Microdroplet Generator Surrogate Enhancement and Inverse Design Through an Uncertainty-Guided Framework
Authors: David Gabriel Harispe, Matías Gerard, Pablo A. Kler
Organizations: Centro de Investigaci´on en M´etodos Computacionales (CIMEC, UNL–CONICET), Predio CCT CONICET Santa Fe, RN 168, Santa Fe, 3000, Argentina · Departamento de Ingenier´ıa en Sistemas de Informaci´on, FRSF-UTN, Lavaise 610, Santa Fe, 3000, Argentina · Research Institute for Signals, Systems and Computational Intelligence (sınc(i)), FICH–UNL/CONICET. Ciudad Universitaria UNL, Santa Fe, Argentina.
Microfluidic droplet generation is critical for many applications, yet efficient device design in high-dimensional parameter spaces remains computationally prohibitive due to complex multiphase flow dynamics. This work presents a unified methodology that couples adaptive surrogate enhancement with simulation-driven inverse design. High-fidelity Volume-of-Fluid simulations provide ground-truth data, while neural network ensembles predict flow regimes and droplet diameters. Ensemble-based uncertainty quantification identifies epistemic uncertainty using consensus levels for regime classification and the coefficient of variation for diameter prediction, revealing that uncertainty concentrates near regime transition boundaries where physics is highly sensitive to parameter perturbations. A physics-informed enrichment strategy targets these critical zones and addresses class imbalances by deploying additional targeted simulations and retraining the surrogates forming the ensembles. A framework that integrates Particle Swarm Optimization for constrained inverse design, balancing accuracy with predictive confidence and physical consistency is therefore proposed. Validation on independent test datasets demonstrates that the enriched ensembles achieve higher accuracy, with the fraction of predictions within the 0-3% error range on droplet diameter nearly doubling, and high-error cases (above 10%) dropping substantially. The inverse design procedure reliably identifies physically realizable configurations meeting target droplet sizes. This approach significantly reduces computational costs while ensuring reliable predictions, offering a framework with potential applicability to the optimization of a broad range of microfluidic systems.
Figures & tables
Figure 1: Schematic of the proposed optimization and inverse design framework. Operational requirements (1) are processed via an optimization algorithm to generate parametric configurations (2). Neural network ensembles evaluate these to provide feedback on droplet and flow regimes (3) to guide the optimization loop, to output eventually the operational parameters (4) for the microfluidic device. The framework also enables parameter customization (5) to achieve the target dripping regime (6).
Figure 2: Computational domain and coaxial droplet generator geometry used in the numerical simulations. The dispersed phase (orange) is injected through a cylindrical inner capillary into the continuous phase (light-blue), which advects the generated droplets through the surrounding concentric outer capillary.
Figure 3: Panel (I) shows three experimentally observed coaxial flow regimes: dripping, narrowing jet, and widening jet. Panel (II) summarizes the tested conditions from Utada et al. Open symbols denote jetting regimes ( Caout≳0.3−0.5 for the narrowing jet and Win≳2 for the widening jet), while filled symbols denote the dripping regime. Dashed orange rectangles mark regions where the expected regime is not observed. Reprinted figures with permission from Utada et al., http://dx.doi.org/10.1103/PhysRevLett.99.094502 . Copyright 2007 by the American Physical Society.
Figure 4: Schematic of the surrogates training and inference procedure. The black arrows represent the inference flow, where the neural networks (NNs) map input physical parameters directly to the predicted flow regime and, conditionally, the droplet diameter. The blue arrows denote the training flow, which compares the predicted regimes and droplet diameters against the ground-truth Basilisk simulation results ( = and Δ respectively) to drive surrogates training. Note that the droplet diameter prediction module is specifically constrained to operate only when the regime prediction NN identifies a dripping regime
Figure 5: Schematic overview of the proposed Adaptive Surrogate-Refinement Pipeline, whose closed loop couples ensemble-based uncertainty quantification with targeted Basilisk VOF simulations and dataset-driven model retraining, to progressively improve surrogate fidelity. The upstream optimization of baseline architectures and hyperparameters lies outside the loop and supplies the optimized models used to build the ensembles.
Figure 6: Overview of the dataset generation and enrichment pipeline. The flow goes from literature-based and training datasets through balancing, synthetic sampling, training dataset enrichment, and uncertainty reduction verification, as well as final physical predictions test datasets. Dashed red arrows indicate interpolation procedures, whereas solid green arrows denote both interpolation procedures and FVM-VOF simulations. DR: Dripping Regime, NJ: Narrowing-Jet regime, WJ: Widening-Jet regime.
Stage 1: Architecture
Stage 2: Hyperparameters
Cross-validation
Stratified 8-fold on 80% of TR1 ; 15% internal validation split
15 configurations; batch: 5–20; epochs: 200–1600; learning rate: 2.5×10−4 – 4×10−3 ; early stopping off (best model retained)
Fixed parameters
Batch =16 ; epochs =1500 ; learning rate =1.3×10−3 ; early stopping off (best model retained)
Architecture from Stage 1
Selected
Layers 7-40-5-1; 10% dropout after the 40-neuron layer
Batch =16 ; epochs =1600 ; early stopping off (best model retained); learning rate =8×10−4
Table 2: Optimization settings and selected configuration for the droplet diameter regression network.
Figure 7: Distribution of the consensus level over the samples in S1 . The horizontal axis shows the number of ensemble members, out of N=20 , that voted for the most voted regime, which corresponds to Nc(x) in Eq. ( 11 ). The vertical axis is in logarithmic scale, and the number above each bar indicates the sample count. Most samples (14,361 of 15,000) received a unanimous vote.
