The success of electric vehicles relies heavily on battery technology. Key performance attributes like range, acceleration, and sustainability are linked to battery packs. Therefore, battery testing ensures adherence to EVs’ performance attributes and reliability, safety, and longevity. However, comprehensive testing of battery packs takes more than 12 months to complete, and challenges are multiplied by the continuous evolution of battery chemistry.
Current Scenarios
During the 1990s and 2000s, testing cells and batteries posed a significant challenge. To execute standardised tests, new equipment and systems had to be developed, which increased the cost (CAPEX and OPEX) for OEMs and cell suppliers. From 2010 onwards, as the evolution in lithium-ion chemistries progressed, cell testing became a little more complicated due to the same testing standard applied to different chemistries. Lithium iron phosphate has been the mainstay solution due to its stability, safety, and lower proneness to thermal runaway. However, it does not help us achieve our extended range and rapid charging aspirations. The NMC chemistry, while helpful in realising those aspirations, is comparatively unsafe due to the presence of cobalt and higher chances of thermal runaway. Advancements in mathematical modelling have enabled the virtualization of module- and battery-pack-level tests based on cell-level test results. CAE toolchains have reduced the need for physical tests for structural integrity, reducing costs and time-to-market. This creates an opportunity to focus on testing cells for safety.
Never has an OEM in the world intentionally developed an unsafe product for its customers. However, high-profile incidences of EV batteries catching fire created an urgency to address safety. While robust battery and thermal management systems are in place, thermal runaway (heat generated by internal reactions being higher than the heat that can be dissipated or taken away), a characteristic of most chemistries, is the centre of attention to address safety due to a gamut of factors creating it, from materials used in the cell to the design of the cell to production processes. Aspirations of rapid charging to enable the higher adaptation of EVs have created an opportunity to focus on thermal runaway rigorously. Therefore, the entire scientific community and test infrastructure are currently focused on testing the cells for thermal runaway through accelerated ageing and abuse approaches. Simulation tools that can predict thermal runaway precisely are yet to mature and be accessible.
Another aspect where the scientific and engineering communities are engaged extensively is to address range anxiety, which necessitates the batteries to perform consistently under diverse operating conditions, for extended distances, and for a considerably long time. Range anxiety is addressed through a robust battery and energy management system, which needs extensive testing of the cells and batteries under diverse test conditions first to understand the cell’s behaviour under such conditions and the degradation in the cell’s performance over a longer period of time. Cycle Life Tests generate all the essential data of capacity loss, retention and changes in cell impedance required for the battery and energy management systems. In addition, integrated test systems like Battery in Loop validate the overall energy storage and management system. And yet, the test coverage does not assure the uniformity of the system’s performance, as the cost and effort for such extensive coverage are enormous.
Data analytics is another challenge in battery testing, with the enormous amount of data collected during the testing calling for extensive resources and efforts. While the engineering and scientific community know what data points are to be extracted and for what purpose, the sheer number of data points collected from various tests and their quantity pose a significant challenge. Even with the automation competency provided by toolchains, it is estimated that more than 60% of human efforts are spent on data analytics.
Emerging Strategies
Gigafactories are popping up worldwide, leading to considerable investments in cell and battery testing. New approaches to validate cells are emerging, accelerating current practices or reducing physical testing significantly.
With the help of artificial intelligence (AI), we can create predictive models of cells that assist in diagnostics and prognostics. AI can automate various aspects of the testing process, accelerating the process. Optimisation algorithms powered by AI can streamline battery testing procedures by identifying the most informative test conditions and parameters. This can reduce testing time and costs while maximising the amount of relevant data collected. Moreover, AI algorithms can analyse large volumes of data collected during battery testing more efficiently and accurately than manual methods. This analysis can identify patterns, anomalies, and trends in battery performance, helping researchers understand battery behaviour under different conditions.
Generative AI has a crucial role in reducing the need for certain types of tests on cells. In addition to abundant data, generative AI can create synthetic data resembling real-world data. This synthetic data can be used to train machine learning algorithms, making them more robust and capable of generalising at much lower costs and efforts. When implemented with laboratory management systems, generative AI (or even AI) can detect anomalies during the testing process in almost real-time. It can quickly detect changes in the test data about a standard dataset. Furthermore, generative AI can mitigate risks such as thermal runaway by generating the data required to simulate extreme conditions. This data can be used in high-fidelity simulation models, increasing testing coverage significantly and at dramatically lower costs than physical testing.
Collaborative Approach to Advancement
Comprehensive battery testing is crucial as transportation electrification accelerates. Collaboration and a holistic approach to innovation can drive transformative change towards a cleaner, sustainable future powered by electric vehicles. Engineering research and development service providers invest in AI and generative AI to accelerate product development for electrification goals. They are coming to the table as significant stakeholders in the business. Therefore, a collaborative approach among all such stakeholders can save a lot of capital and get much safer products to market more quickly. Government and regulatory bodies need to understand the pain areas of different stakeholders and act as enablers for such collaboration.
