Emerging Technologies

Fully charged: how AI-powered battery testing can support the EV boom

An EV battery factory in South Korea.

An EV battery factory in South Korea. Image: Reuters/Kim Hong-Ji

Richard Ahlfeld
Chief Executive Officer, Monolith AI
  • A new automotive industry survey reveals widespread dissatisfaction with EV battery testing, a problem that could be solved by AI.
  • AI can accelerate battery validation by trialling different use cases faster than physical tests.
  • Thoughtfully designed AI will surmount the 'trust gap' the technology currently faces.

The automotive industry faces immense pressure in the fight to achieve net zero. Engineers must overcome incredibly complex – intractable – physics underpinning battery technology to deliver competitive, sustainable electric vehicles (EVs).

Cost-effective validation and optimization of EV batteries have become pivotal. Industry leaders consistently tell us that their current toolset, including physical simulation, falls short. To better understand these challenges, we commissioned a study with Forrester Consulting titled 2024: AI for EV Battery Validation with 165 senior decision-makers in automotive engineering across North America and Europe. The findings proved illuminating and, frankly, unsurprising.

The study revealed widespread dissatisfaction with existing validation tools. Over 60% of respondents stated that their current methods, dominated by physical testing and simulations, cannot meet the rigorous demands of EV battery development. These shortcomings stand out starkly when the goal is to accelerate EV market entry while ensuring safety, reliability and sustainability. Engineers must embrace a new approach that alleviates these bottlenecks without compromising trust in the process.

Have you read?

At Monolith, we believe AI and machine learning (ML) offer the solution. However, introducing AI into established industry workflows presents its challenges, with trust being paramount. Our conversations with engineers reveal that scepticism often stems from a fundamental misunderstanding of AI's workings. Many engineers express concerns about “black box” systems that obscure the reasoning behind their outputs. How can engineers who rely on tangible physical tests and proven simulations trust conclusions from algorithms they cannot fully explain?

This trust gap, while significant, is not insurmountable. Thoughtfully designed AI enhances transparency rather than diminishes it. Machine learning models, particularly those with explainable AI capabilities, provide clear traceability of their inputs and decisions. Engineers can immediately see the data patterns and factors that inform an algorithm's recommendations, saving time compared to piecing together insights from colleagues. Most importantly, AI reinforces engineering expertise rather than replacing it, reducing time spent on repetitive tasks and freeing talent for creative problem-solving.

The intractable physics of EV battery validation remains the core challenge. Unlike most automotive components, batteries resist precise modelling due to highly non-linear and application-specific interactions between chemical, thermal and electrical processes. Engineers must meticulously test batteries under specific use-case scenarios to validate performance – a labour-intensive, time-consuming and expensive process. The industry clearly needs a new validation approach, which machine learning provides.

We see important parallels between early simulation tools and current AI adoption. Computational fluid dynamics (CFD) met similar resistance when it entered engineering workflows 30 years ago. The scepticism stemmed from the same issue: unfamiliarity. Today, such tools have become indispensable. AI is undergoing the same transformation. Innovation now requires a mindset shift: recognizing AI not as a threat, but as engineering methodology's natural evolution.

AI offers profound benefits in EV battery development. Self-learning algorithms accelerate product validation, optimize battery designs and select optimal materials based on thousands of variables that no human could analyze simultaneously. These algorithms simulate scenarios faster than traditional physics-based models, offering engineers new insights into potential failure modes and performance improvements. Instead of testing battery cells under every possible condition, engineers can use AI to predict behaviour, reducing necessary physical tests. This approach saves time, conserves resources, reduces waste and cuts costs.

The past few years have laid crucial groundwork for this transformation. Leading battery manufacturers have invested heavily in AI capabilities, recognizing the competitive advantage that accelerated development cycles provide. Industry giants like Samsung SDI and CATL have begun integrating machine learning methods into their development processes, significantly reducing their time-to-market while maintaining rigorous quality standards. Their success has sparked a wave of adoption across the industry, with research laboratories and testing facilities following suit.

Testing laboratories, in particular, have emerged as early adopters of AI technology. These facilities face mounting pressure to accelerate product development, while maintaining accuracy and reliability. AI helps them achieve both goals by optimizing test sequences and identifying potential failure modes early in the development cycle. The increasing volume of published research on AI applications in battery testing demonstrates the industry's shifting mindset. Each month brings new papers and case studies showcasing successful implementations and methodological improvements.

We have witnessed this transformation first hand through our work with automotive industry partners. Their growing reliance on machine learning to tackle previously insurmountable challenges produces consistently impressive results. Whether shortening validation timelines by months or uncovering overlooked design insights, AI delivers tangible value. The industry's large-scale adoption of these tools is now not a question of if, but when.

Yet, adoption requires more than deploying technology. Building trust in AI demands collaboration between developers and engineers. We must demystify machine learning and help engineers confidently incorporate it into their decision-making processes. AI enhances human expertise, enabling engineers to focus on innovation rather than repetition.

The automotive industry faces an exciting yet challenging road ahead. While the transition to sustainable mobility takes time, AI accelerates the journey. Engineers with powerful tools to decode EV battery physics complexities drive this transition. Electric vehicles' future depends on their ability to innovate; AI, when trusted and understood, becomes their most powerful ally. Together, we can solve the intractable and build a cleaner, more sustainable automotive future – faster.

Don't miss any update on this topic

Create a free account and access your personalized content collection with our latest publications and analyses.

Sign up for free

License and Republishing

World Economic Forum articles may be republished in accordance with the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International Public License, and in accordance with our Terms of Use.

The views expressed in this article are those of the author alone and not the World Economic Forum.

Stay up to date:

Artificial Intelligence

Related topics:
Emerging TechnologiesEnergy Transition
Share:
The Big Picture
Explore and monitor how Artificial Intelligence is affecting economies, industries and global issues
World Economic Forum logo

Forum Stories newsletter

Bringing you weekly curated insights and analysis on the global issues that matter.

Subscribe today

Why Generative AI leaders must blend thinking, building and creating value

Chetan Kapoor

December 19, 2024

How investing in connectivity and digital infrastructure can be a catalyst for inclusion and empowering people

About us

Engage with us

  • Sign in
  • Partner with us
  • Become a member
  • Sign up for our press releases
  • Subscribe to our newsletters
  • Contact us

Quick links

Language editions

Privacy Policy & Terms of Service

Sitemap

© 2024 World Economic Forum