Winston J. Hu, Ph.D.

Email: win.j.hu@proton.me

PORTFOLIO

Rapidly-deployable wind energy system

Co-founder, Lead Engineer, Rival Lab Inc.

Rival Lab Inc. aims to design a modular, rapidly-deployable wind energy system for the Canadian Department of National Defence (DND) and remote indigenous communities to decrease reliance on diesel fuel. Our start-up has received $500k in funding.

As the lead engineer, I am responsible for:

development of the technology demonstration unit, which is currently under evaluation in the harsh winter of Mont-Tremblant, QC
development of a fully automated wind turbine rotor testing platform capable of characterizing a blade design in under 50 minutes

Unsteady "Wind Pixel" Tunnel

Postdoctoral fellow, Queen's University

Developed a novel wind tunnel consisting of an array of "wind pixels" (high-performance server cooling fans). It can create extreme unsteady gust flows due to its low inertia, a flow condition that was otherwise impossible by using a traditional single-fan wind tunnel. In addition, the "wind pixels" provide the flexibility of digitally controlled spatial velocity profile generation.

Data assimilation for RANS CFD

Postdoctoral fellow, Queen's University

Creating a data assimilation algorithm with OpenFOAM that utilizes data from real-world surface-pressure measurements to improve the accuracy of RANS CFD simulations.

Aerodynamic load estimation from sparse surface-pressure sensors for aerial vechicles

Postdoctoral fellow, Queen's University

Developing a machine learning algorithm to estimate the aerodynamic forces of aerial vehicles using limited real-world data from sparse surface-pressure sensors. An automated wind tunnel apparatus was also created to gather a significant amount of experimental data necessary for neural network training.

Local flow state estimation from sparse surface-pressure sensors for surface vechicles

Postdoctoral fellow, Queen's University

Creating a machine learning algorithm to predict the local flow condition surrounding a surface vehicle with sparse surface-pressure sensors as part of the National Research Council Canada maritime autonomous surface ships program. While simultaneously conducting the laboratory investigation, an autonomous platform is also being developed to collect real-world data for neural network training.

Truck platooning opportunity assessment with sparse surface-pressure sensors

Postdoctoral fellow, Queen's University

Evaluating the feasibility of identifying and assessing truck platooning opportunities with a machine learning algorithm and sparse surface-pressure sensors. A mobile research platform was developed to collect real-world data on Ontario Highway 401 for data analysis and neural network training. The success of the pilot study led to a collaboration with the connected vehicle team of the National Research Council Canada to further develop the mobile research platform.