Profile Picture

Michael Shimizu

Ph.D. Student in Applied Mathematics & Statistics
Stony Brook University | Sandia National Laboratories
Focus: Machine Learning, Partial Differential Equations

About

I am a second-year Ph.D. student in Applied Mathematics and Statistics at Stony Brook University, where I am advised by Professor Hong Qin. My current research focuses on the intersection of Machine Learning and Parital Differential Equations, with an interest in leveraging Mamba state space models for improved PDE solutions.

In addition to my work at Stony Brook, I am a Research Intern at Sandia National Laboratories, working on projects to better understand the damage and risk of hail impacts on Photovoltaic modules.

Previously, I completed my undergraduate studies at Cornell University, majoring in Engineering Physics with a minor in Mathematics. My research background includes working with quantum algorithms in the McMahon Lab at Cornell and conducting atomistic polymer simulations with Professor Ting Ge at The University of South Carolina.

Publications & Research

Atomistic Simulations of Mechanical Properties of Lignin Polymers (2024) +

Summary: In order to better understand lignin, a sustainable polymer alternative, we utilized molecular dynamics (MD) simulations to characterize its stress-strain response under various conditions. This study evaluates the impact of temperature, strain rate, and oligomer size (N-values) on the material's rigidity. Our findings establish that lignin exhibits significant strain hardening and maintains a Young’s modulus comparable to standard plastics, validating the use of modified CHARMM force fields for accurate biopolymer modeling.

My Contribution: Constructed and executed the HPC molecular dynamics simulations to analyze stress-strain relationships, temperature effects, and oligomer size variations.

DOI: 10.3390/polym16243552
Hamiltonian-reconstruction distance as a success metric for VQE Quantum Science and Technology (Submitted) +

Summary: Variational Quantum Eigensolvers (VQE) often lack a reliable metric for convergence when the true ground state energy is unknown. This paper proposes the Hamiltonian-reconstruction distance as a new convergence metric that can be estimated efficiently on quantum hardware. Using simulations and trapped-ion experiments, we demonstrated that this metric provides a distinct signal of convergence, allowing for the distinction between local minima and true ground state solutions.

My Contribution: Developed and executed parallelized density matrix simulations to model hardware noise, generated the primary correlation analysis for Figure 3, and contributed to manuscript drafting.

arXiv: 2403.11995
Measurement of PV Module Deformation During Hail Impact IEEE J. Photovoltaics (2024) +

Summary: This study introduces a novel application of Digital Image Correlation (DIC) to capture the transient deformation of photovoltaic modules during hail impact events. By utilizing high-speed stereoscopic cameras, we quantified the out-of-plane displacement fields and calculated derived strain metrics in real-time. The analysis reveals that impact energy scales with the square root of deformation depth and identifies specific damping behaviors in the module's glass-polymer layered structure.

My Contribution: Developed the interpolation code and visualization pipeline responsible for generating the primary deformation and strain analysis plots (Figs 4–7, 9).

DOI: 10.1109/JPHOTOV.2024.3405377
Laboratory Hail Damage of PV Modules: Electroluminescence and High-speed DIC Analysis IEEE PVSC (2024) +

Summary: This conference paper presents a multi-modal analysis of hail damage initiation in PV modules by correlating high-speed mechanical deformation data with Electroluminescence (EL) imaging. The study investigates the threshold at which physical impact results in electrically active cell cracks. By spatially mapping the transient deformation peaks to subsequent EL crack patterns, we provide a clearer understanding of how impact energy translates into internal structural failure in commercial solar panels.

My Contribution: Processed experimental data to visualize the correlation between peak deformation and crack initiation (Fig 4).

DOI: 10.1109/PVSC57443.2024.10749223

Projects

Personal Website

A static minimalist portfolio built with vanilla HTML/CSS/JS. Focused on clean typography and performance.

View Code

More Coming Soon

Cleaning / updating projects...

Teaching

  • TA, AMS 261: Applied Calculus III (Fall 2024, Spring 2025, Fall 2025)