Projects

In this project, I engineered advanced bone-mimicking scaffolds by integrating computational geometry analysis with implicit modeling techniques. My primary role was to define and replicate the precise topological features of native bone to create functionally superior lattice structures for tissue engineering. Advanced Geometric Analysis: ⦁ Processed high-resolution µCT scans of bone tissue to generate high-fidelity triangle meshes. ⦁ Implemented a multiscale fitting method using the C++ library libigl to calculate principal (κ1​,κ2​) and Gaussian (K) curvatures at each vertex, quantifying the local surface shape. ⦁ Characterized the bone’s unique Interface Shape Distribution (ISD) by applying bivariate Kernel Density Estimation (KDE) to the curvature data, providing a quantitative benchmark for bio-mimicry. Our analysis confirmed the prevalence of hyperboloidal (saddle-shaped) surfaces, a key driver for osteogenesis. Implicit Modeling & Lattice Design: ⦁ Utilized MATLAB to programmatically generate complex, porous architectures based on Triply Periodic Minimal Surfaces (TPMS), including Gyroid and Diamond structures. ⦁ We developed gradient and hybrid-gradient (hybrid-G) lattices by manipulating the implicit TPMS equations to vary porosity and combine geometries, directly mimicking the heterogeneity found in native bone. ⦁ Engineered all models to match the target average porosity (70%) of the bone specimen by precisely offsetting the implicit surfaces. Computational Verification & Visualization: ⦁ Validated our lattice designs by performing the same rigorous curvature analysis on them and comparing their ISDs against the original bone data benchmark. ⦁ Employed Paraview to create detailed 3D visualizations of the Gaussian curvature fields, confirming that hybrid-G design most accurately replicated the critical hyperboloidal characteristics of the target tissue. Publications related to this project: 10.1016/j.mser.2024.100876

Jun 1, 2024

In this project, I developed a novel method to overcome the fundamental resolution limits in projection stereolithography (SLA) 3D printing, enabling the fabrication of features smaller than a single hardware pixel. Impact: This work redefines the print capabilities of projection SLA. The technique enables the direct printing of complex, assembly-free components from both polymers and ceramics with unprecedented precision, eliminates the need for post-processing and assembly, and makes previously impossible designs (such as preassembled gearsets) achievable. Publications related to this project: Coming soon.

Apr 1, 2024

This project focuses on developing in-situ monitoring methods for Stereolithography (SLA) 3D printing, with an emphasis on porosity detection and 3D reconstruction of printed part. We’re on track to publish the first paper of this project by the end of 2025, so stay tuned! Publications related to this project: Coming soon.

Aug 1, 2023

In this project, we developed and validated a novel design methodology to enhance the manufacturability and buckling strength of metallic plate lattices. The core challenge was to introduce micro-holes for powder removal in additive manufacturing without the typical reduction in mechanical performance, particularly against buckling failure in low-density structures. Theoretical & Numerical Method Development: ⦁ We established a design framework rooted in a Rayleigh quotient-based theoretical criterion to identify the optimal locations for micro-holes that actively increase the critical buckling load of a plate. ⦁ To implement this, we developed a robust numerical method to evaluate the second-order derivatives of the buckling eigenmode. This was achieved by directly utilizing the shape functions of 8-node second-order shell elements (S8R) within Abaqus. Finite Element Analysis & Design Optimization: ⦁ We applied this methodology to design optimized Simple Cubic (SC), Body-Centered Cubic (BCC), and Face-Centered Cubic (FCC) perforated plate lattices. ⦁ Using Abaqus/Standard and Abaqus/Explicit, we performed a comprehensive suite of simulations, including linear eigenvalue buckling on Representative Volume Element (RVE) models with periodic boundary conditions, nonlinear post-buckling analysis incorporating geometric and material nonlinearities, and large-deformation compression simulations on full-scale models. ⦁ Our analysis successfully demonstrated that the optimized designs increased critical buckling stress by up to 15.1% compared to unperforated lattices, while maintaining comparable post-buckling compressive strength. Experimental Validation & Characterization: (by our collaborators) ⦁ We fabricated the optimized lattice specimens using μ-LPBF with SS316L powder. ⦁ Through quasi-static compression testing, we experimentally verified that the fabricated lattices exhibit superior mechanical properties. Publications related to this project: 10.1016/j.matdes.2024.113544

Sep 1, 2022

This project focuses on analysis and optimization of thin struts in lattice core sandwich panels. Lightweight lattice cores are easy to fail by buckling due to its long slenderness ratio. We optimized thin struts in lattice core sandwich structures to improve buckling resistance. The key innovation was an efficient, bottom-up methodology that avoids computationally expensive optimizations on entire structures. Methodology & Strut Optimization We developed an efficient, bottom-up design methodology to improve the buckling strength of lattice core sandwich structures. The strategy involved first identifying the optimal, non-uniform cross-sections of individual struts by maximizing their buckling eigenvalues using a Python-Abaqus scripted workflow. Strut profiles were modeled with a Fourier Series (FS) representation, and a key finding was that the optimal shapes are proportionally scalable, creating a reusable library of high-performance struts. Lattice Design & Simulation We designed four distinct non-uniform lattice cores (Kagome, BCC, BCCZ, and Six-fold) by strategically replacing uniform struts with pre-optimized shapes based on their specific, identified buckling modes (first or second-order). Comprehensive eigenvalue and nonlinear post-buckling (Riks method) analyses in Abaqus confirmed these designs achieved a 16-21% improvement in critical buckling load over uniform counterparts. Experimental Validation We validated the designs by fabricating BCC and Six-fold specimens from Nylon 12 powder using Selective Laser Sintering (SLS). Quasi-static compression tests demonstrated a measured compressive strength improvement of 18.77% for the BCC structure and 10.00% for the Six-fold. We also developed a high-fidelity nonlinear FE model that incorporated material plasticity and a measured 80µm surface roughness layer, achieving excellent agreement with experimental peak loads and failure mechanisms. Publications related to this project: 10.1080/0305215X.2022.2163239

Sep 1, 2021