Aeroil: An Oil Cleanup System Using Aerogel and a Deep Learning Prediction Model
In collaboration with Naila Moloo, Sophia Moloo, Marko Youngson, and Andrew Krechmer
In a recent hackathon, my team and I were given the goal of solving an important problem using the intersection of nanotechnology and artificial intelligence, where we placed in the top 5 projects. We were particularly interested in oil, and its detrimental impacts on the environment. Current alternatives for cleanups like oil skimming, booms, sorbents, dispersants, and bioremediation are generally inefficient, not reliable, or toxic.
We came up with Aeroil, a business idea that utilizes the oleophilic properties of aerogel to be a cleaning agent for oil spills. There have been tons of research papers on this, and aerogel’s capability to give us the solution to save millions of habitats, wildlife, and humankind, but there has been no way to distribute these aerogels effectively, and do so in an efficient and thought-out manner. That is precisely what Aeroil does.
How Aeroil Works — High-Level Overview
When a spill occurs, Aeroil will be informed, and bundles of aerogel are distributed to the location with helicopters. An AI prediction model is used to locate the aerogel after it has absorbed the oil, at which point a helicopter collects it from the ocean. This infographic demonstrates the four steps of the Aeroil process.
Aeroil is simplistic and easy to use, with an intuitive application for the geographic tracking of oil spills and aerogels. When a spill occurs, anyone can easily log onto the app and alert the company to drop off the aerogel. Further details are also to be provided, for calculating specifics, like how much aerogel is needed. After Aeroil is alerted, they send helicopters to release the aerogels into the oil spill area. Each step of the process is outlined in the Aeroil application, promoting transparency and understanding.
Aerogel is an ultralight material that can be used for absorbing oil at a rapid rate. It is a synthetic porous derived from a gel that is hydrophobic and oleophilic, meaning it has a stronger affinity for oils rather than water. It is the lightest material in the world, made up of around 99.8% of gas-filled nanopores, and is exceptionally absorbent, allowing the absorbency of 4 grams of oil per gram of aerogel at a rate of 1 gram per 3 minutes. Aerogel works like a super-efficient sponge. Sponges absorb water because they’re made of loose fibres with a ton of space between them, so the holes between the sponge fibres absorb the water. It’s the same thing with aerogel; because of all the space, it’s a really good absorber.
Locating the Aerogel with an ANN
Aeroil uses deep learning to predict the coordinates of where the aerogel will be when all the oil has been picked up. The model can be broken down into two parts. The first part uses a simple perceptron that takes inputs — the type of oil, the amount of oil, and the rate at which aerogel can absorb oil — and outputs the amount of time it will take to absorb all the oil.
The output of the first model is passed as the input to the second model, which utilizes a deep artificial neural network to predict where the aerogel will be located once it finishes absorbing the oil.
The data from our AI system is presented in real-time through the Aeroil application. The emergency cleanup team can access constantly updated information about where the oil spill is, and when it will be absorbed.
Our futures are at risk; the global climate crisis, and our insatiable need for oil, is killing us. With Aeroil, oil spill crises can be prevented, ocean waters kept clean, and millions of lives saved. If you have any questions, reach out to the team here!
Kimberly — firstname.lastname@example.org
Naila — email@example.com
Sophia — firstname.lastname@example.org
Marko — email@example.com
Andrew — firstname.lastname@example.org