Large Scale Multiphysics Computations

Intelligent Multiphase Modeling in Microsystems IM³

Principal Investigator

Pavel Ryzhakov

Back

Overview

Research

Staff

Projects

Publications

The IM³ group advances research in multiphase flow modelling, with a particular focus on microsystems. It combines numerical simulations with experiments and AI to tackle real-world challenges.

The Intelligent Multiphase Modelling in Microsystems (IM³) group, led by Dr. Pavel Ryzhakov, is dedicated to advancing research in multiphase flow modelling, with a particular focus on microsystems. The group develops and enhances cutting-edge computational models tailored to address the unique challenges of microfluidic environments, bridging the gap between fundamental research and practical engineering solutions. IM³’s integrated approach combines state-of-the-art numerical modeling, experimental studies, and artificial intelligence, reflecting the complexity of real-world systems and enabling deeper insights into underlying physical phenomena.

Droplet dynamics is central to the group’s work, with the development of innovative computational models and simulations of free droplets, droplet generation, wetting behavior, and multi-droplet interactions. Substantial effort is also dedicated to the development of techniques for simulating the propagation of multiphase flows in complex and/or porous media. Artificial intelligence enhances our activities, laying the groundwork for digital twins and smart control systems for microfluidic devices. IM^3’s projects span digital manufacturing, inkjet printing, water purification, fire safety, and fuel cell water management, delivering innovative, sustainable solutions.

At IM³, we develop advanced computational and experimental tools to model, analyze, and optimize multiphase microfluidic systems. Our work combines physics-based simulations, data-driven methods, and lab-scale experiments to tackle challenges in energy, manufacturing, and biomedical applications.

Research areas

Multiphase Flow Modeling

We develop innovative high-fidelity CFD models for complex microfluidic flows involving multiple phases and materials. Our research includes:

- Enriched finite element methods for multimaterial and multiphase problems
- Interface-capturing techniques (conservative level set, NURBS)
- Coupled physics models (fluid dynamics, electrodynamics, thermodynamics)
- Wetting dynamics and contact line modeling
- Lagrangian particle-based methods (PFEM, SPH)

Outcomes: Modeling of water evacuation in fuel cell flow channels. Simulation of multiphase transport in porous media such as gas diffusion layers. Study of droplet dynamics for applications in inkjet and drop-on-demand technologies.

Applications: Fuel cell water management, porous media transport, inkjet, heat exchangers and boilers, flows in aorta and blood vessels, bioprinting.

AI-Driven Multiphase Simulation

We leverage artificial intelligence and data-driven tools to accelerate simulation and control of droplet and bubble dynamics:

- Image-based analysis and machine learning
- Surrogate modeling and multi-fidelity simulation
- Physics-informed neural networks (PINNs)
- Dimensionality reduction and manifold learning

Outcomes: Optimization and control of printing devices for digital manufacturing, with applications such as bioprinting, pharmaceutical dose control, and lab-on-a-chip systems. Analysis of bubble behavior in electrolysis and hydrolysis processes.

Applications: Digital manufacturing, microdroplet control, bubble behavior in hydrolysis, boilers and heat exchangers.

Microfluidics Experiments

Experiments conducted in our in-house lab provide first hand insights into complex mircofluidic processes.

- Droplet and inkjet generation experiments
- Electrospray (EHDA) and electrospinning experiments
- Sophisticated tools for image processing and data analysis

Outcomes: Large openly available databases for models training and validation. Creation of the core for digital twin development. Determination of optimal manufacturing regimes.