Neural-Network-based Viscosity Closure for Non-Newtonian Multiphase Flows

Suresh Murugaiyan, Claire L. Nelson, Dhruv Gamdha, Austin Cunniff, Cheng-Hau Yang, Abraham Wiletsky, Kaitlyn W. Dilley, Patrick Babb, Andrew Rhode, Christopher M. Bates, Angela A. Pitenis, Michael L. Chabinyc, Adarsh Krishnamurthy, Baskar Ganapathysubramanian · 2026

Read Neural-Network-based Viscosity Closure for Non-Newtonian Multiphase Flows by Suresh Murugaiyan, Claire L. Nelson, Dhruv Gamdha, Austin Cunniff, Cheng-Hau Yang, Abraham Wiletsky, Kaitlyn W. Dilley, Patrick Babb, Andrew Rhode, Christopher M. Bates, Angela A. Pitenis, Michael L. Chabinyc, Adarsh Krishnamurthy, Baskar Ganapathysubramanian (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-30659.txt.

Introduction

Murugaiyan and colleagues replace handmade viscosity formulas for 3D printing with a neural network trained on lab data. Carreau-Yasuda and power-law models capture shear behavior well, but each new material demands a derivation and solver rebuild. Their network serves as the viscosity closure inside a CHNS solver, with viscosity depending only on the local shear rate. Lipschitz regularization keeps predictions smooth, and the network ships in ONNX format for runtime queries. Benchmarks and silicone ink tests validate the workflow, with simulated speeds and droplet shapes matching observations.

Catalog record

CREATORSuresh Murugaiyan, Claire L. Nelson, Dhruv Gamdha, Austin Cunniff, Cheng-Hau Yang, Abraham Wiletsky, Kaitlyn W. Dilley, Patrick Babb, Andrew Rhode, Christopher M. Bates, Angela A. Pitenis, Michael L. Chabinyc, Adarsh Krishnamurthy, Baskar Ganapathysubramanian
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE53648
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.30659