Figure 8: Kernel density estimates of the samples in SH1 along three input parameters. For each parameter, the black curve corresponds to all samples in the set and the red curve to the samples with non-unanimous votes ( c(x)<1 ), whose percentage is indicated in the legend. The red marker indicates the maximum of the red curve.
Figure 9: Distribution of the cv for droplet diameter predictions. The dashed green line indicates the 20% threshold used to define excessive uncertainty. The inset provides a magnified view of the distribution’s tail, illustrating the presence of high- cv samples reaching up to 70%, which characterize the pre-enrichment state of the ensemble.
Figure 10: Spatial distribution of high predictive uncertainty ( cv>20 %) for droplet diameter predictions on the ( Caout , Win ) plane. High-uncertainty samples are primarily clustered along the regime transition boundaries, whereas the center of the dripping regime region exhibits mostly high ensemble confidence.
Figure 11: Spatial distribution of SH2 synthetic cases with cv>20 % (light green) and the ∼ 200 interpolated combinations (red and blue) on the ( Caout , Win ) plane. The red dots indicate cases that produced dripping regime on the simulation, while blue dots are for widening jet. The black line represents the documented approximate boundary of the dripping regime. Interpolated points for the enrichment process are situated in the main regions of high epistemic uncertainty.
Figure 12: Kernel density estimates of the samples in SH3 along three input parameters, before and after enrichment. Each panel corresponds to the subset of SH3 in which only the indicated parameter is sampled uniformly. The black curve corresponds to all samples in the subset, and the red and orange curves to the samples with non-unanimous votes ( c(x)<1 ) for the baseline ( Er1 ) and retrained ( Er2 ) regime ensembles, respectively. The percentage of non-unanimous samples is indicated in the legend, and the markers indicate the maximum of each curve.
Figure 13: Distribution of the coefficient of variation cv of the diameter predictions before ( Ed1 , blue) and after ( Ed2 , red) enrichment. I) Evaluation on SH4 , generated from TD1 and filtered with Er1 . II) Evaluation on SH5 , generated from T2 and filtered with Er2 . Each dataset contains 1.2 million synthetic samples. The dashed line marks the threshold of 20%. The boxes give the 95th percentile and the maximum of cv for each ensemble.
Figure 14: Classification of the configurations of GR on the ( Caout , Win ) plane by the baseline ( Er1 ) and retrained ( Er2 ) regime ensembles. Fixed configurations were misclassified by Er1 and correctly classified by Er2 . Unaffected configurations were correctly classified by both ensembles. Unfixed configurations were misclassified by both ensembles. Degraded configurations were correctly classified by Er1 and misclassified by Er2 . The black line marks the regime-transition boundary between dripping regime and narrowing-jet or widening-jet regimes ( Utada et al., 2007 ) .
Figure 15: Change in the magnitude of the relative diameter error, Δ=∣εd2∣−∣εd1∣ , for the configurations of GD on the ( Caout , Win ) plane. Green indicates a lower error of the retrained ensemble, and red a higher error. The color scale is nonlinear and saturates at ±10 percentage points. The black line marks the regime-transition boundary. The configurations were sampled preferentially in the high- cv regions of Section 4.3.2 .
∣εd∣ (%)
Ed1
Ed2
[0,3]
47
91
(3,5]
39
40
(5,10]
47
29
(10,∞)
38
11
Total
171
171
Table 3: Number of configurations of GD in each range of the magnitude of the relative diameter error, ∣εd∣ , for the baseline ( Ed1 ) and retrained ( Ed2 ) diameter ensembles. Square brackets denote inclusive boundaries, while parentheses denote exclusive boundaries.
Parameter
Symbol
Range
Capillary number of the outer phase
Caout
[0.047,0.8]
Weber number of the inner phase
Win
[0.035,20.0]
Viscosity of the outer phase
μout
[8.9×10−3,9.5×10−2] Pa ⋅ s
Viscosity of the inner phase
μin
[8.9×10−4,8.9×10−2] Pa ⋅ s
Radius of the inner capillary
rin
[5.0×10−6,2.5×10−5] m
Table 4: Ranges of the free parameters in the inverse design problem.
Quantity
Symbol
Value
Number of particles
N
50
Inertia weight
w
U(0,1)
Cognitive acceleration constant
c1
2
Social acceleration constant
c2
2
Velocity clamping coefficient
δ
0.9
Boundary rebound factor
αb
−0.7
Table 5: Parameters of the PSO algorithm and termination criteria.
Case
ddroptarget
Caout
Win
μout
μin
rin
dpred
dsim
Error
( μ m)
(mPa ⋅ s)
(mPa ⋅ s)
( μ m)
( μ m)
( μ m)
(%)
1
21.75
0.323
0.164
22.6
0.890
5.01
21.75
21.44
-1.46
2
0.275
0.235
15.1
1.60
5.00
21.75
20.81
-4.53
3
0.289
0.0356
9.21
1.16
8.00
21.75
22.22
2.10
4
28.52
0.337
0.110
43.6
0.893
6.61
28.52
29.22
2.38
5
0.363
0.0600
56.0
0.890
6.89
28.52
29.15
2.15
Table 6: Configurations obtained by inverse design and their validation with VOF simulations. For each target diameter ddroptarget , three independent PSO runs were performed. Values in bold were fixed during the search. dpred is the diameter predicted by Ed2 , and dsim the simulated diameter. The error is computed as (dsim−dpred)/dsim×100% . Simulations exhibited dripping regimes in all the configurations.
Center for Machine Learning (ZML) 2Institute for Flow in Additively Manufactured Porous Structures (ISAPS) Heilbronn University of Applied Sciences, 74081 Heilbronn, Germany
Department of Mathematics and Computer Science, Eindhoven University of Technology · Department of Chemical Engineering and Chemistry, Eindhoven University of Technology