About

Through high-performance computing (HPC) and state of the art parallel computing techniques, I focus my research on the reconstruction, characterization, and optimization of heterogeneous materials. On the continuum scale, there is an instrinsic and delicate relationship between the geometry of the pore/particle space, which determines bulk properties. I am interested in physics-based simulation and characterization of bulk properties and their optimization through changes to the phase-space. A good example of one such model, from material reconstruction to topology optimization, can be found in this publication.

More broadly, below are some of the topics I have worked on:

1. 3D Microstructure reconstruction via the Yeong-Torquato (Y-T) algorithm.
2. Pore-size distribution estimation of digital media via erosion-dilation.
3. Correlation function and descriptor based characterization of porous media.
4. Bulk property prediction (effective thermal conductivity, effective diffusivity, and permeability) in porous media via the finite volume method (FVM).
5. Data-driven Convolutional Neural Networks for property prediction.
6. Gradient based and non-gradient based topology optimization.

Mission

Here are broad long-term research objectives:

1. Unified physics platform based on the FVM for generating large machine learning datasets.
2. Computational framework for thermal management material microstructure optimization with all modes of heat transfer.
3. Transport phenomena simulations for energy-related materials characterization and optimization.
4. Open source datasets and benchmarks for physics-related ML projects.

The aforementioned long-term objectives have a plethora of potential applications, and therefore broad impact. The applications include, but are not limited to: geology, petroleoum engineering, geothermal power, nuclear power, batteries, fuel cells, supercapacitors, biomedical implants, carbon capture, energy storage, grid efficiency, HVAC, food science, among many others.

Impact

To the stars and beyond!

Check out other metrics on Scholar and ResearchGate.