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Neural-Network-based Viscosity Closure for Non-Newtonian Multiphase Flows Suresh Murugaiyan, Claire L. Nelson, Dhruv Gamdha, Austin Cunniff, Cheng-Hau Yang, Abraham Wiletsky, Kaitlyn W. Dilley, Patrick Babb, Andrew Rhode, Christopher M. Bates, Angela A. Pitenis, Michael L. Chabinyc, Adarsh Krishnamurthy, Baskar Ganapathysubramanian Abstract Materials used in polymer-based additive manufacturing processes, such as Digital Light Processing (DLP) and direct ink writing (DIW), typically exhibit non-Newtonian rheology. Carreau–Yasuda and power-law models describe basic shear-thinning and shear-thickening behavior well, but applying them to a new material requires choosing a functional form, deriving it, and re-implementing it inside the flow solver. We present a deployment workflow in which a neural network trained on experimental rheometry data serves as the viscosity closure inside a Cahn–Hilliard–Navier–Stokes (CHNS) finite element solver. Lipschitz regularization during training produces smooth viscosity predictions, and the trained network is exported in the Open Neural Network Exchange (ONNX) format and queried by the solver at runtime via the ONNX runtime, without solver modification or network reimplementation. The framework is built on a parallel octree-based adaptive mesh refinement infrastructure that concentrates resolution at the fluid interface. We validate the CHNS solver against benchmark shear-thinning bubble-rise cases from the literature, reproducing reported bubble shapes across varying power-law indices and Weber numbers. We characterized two silicone ink formulations, recorded their rise dynamics in perfluorodecalin on high-speed video, and used the resulting data to test the full workflow. Simulated rise velocities fall within the experimentally measured spread, and the simulated steady-state droplet shape agrees with the observed one. This work contributes to a growing body of literature on integrating neural constitutive closures into multiphysics simulations, and demonstrates a practical path for deploying experimentally trained rheological surrogates inside finite element solvers. 1 Introduction Multiphase flows involving non-Newtonian materials are central to a broad class of polymer-based additive manufacturing processes. In Digital Light Processing (DLP), patterned light from a projector cures a photosensitive resin layer by layer; the resin must flow into place for each new layer while solidifying under exposure. In direct ink writing (DIW) and embedded 3D printing, a non-Newtonian ink is extruded into a support bath, and the interaction between the ink and the surrounding fluid governs filament shape, placement, and print fidelity. In both cases, the coupling between flow, solidification, and the shear-dependent rheology of the printing material determines print resolution and the mechanical properties of the final part. This coupling is difficult to probe directly during printing. As a controlled proxy, we analyze the buoyant rise of a non-Newtonian silicone ink droplet through a viscous fluid, a configuration that imposes shear histories comparable to those encountered during extrusion- and resin-based printing. Many empirical constitutive laws have been developed to describe non-Newtonian viscosity. The Carreau–Yasuda [ 41 ] and power-law [ 31 ] formulations are the standard tools for shear-thinning and shear-thickening behavior in computational studies, capturing the rheology of many polymeric and soft-matter systems with a small number of interpretable parameters. The fixed, low-dimensional form that gives these models their interpretability also limits them: extending them to viscoelastic, thixotropic, or yield-stress behavior generally requires deriving and implementing a new constitutive law inside the flow solver, and refitting parameters for each new system. 9 used Bayesian inference to show that good fits to rheological data are insufficient grounds for model selection; models with fewer, well-constrained parameters are more credible than empirical forms that achieve closer fits through added flexibility. The same concern applies to data-driven constitutive surrogates, including neural networks, which carry genuine model-selection risk and require appropriate regularization when used as replacements for parameterized constitutive forms. The literature on bubble rise in non-Newtonian fluids and on data-driven constitutive modeling has grown substantially, but the two have rarely been coupled: few studies deploy constitutive surrogates trained on experimental rheometry inside high-fidelity multiphase solvers. The remainder of this section reviews three relevant areas: numerical modeling of bubble dynamics in two-phase flow solvers, bubble dynamics in non-Newtonian fluids with established constitutive relations, and recent work on data-driven or machine-learning-assisted constitutive modeling. Bubble dynamics in immiscible fluids has been studied extensively with interface-capturing methods. Two-phase simulations have progressed from qualitative shape comparisons to quantitative benchmarks that track bubble circularity, center-of-mass height, and mean rise velocity, providing reproducible targets for code verification in capillary-sensitive flows [ 13 ] . Three-dimensional studies [ 4 , 30 , 11 , 25 ] have captured complex deformation, instability, and breakup behavior under varied flow conditions. 35 mapped the transitions between axisymmetric rise, skirt formation, path instability, and breakup as functions of the Galilei and Eötvös numbers, and the same group subsequently showed that thermocapillary effects in self-rewetting fluids can reverse or arrest bubble motion [ 34 ] . The Cahn–Hilliard–Navier–Stokes (CHNS) formulation [ 5 , 14 , 1 ] has become a standard diffuse-interface approach for interfacial evolution, coalescence, and breakup with thermodynamic consistency [ 16 , 15 , 27 ] . CHNS solvers paired with adaptive mesh refinement [ 2 , 29 , 40 , 39 ] resolve steep gradients at phase boundaries efficiently [ 16 ] . The studies above mostly concern Newtonian fluids. In non-Newtonian systems [ 6 , 36 , 7 , 42 ] , viscosity varies with local shear rate and affects bubble deformation, rise velocity, and interfacial stability. 26 studied an air bubble rising in shear-thinning and shear-thickening liquids with the Carreau–Yasuda model, finding that stronger shear-thinning raises rise velocity and reduces deformation, and that higher Galilei numbers increase shape distortion. 42 combined experiment and simulation for a single bubble rising in shear-thinning Carreau fluids and reported that stronger shear-thinning lowers local viscosity near the bubble surface, enhances rise velocity, and produces oblate shapes with high-viscosity wakes. 24 examined bubbles rising in shear-thinning power-law fluids under varied gravity conditions and showed that a lower power-law index produces a detached high-viscosity region behind the bubble, while higher Galilei and Eötvös numbers produce larger wake structures and stronger deformation. 10 added that the liquid-to-bubble viscosity ratio also affects rise velocity, with high ratios inducing unsteady deformation at low power-law indices. Data-driven constitutive modeling has emerged as an alternative to fixed parametric forms when the rheology is difficult to capture in closed form. Physics-informed neural networks (PINNs) have been used to infer unknown constitutive relations and material parameters from sparse or indirect flow observations. 32 showed that PINNs can recover nonlinear diffusion and transport relations from limited state measurements without prescribing a functional form. 28 extended this approach to non-Newtonian fluids, learning shear-dependent viscosity distributions from velocity data in three-dimensional flows. For viscoelastic fluids, where stress depends on both kinematics and material memory, 33 introduced ViscoelasticNet, which infers stress fields and material parameters from velocity data alone, even with sparse or noisy inputs. 21 introduced Rheology-Informed Neural Networks (RhINNs), which embed systems of ordinary differential equations describing thixotropic and viscoelastic behavior into the network architecture. The same model predicts stress evolution under varied flow protocols and infers hidden material parameters from limited experimental data, providing accurate forward and inverse solutions for thixotropic elasto-viscoplastic fluids even under sparse or noisy inputs. A later extension [ 22 ] combines sparse regression with automatic differentiation to construct closed-form constitutive relations from experimental data without relying on a predefined model, identifying the dominant physical terms

Read the complete text