ENGINEERING PROJECTS
Airship for atmospheric pollution monitoring
Motivation: Living next to a construction site is extremely painful: noisy and very polluting with all kinds of particulate matter being thrown around. Many children and elderly in my housing society started complaining of chest congestion and breathing issues. Efforts to talk to the construction site team yielded no results. To understand the nature and intensity of the pollution, I designed and built an atmospheric pollution monitoring airship.
Technical details:
- Shape was optimised using fluid flow analysis on Ansys
- Balsa wood internal structure for light weight and stability
- Mylar casing with air outlet
- 3D printed propulsion system
- DHT 11, MQ7, MQ135 sensors for pollution monitoring
- 3.7V 1200 mAH LiPo battery to give 5 hours of flying time
I am currently using the prototype to collect data on the pollutants at different altitudes near my housing society before sharing it with the local municipal corporation for further action.
Recognition: Research paper on how to optimise the shape of the airship received the British CREST Gold Award. It was presented at FLAME 2024 by Amity University and will be published in Springer’s ‘Advances in Engineering Design’ journal.
Centrifuge to simulate Martian gravity
Motivation: Testing equipment designed for Martian conditions is difficult. I am working on designing a centrifuge to simulate Martian gravity on Earth.
Mentored by: Prof. Sandeep Saha from the Aerospace Engineering department at IIT Kharagpur.
Technical details:
- Evaluated different gravity simulation options and zeroed in on a centrifuge
- CAD designs on OnShape + Ansys Modeller
- Ansys structural and fluid flow simulations for deformation, stress, vibrations, natural frequencies, etc
- Used rigid body mechanics concepts to calculate centrifuge RPMs for Martian gravity
I also verified Ansys results by hand calculations using Physics concepts beyond my school curriculum.
NASA Lunar Loo
Motivation: In 2020, when the pandemic hit, NASA announced a special challenge to design a toilet for use on the Moon – a “Lunar Loo” 💩. With a couple of friends, I designed a toilet that could be used in low gravity conditions on the Moon.
Technical Details:
- Designed on Fusion 360
- Detachable funnels for urine collection
- Rocking bowl for fecal matter
- Automatic wet wipe cleaning system
- Automatic seat cleaning
- Hand and foot restraints
While we did not win, it was a project that helped me apply my CAD skills to something with real-life application!
Denoising Grayscale Images
Motivation: Eager to build on my signal processing work at Sisir Radar, I worked on image denoising algorithms.
Mentored By: Prof. Maryam Yashtini from the Mathematics Department at Georgetown University.
Technical details:
- Learned introductory multivariable calculus and partial derivates to computationally minimise the value of a multi-dimensional function
- Minimised variation between neighbouring pixels to eliminate Gaussian noise, while ensuring outlines were preserved
- Tuned algorithm on images from different sources including those taken from my mobile phone.
I enjoyed pushing myself to learn advanced calculus concepts. The application of math to something so commonplace fascinated me!
Prof. Yashtini’s evaluation. She placed me in the top 5% of students she has worked with.
Asteroid Deflection Methodology
Motivation: I led our school delegation for the IB League STEM Challenge hosted at JPIS, Jaipur. We attempted to create a solution to deflect a fictional Earth-bound asteroid. Within 5 hours, we researched, brainstormed, and make a presentation deck.
Technical Details:
- Numerous solutions considered e.g. nuclear explosion, gravitational deflection, etc before deciding on swarm of ion beam emitting spacecraft because of safety and feasibility
- Xenon gas ionised and accelerated outwards at high speed
- Thrusters on both sides of spacecraft to keep constant position
- Coordinated swarm of spacecraft for maximum impact
- Backup plan – hit the asteroid with the spacecraft to deflect
Recognition: After our presentation, the host school principal and one of the judges (a Princeton aerospace engineering graduate) personally came up to me to commend the thinking and detail behind our presentation! We won the Most Innovative Presentation award.
BOOM - Breathing Oxygen on Mars
Motivation: Being forever enamoured by human life on Mars, I tried solving the problem of oxygen generation on Mars. After months of reading research, evaluating existing solutions, and meeting experts, I managed to create BOOM, an industrial-scale system to solve this problem. It was my first taste of writing a paper too!
Technical Details:
- Modified Electrolysis of Water
- Utilises CO2 abundant in Mars’ atmosphere to recycle by-product hydrogen in the Sabatier Reaction to reproduce water.
- Solar panels for sustainable power
- Aluminium oxide coating on electrodes for faster electrolysis
- Use of nickel as a catalyst for hydrogenation of CO2
Through BOOM, I learnt the multidisciplinary nature of engineering!
Recognition: Silver Medal at National Science Concours, 2019
Indian Stock Exchange Simulator
Motivation: Coming from a family of business professionals, I had heard a lot about the stock market growing up. And having just completed HarvardX’s Web Programming course, I decided to build an Indian stock exchange simulator using Django.
Technical Details:
- Authentication system to create accounts for new users
- Database connection to store stocks bought and portfolio value using Django models
- Yahoo Finance API calls using JavaScript to pull live stock prices
- Leaderboard of active players
- Django template language for dynamic UI
It was challenging to implement some of the features as I did not have any experience. Yet, I had a lot of fun working on this project!
Diseased Cotton Leaf Detection
Motivation: To apply my learnings from UC SanDiegoX’s Python for Data Science course, I built a tool to detect disease in cotton plants. This would be very helpful for cotton farmers.
Technical details:
- Numpy and Pandas ingest data, simplifying the 3D image to 2D and then to 1D Histogram-Oriented Gradients
- Tested 12 ML classification algorithms using scikit-learn library.
- Timed each algorithm and calculated accuracy scores
- Selected the best ML classifier for the final algorithm
I submitted this project to Microsoft Imagine Cup as well. Even though it didn’t win, it was a fascinating introduction to ML and image processing. Code Explanation.