Microstructure-aware fracture and damage
How heterogeneities, pores, inclusions, grain boundaries, and phase distributions steer crack initiation, growth, localization, and toughness.
Technion · Mechanical Engineering
Failure, but make it useful.
CEµM² studies how microstructure controls deformation, damage, and fracture in engineering materials. We combine experiments, mechanics-based modeling, numerical simulation, statistical analysis, and data-driven methods to understand failure across scales and to design more reliable materials.
Research themes
Materials do not fail randomly. They merely hide the evidence poorly.
Our research themes connect experiments, modeling, simulation, and data analysis across the scales where materials deform, localize, and fail.
Dynamic failure and extreme loading
Experiments and modeling for high-rate deformation, adiabatic shear, thermal conversion, localization, and impact-driven fracture.
Quantitative and machine-learning fractography
Using surface morphology, computer vision, and statistical descriptors to turn fracture surfaces into mechanical evidence.
Computational micromechanics and simulation
FEM, FFT, homogenization, constitutive modeling, and data-driven workflows for heterogeneous materials.
Process-structure-property relations
How processing routes and microstructural evolution influence mechanical behavior, reliability, and failure.
Open mechanics education
Course notes, computational notebooks, and teaching material for mechanics, FEM, fracture, and constitutive modeling.
Teaching
Open mechanics teaching resources.
Current course notes and archived teaching material are collected here as reusable resources for students and collaborators.
Publications
Selected publications from the group.
2021Ti-6Al-4V Hybrid Structure Mechanical Properties – Wrought and Additive Manufactured Powder-Bed Material
2021M7C3: The Story of a Misunderstood Carbide
2021Experimental and numerical study of the interaction between dynamically loaded cracks and pre-existing flaws in edge impacted PMMA specimens
2021Unsupervised Machine Learning in Fractography: Evaluation and Interpretation
2020Toward quantitative fractography using convolutional neural networks
2020The BCC-FCC Phase Transformation Pathways and Crystal Orientation Relationships in Dual Phase Materials From Al-(Co)-Cr-Fe-Ni Alloys
Latest from the blog
Recent notes from the lab.
PINNs Are Mesh-Free, Not Cost-Free
A lifecycle-accounting note on physics-informed neural networks, solver baselines, and the hidden cost of pretending inference time is the whole workflow.