Local GPU workstations
Three Dell Precision 3680-class workstations with Intel Core i9-14900 processors and NVIDIA GeForce RTX 4090 GPUs with 24 GB GPU memory support data processing, simulation, visualization, and machine-learning development.
From intelligence
to impact.
Scientific AI, mechanics, and materials meet in our research on predictive engineering and advanced manufacturing.
01 / AGENTIC AI
Agentic AI for Advanced Manufacturing
We develop agentic AI frameworks that connect reasoning, planning, and engineering simulation for advanced manufacturing. By bringing physics-based models and computational tools into coordinated workflows, we aim to support autonomous exploration of process conditions, informed decision-making, and adaptive manufacturing.
02 / SCIENTIFIC AI
Mechanistic computational intelligence for foundation models
We develop separable neural architectures that bring training, solving, and calibration into one computational framework. This work builds on HiDeNN and interpolating neural networks, linking machine learning with engineering simulation.
03 / MANUFACTURING CONTROL
Real-time inverse control of additive manufacturing
KHRONOS explores manufacturing pathways that achieve a desired material response. The architecture supports real-time inverse control without retraining or calibration.
04 / INVERSE DESIGN
Designing the microstructure
JANUS uses a deterministic neural architecture to generate target microstructures from physics-informed criteria. We study the process–structure–property relationships behind 3D-printed alloys.
05 / BEYOND EARTH
The physics of manufacturing in space
We study how molten flow and microstructure interact during welding in microgravity. Fundamental models help reveal the physics needed for reliable manufacturing beyond Earth.
06 / MATERIALS PHYSICS
Understanding materials at the nanoscale
Molecular dynamics, density functional theory, cellular automata, and scientific machine learning help us understand deformation mechanisms and connect terrestrial and extraterrestrial materials applications.
07 / AEROSPACE ENGINEERING
Inverse design in high-dimensional space
Multi-fidelity surrogates capture full-field physics across high-dimensional design spaces. We apply these tools to inverse design for next-generation wing structures.
08 / ADVANCED MANUFACTURING
Additive friction stir deposition
Working with industry, we develop high-fidelity digital twins for additive friction stir deposition. These models aim to predict defects and improve the reliability and affordability of large aerospace components.
Research facilities
Phillips Haas–Meltio VF3
Access in collaboration with Prof. Rao to a five-axis hybrid laser-wire directed energy deposition system with dual-wire feed, a 1.2 kW laser, and open controls for feedback-control research.
Bambu Lab H2D printers
At least three machines are available across the lab and VT-MADE center. Integrated cameras and sensor monitoring support manufacturing experiments and hands-on student training.
Nikon VOXLS 40 C 450 X-ray CT
Dual-source, dual-detector computed tomography with resolution down to 10 micrometers and Volume Graphics analysis tools for evaluating additive-manufacturing defects.
eXpert 9313 fatigue tester
A 1 kN rotating-beam fatigue tester with servo-controlled bending moment, operation up to 6,000 RPM, and an elevated-temperature option supports high-cycle fatigue characterization.
Aerospace Structures and Materials Laboratory
Shared access to mechanical and impact testing, microscopy, digital image correlation, thermal imaging, real-time control, and data-acquisition systems supports materials and structures research.
Virginia Tech Advanced Manufacturing Center (VT-MADE)
IMMPACT Lab is a member laboratory, and Dr. Saha is affiliated faculty. Center resources include additive and hybrid manufacturing systems across metals, ceramics, polymers, and composites, plus the Learning Factory for hands-on education.
NIST AM Bench 2022
Our group received awards in five major categories of the NIST Additive Manufacturing Benchmark Challenge.
Watch our research talks ↗Our sponsors
