Fluid Dynamics & Numerical Analysis

We investigate a wide range of flow phenomena—flow separation and wakes around complex geometries, fluid–structure interaction (FSI), and multiphase flows—using high-fidelity numerical methods including direct numerical simulation (DNS). Our goal is to uncover the underlying physical mechanisms and to predict these phenomena with high accuracy.

HPC / GPU Computing

Media coming soon

We leverage massively parallel computing and GPU acceleration to run high-resolution flow simulations on grids of hundreds of millions of cells at high speed. By maximizing computational efficiency and scalability, we enable analyses of large-scale problems that were previously out of reach.

Turbulence & Transition Control

We control the laminar-to-turbulent transition process and near-wall flow separation through active and passive techniques. By designing and validating control strategies, we target the performance gains required in real-world applications, such as drag reduction and flow stabilization.

Design & Optimization of Aero and Power-Generation Gas Turbines

We characterize the internal flow and combustion of gas turbines used in aircraft propulsion and power generation through precise numerical analysis. Based on these results, we optimize blade geometry and operating conditions to improve efficiency and stability.

Steam Turbine

We analyze the complex flow structures and loss mechanisms that develop inside steam turbines using numerical simulation, thereby evaluating stage performance and efficiency and identifying directions for improvement.

Reinforcement-Learning (AI) Control of Wind-Turbine Pitch and Yaw

We apply reinforcement-learning (AI)-based control to optimize wind-turbine blade pitch and yaw angles in real time. The goal is to maximize power output while reducing structural loads, even under fluctuating wind conditions.

Optimal Design of the Porous Transport Layer (PTL) for Hydrogen Fuel-Cell Vehicles

We optimize the microstructure of the porous transport layer (PTL), a key component of hydrogen fuel-cell electric vehicles. By improving mass and heat transport, we enhance fuel-cell efficiency and durability.

Turbulence Prediction

We predict the spatiotemporal evolution of turbulent flow fields using neural-network–based models. Prediction results are compared with ground-truth simulations to verify the model’s accuracy and reliability.

Optimal Analysis & Design

We introduce machine learning to accelerate flow analysis and efficiently explore large design spaces, substantially reducing computational cost while deriving optimal designs.

Turbulence Modeling

We develop data-driven, neural-network–based turbulence models to overcome the limitations of conventional models, aiming for reliable turbulence analysis with improved predictive accuracy and generalization across diverse flow conditions.

Quantum-Native Numerical Algorithms

We develop fluid-dynamics numerical algorithms optimized for the characteristics of quantum computing hardware, targeting quantum acceleration of large-scale flow problems that are intractable for classical computing.

Quantum Noise Effects

Media coming soon

We quantitatively analyze how the noise inherent in today’s quantum hardware affects the accuracy of numerical results, and study analysis techniques that remain reliable in noisy environments.

Quantum Machine Learning

We apply quantum machine learning to fluid-dynamics problems to explore new analysis and prediction methodologies, aiming for a next generation of computational fluid dynamics that combines quantum computing with data-driven approaches.