local Sherwood number for selected time instances.

Inlet boundary condition - first model:

$$ \mathbf{u}_{in} = -\tilde{U}_b\mathbf{e}_y $$

$\tilde{U}_b$ - the bubble's terminal velocity

What would be a more suitable Ansatz for $\tilde{U}_b$?

$\tilde{U}_b = f_\theta(\tilde{t})$
$\tilde{U}_b = f_\theta(\tilde{t})\tilde{t}$

$\tilde{t}$ - dimensionless time; $f_\theta(\tilde{t})$ - neural network model.

Interface velocity boundary condition - second model:

$$ \mathbf{u}_s = \tilde{u}_t\mathbf{t} $$

$$ \tilde{u}_t = \tilde{u}_t(\vartheta, \tilde{t}) $$

$\tilde{u}_t$ - tangential component of interface velocity vector.

Sine and cosine functions with vertical lines at $0$ and $\pi$.

How can we enforce the correct symmetry in the model?

$\tilde{u}_t = f_\theta (\vartheta, \tilde{t})\mathrm{sin}(\vartheta)$
$\tilde{u}_t = f_\theta (\vartheta, \tilde{t})\mathrm{cos}(\vartheta)$
$\tilde{u}_t = f_\theta (\mathrm{sin}(\vartheta), \tilde{t})\mathrm{sin}(\vartheta)$
$\tilde{u}_t = f_\theta (\mathrm{cos}(\vartheta), \tilde{t})\mathrm{sin}(\vartheta)$