IML | AI | Intelligent Manufacturing Lab

AI for Engineering

Our AI research focuses on developing advanced machine learning and optimization methodologies for engineering applications. We combine physics-based modeling with data-driven approaches to create intelligent systems that solve complex engineering challenges. Our work spans physics-informed machine learning, generative design, uncertainty quantification, surrogate modeling, and engineering design optimization, enabling breakthroughs across diverse domains including energy storage, manufacturing, and thermal management.

Physics-Informed Machine Learning

We develop physics-informed machine learning models that integrate physical laws and constraints with neural networks to enable accurate predictions while respecting fundamental conservation principles. Our research includes multi-fidelity physics-informed convolutional neural networks for thermal prediction, physics-informed neural networks for battery pack temperature estimation, and hybrid approaches that combine experimental data with physics-based models.

Generative Design

We leverage generative AI and diffusion models to explore novel design configurations and discover innovative solutions. Our generative design approaches enable rapid exploration of design spaces, automatic generation of optimal geometries, and discovery of previously unexplored design topologies for applications ranging from battery thermal management to cellular structures.

Uncertainty Quantification

We develop advanced uncertainty quantification methods to assess and propagate uncertainties in engineering systems. Our research includes statistical shape modeling, probabilistic frameworks for design under uncertainty, and multi-fidelity approaches that combine high and low-fidelity information sources to quantify and reduce uncertainty in engineering predictions.

Surrogate Modeling

We create efficient surrogate models and machine learning-based approximations that replace computationally expensive simulations with fast and accurate predictions. Our research includes adaptive surrogate modeling, multi-fidelity surrogate models, and transfer learning approaches that enable rapid design exploration and optimization while maintaining accuracy.

Engineering Design Optimization

We develop optimization frameworks that combine AI, physics-based modeling, and engineering constraints to solve complex design problems. Our research includes reliability-based design optimization, control co-design optimization, multi-objective optimization, and multi-task learning approaches that simultaneously optimize multiple design objectives and constraints.

Related Publications

Internal flow cooling of cylindrical lithium-ion batteries

Dec 1, 2025

Integrating heat transfer and control optimization: A comprehensive review of battery thermal management systems

Sep 1, 2025

Single‐Step Electrochemical Battery Recycling

Aug 1, 2025

Life Cycle Assessment of Direct Recycling for Cathode Active Materials

Jun 1, 2025

Multi-fidelity physics-informed convolutional neural network for heat map prediction of battery packs

Apr 1, 2025

Physics-Informed Machine Learning for Battery Pack Thermal Management