Design Teams & School
Masters Thesis:
Topology Optimization of Weldment Structures with Cold-Worked Members
April 2024
-
Constructed a complete topology optimization (TO) solver in Python, utilizing OptiStruct as FEA blackbox.
-
Designed GUI for initial condition construction and viewing results.
-
Utilized geometric primitives to represent weldment members, imposing novel constraints that ensured feasible manufacture of these members via cold-work technology. Additional constraints provided control of process-related material costs.
-
Developed these constraints to be conducive with gradient-based optimization, and within that, usable for both fixed and adaptive meshing techniques.
-
Optimized back-end subroutines that decreased computational time by 65% compared to previous efforts.
Undergraduate Thesis:
Artificial Neural Networks for Inverse Kineto-Static Solutions to Cable-Driven Parallel Robots
April 2021
-
Cable-driven parallel manipulators are a subset of parallel manipulators that are used when large workspaces are required. Beyond a certain threshold workspace size, the weight and elasticity of the cables become mathematically significant; the inverse kinematics and inverse force analysis become intertwined, necessitating nonlinear optimization techniques. This complicates vehicle control.
-
Instead of relying on computationally intensive optimization schemes to conduct kineto-static solutions in real time, Neural Networks were trained to handle the solution process. With the reduction in real-time computational burden, NNs allow use of cheaper on-board processors.
AE505 - 2D Topology Optimization Using Primitive Genetic Algorithm
April 2022
-
The goal of this project was to conduct small-scale topology optimization on 2D shapes.
-
Using a Genetic Algorithm to conduct optimization was known at the outset to be computationally inefficient. The undertaking was therefore one of interest in biological evolution and how analogous rules can be applied in the field of mechanical engineering.
- Rows of boolean values (i.e. bit arrays) were iteratively manipulated via the selection, crossover and mutation evolutionary mechanisms.
- Problem Statement: Minimize compliance subject to a volume equality constraint
- Code can be found on GitHub.
MINE 472 - Path Tracking Controller for Epiroc PV-271 - Matlab
April 2020
-
Designed and validated a full-state feedback controller in MATLAB, to be used in path tracking applications of an autonomous off-road vehicle. Principles utilized during the design included the Extended Kalman Filter, Gaussian sensor noise and vehicle pose error, with the assumption that mapping data is pre-existing (SLAM navigation not integrated).
Mechanical Supervisor - Queen's Space Engineering Team
April 2020 - December 2020
-
Engineering lead for the redesign of the Mars Rover’s chassis, along with design for integrating control systems hardware.
-
Involved in optimizing the Rover’s autonomous navigation system using ROS software.
Senior Propulsion Engineer - Queen's Hyperloop Design Team
June 2018 - August 2019
-
Responsible for the design of the vehicle’s battery enclosure and housing using Solidworks.
- Responsible for calculations determining the propulsion system’s gear ratios for maximum acceleration, and drive-train stress/fatigue analysis, both utilizing MATLAB optimization scripts.
-
Responsible for the design of a mechanism to dynamically adjust the vehicle’s axle height and chain tension using Solidworks.
-
Involved in selection of the propulsion system’s electric motors.
MECH 328 - Mechanical Conveyor Linkage - Kinematic and Dynamic Analysis - Mechanical Design
Fall 2018
-
Completed mathematical analysis and modelling on the kinematic and dynamics characteristics of a reciprocating push rod's linkage system, used in transferring items from one conveyor belt to another.
MECH 323 - RC Gearbox Optimization, Design and Fabrication
Spring 2019
-
Programmed optimization scripts in Matlab to determine an RC vehicle's optimal gear ratios for two scenarios; hill climbing and drag race.
-
The resulting gearbox was able to shift between the two gear-trains, each tailored to their specific application.
-
Shaft and gear static and fatigue stresses analyzed in Matlab, and were designed in Solidworks.
-
All components were 3D printed and assembled prior to testing.
