Artificial intelligence (AI) is here and is already making humans more effective in practically everything they do. In healthcare, AI diagnostic tools have improved patient outcomes and streamlined medical workflows. In finance, algorithms enhance risk management and fraud detection. In manufacturing, AI-powered automation and predictive maintenance are improving cost-effectiveness and efficiency.
For the battery industry, AI may be the key to accelerating the research and development necessary to replace flammable, toxic lithium-ion batteries with higher performing and sustainable technologies. New chemistries have typically required intensive testing in carefully controlled environments, leading to high costs for time, people, materials, and equipment. Many promising technologies fail to make it to market because they lack the funding for development. Properly harnessed, AI could be the force that changes the equation.
What Can AI Do for the Battery Industry?
Discover new materials
AI is able to filter potential materials based on their electrical and chemical properties faster than any human researcher can. While AI typically can’t churn out entirely new battery chemistries, it can narrow down candidates to enable researchers a much, much smaller pool to sift through.
A recent example of this was from researchers at Microsoft and Pacific Northwest National Laboratory, who used AI to accelerate their search for a new solid electrolyte. The AI driven process identified 23 promising materials from 32 million candidates in just 80 hours. The AI was key in filtering the materials based on their electrical and chemical properties to identify which ones can feasibly work together in a battery.
In another study, Carnegie Mellon researchers mixed various solvents, salts, and other chemicals together, then measured how the solution performed on critical battery benchmarks. A machine-learning system called Dragonfly then used the data to propose different chemical combinations that might work even better. In the end, the system produced six electrolyte solutions that outperformed a standard solution.
Focus research on desirable properties
AI can help develop batteries with specific characteristics, such as ability to withstand high temperature ranges or to exclude certain problematic materials. This opens up the possibility for nuanced battery technologies optimized for different applications. For example, EV battery developers may choose to prioritize electrode materials that enable quick-charging technology whereas battery energy storage system (BESS) researchers will seek out materials associated with long service life.
Run simulations using existing data
Battery development requires extensive trial and error before new iterations are made. Identifying areas of improvement can be a slow process, as researchers need to test their product, gather data, then conduct analysis. AI can enhance the accuracy and speed of simulations and models used in battery research, providing deeper insights into battery behavior and performance.
One early-prediction model helped reduce the time it took to identify fast-charging techniques from over 500 days to 16 days, in part by predicting the final cycle life using data from the first few cycles. Predictive analytics can guide the strategic direction of research efforts, focusing resources on the most promising areas.
Leverage industry research
Language models trained on textbooks and published research in battery chemistry can help answer questions about chemical properties or give suggestions for solving problems in the lab. To prevent inaccurate information, the model can even be configured to include references to published work in their responses.
This helps scientists make the most of existing industry research, so developers can learn from each other’s findings to advance battery technology. Rather than being hindered by the time constraints of human researchers scanning for articles, AI can instantaneously generate resources with potential answers.
Create virtual prototypes
Physical prototypes are expensive and time-consuming to produce. New technologies often require specialized equipment, and there are numerous steps to refine and prepare battery components before cell assembly even begins. When a prototype is made, it then needs to be tested, cycled, and stored for a period of time to see if defects arise.
In contrast, AI can create and test multiple virtual prototypes in a fraction of the time it would take to build physical prototypes. These simulation models help identify potential design flaws and performance issues earlier in the development process, saving researchers time and enabling them to work proactively on areas that need improvement. Virtual prototypes can undergo rigorous testing in simulated environments, ensuring that only the most promising designs proceed to physical testing.
Virtual prototypes also reduce costs because they do not require expensive physical materials and are less likely to use specialized equipment and labor, conserving materials that would otherwise go toward trial-and-error testing.
Artificial Intelligence, Real Innovations
To be clear, it is unlikely AI will solve all our R&D challenges or spit out the holy grail of battery technologies. But it is going to help humans pursue new materials faster and enable a new generation of battery chemistries tailored to desired characteristics and use cases.
By accelerating research, reducing costs, and digitizing testing, AI creates a more equal playing field for developers, so even companies with limited resources can work on building a better battery. Just as in other sectors, artificial intelligence is set to help the battery industry unlock a more advanced and resilient world.
