
Nuwantha Kumara
I build computational geometry and physics solvers - software that does the hard mathematical work inside engineering design tools. My background is mechanical engineering, but most of my work lives at the intersection of numerical methods, C++, and CAD kernel development.
Colombo, Sri Lanka
My Journey
Education • Internships • Work Experience
Skills
Explore my technical skill set through an interactive force-directed graph. Hover to highlight related technologies, drag to rearrange, and scroll to zoom in or out.
Skill Hierarchy
Click categories to expand/collapse. Hover items to highlight connections in the graph.
C++
Python
MATLAB
Java
Processing
Arduino
SolidWorks
AutoCAD
ANSYS
COMSOL
OpenFOAMNumPy
SciPy
OpenCV
Qt
VTK
Matplotlib
p5.js
ParaViewJupyter
LaTeX
MySQL
Publications
Research papers and academic contributions.
A Parametric Compact Thermal Modeling Framework for Rapid Design Space Exploration of Radial Multistage Thermoelectric Coolers
Presents a Compact Thermal Model (CTM) for a novel radial multistage thermoelectric cooler (TEC) under 1D heat transfer assumptions. By constructing an equivalent thermal resistance network solved via a linear system of equations, nodal temperature profiles are generated rapidly without FEA overhead. Verified against high-fidelity COMSOL Multiphysics simulations (<5% relative error in nominal design regions; Spearman correlation ρ > 0.93 across the global design space), the framework enables statistically sound, high-throughput design space exploration and multi-fidelity optimization for MEMS TECs.
Linear Matrix Formulation for Directional Tangency Constraints in Parametric Curve Design
In Computer-Aided Design (CAD), parametric curves enable the creation of arbitrary yet controllable shapes. During sketching, designers often apply geometric constraints such as fixed points or tangency, leading to non-linear systems that are costly to solve in real time. This work presents a linear matrix formulation for enforcing directional tangency constraints without optimization algorithms. The method was implemented in Python and C++, and integrated into a commercial turbomachinery design tool to maintain blade-edge tangency constraints interactively.