Many experts think fast-charging batteries will be critical for the adoption of electric vehicles. The main goal of the battery improvement process is to find the balance between the high charging speed and long battery lifetime. Artificial intelligence is accelerating this process. In order to know how a battery can be improved, current battery performance needs to be analysed. Battery performance data is collected by charging and discharging the batteries as fast as possible. Nowadays it might take months to run this process enough times to collect sufficient data, as there are a lot of possibilities for charging a battery and all these possibilities need to be tested in order to find the best one. New research made by Stanford University in 2019 managed to optimize a fast-charging protocol for a lithium-ion battery in less than a month; to achieve those same results without the aid of AI would usually take around two years.
In order to implement AI into the battery manufacturing process people need to know what data can be collected about the system and what information can be inferred from its analysis. Machine Learning approaches which become more and more popular require good quality data. One of the main issues for implementing AI in the battery-building process is the lack of data on various battery performance indicators. This sort of data has been very difficult to acquire because it’s not shared between researchers and companies. There’s a high level of secrecy or proprietary information. Also, the majority of the studies on the dynamic behaviour of battery materials are made in special labs and their number is limited. The explanation is that these studies require special expensive equipment and cannot be easily replicated. With the introduction of low-cost sensing techniques in the future, it is possible that the analysis of behaviour of battery material can also occur in real-world circumstances on a bigger scale. Currently, AI uses mostly the data on voltage, current and temperature. This might seem limited, but with the right analysis and post-processing, there are still some important electrochemical insights that can be inferred. If this data is combined with other variable types such as stress and strain, it can lead to even more powerful estimation. (Wu, Widanage, Yang, Liu, 2020) It is likely that in order to solve the issue with the lack of data in the future we will observe more collaboration between battery scientists from a dozen research institutions and companies for facilitating sharing stats across organizations. The University of Chicago has already developed a platform called Data Station that allows researchers to train machine learning models on a pool of information contributed by different groups without ever giving outsiders direct access to their data. (Oberhaus, 2020)
Other emerging Machine Learning techniques such as generative adversarial networks (GANs) can be an alternative solution for the lack of data problem. These techniques can generate synthetic data types to augment real data. Furthermore, this data gap could also be filled through the development of surrogate models.(Wu, Widanage, Yang, Liu, 2020)
From a material science point of view, people are working with AI to try to identify new chemical couples, be it electrolyte additives, or other compounds. That’s AI intersecting material science. AI can accelerate the rate of discovery of new couples or new compounds or stable chemistry because the classical chemistry approach takes a lot of time. For example, some experts are already arguing for usage of magnesium-ion batteries for electric vehicles instead of lithium-ion batteries. Lithium-ion batteries have been the best practical option so far; in theory, however, other materials have the potential to perform better. Scientists have investigated different metals to replace lithium in batteries, for example, sodium, potassium, aluminum, zinc, and calcium. Though several of these metals showed promise, magnesium came out as having many of the properties that would make for an attractive replacement for lithium. (Shah, Mittal, Matsil, 2021).
2021-3-28-1616945594