Connect with us

Energy

AI Models Used to Find Deposits of Battery Materials and Identify Replacements

mm

Published

 on

AI researchers are in the process of developing AI models to reduce the environmental impacts associated with the extraction of materials used in batteries. The mining exploration startup Kobold is developing an AI model capable of detecting materials used in the creation of batteries in the ground. Meanwhile, a team of researchers from IBM is utilizing AI algorithms to determine which alternative materials could be used to create batteries.

The demand for materials to create batteries is increasing all the time as more and more objects become powered by electricity. To meet this increased demand, more mining will have to take place and researchers are looking for ways to reduce the environmental impact of these resource extraction operations. AI has the potential to improve current methods of extracting ore or even replace these methods with techniques that are more sustainable.

According to IEEE Spectrum, KoBold Metals’ new AI project aims to detect ore deposits in areas where extracting the ore would do relatively minor damage, compared to current resource extraction methods. Kobold explained the AI models they are developing could dramatically reduce the need for invasive, expensive mineral exploration missions, which typically require many explorations and scans to find rare materials. According to KoBold, most of the easily accessed materials have already been found, even though new mineral deposits will be required to change the current energy system.

KoBold is working alongside Stanford’s Center for Earth Resource Forecasting to develop an AI agent that can make recommendations for where to find certain minerals. The startup wants an AI capable of recommending areas that may contain deposits of lithium copper, cobalt, nickel, and other minerals.

A professor of geological sciences at Stanford, Jef Caers, explained that the concept behind the AI is that it will help geologists evaluate multiple sites for potential mineral deposits and expedite the decision-making process. According to Caers, the AI model operates like a self-driving car in the sense that the model both gathers and acts on data collected from the surrounding environment.

As society transitions away from fossil-fueled powered cars to battery-powered cars, aiming to reduce overall greenhouse gas emissions, more battery capacity will be needed. According to a paper published in the journal Nature this past December, there could be over 2 billion electric vehicles on the road by 2050, requiring around 12 terawatt-hours of annual battery capacity, which is approximately ten times the current existing capacity of the US.

Kobold’s AI-driven mineral discovery approach is supported by a data platform that stores information on potential mining sites taken from a variety of sources. Soil samples, drilling reports, and satellite imagery are collected and used as features for the AI model, which makes predictions about the locations of highly-concentrated ore deposits. It’s hoped that the AI model will make accurate predictions about which sites should be mined, the predictions coming much faster than those made by a human analyst.

While Kobold is designing AI models to find more minerals for batteries, researchers from IBM are trying to find materials that can substitute for common battery ingredients like lithium and cobalt. The IBM researchers are employing AI models to identify solvents that could outperform current lithium-ion batteries. This IBM AI project is focusing on currently existing and currently available materials, but a different IBM project aims to synthesize new molecules that can replace common battery materials.

The IBM research team employed generative models to understand the molecular structure, melting point, viscosity, and other attributes of existing materials. Training a generative model on these types of features allows the researchers to generate molecules with similar properties.

IBM has already used their AI system to engineer new molecules dubbed “photoacid generators”. These photoacid generators could assist engineers in developing computer chips using more environmentally friendly materials and techniques. The IBM research team aims to do the same for battery technology.

Blogger and programmer with specialties in Machine Learning and Deep Learning topics. Daniel hopes to help others use the power of AI for social good.