Presented by Accenture, this talk explores AI's transformative potential in the battery industry. The presentation examines the influence of AI on different industries and the opportunities it creates, discusses the adaptation of AI solutions to the battery industry, and highlights use cases along the battery lifecycle. Going beyond the current state of possibilities, future promises and limitations of battery AI are presented together with practical development recommendations for businesses.

AI's potential in the battery industry has fallen short, with issues like inaccurate recommender systems, poor lifetime predictions, and lengthy data entry processes. The key to unlocking better insights lies in clean, real-time data, which is critical for scaling AI in this space. Given the complexity of battery chemistries, supply chains, and production, standardized data collection is essential for effective AI in production environments.

Predictive and Generative AI are commonplace in battery research, but as Agentic AI emerges, are we prepared to embrace Battery Intelligence driven by AI Agents? In this contribution, we present two critical technologies we develop to enable agentic AI in battery research: Semantic Technologies, which embed data with context and meaning understandable by agents, and Explainable AI, which brings transparency to the decisions agents make autonomously. We argue that communication goes both ways: humans and agents must understand each other to ensure productive and reliable Battery Intelligence.

Physics-based simulations allow to make predictions about the thermo-electrochemical behaviour of Li-ion batteries and provide a tool for cause-and-effect analysis for experimental observations. In this talk we are going to discuss how such models can be used to perform digital performance analysis of Li-ion batteries. Due to their physical nature, the models allow to conduct studies under exactly controlled conditions and potentially extend existing experimental data sets.

Recently, Large Language Models (LLMs) have seen unprecedented growth, marked by significant advancements in foundation models, generative AI, reasoning capabilities, and intelligent agents. This progress has enabled innovations in battery conditioning monitoring, diagnostic and optimised control, enabling 1) NVIDIA GPU AI acceleration of complex physical models using Physics-Informed Neural Networks (PINN) 2) online optimisation of ultra-fast charging and V2G curren ts to minimize additional degradation effectively. The presentation will demonstrate spatial-distribution of temperature and electrodes concentration AI accelerated prediction for 4680 cylindrical cells with a discussion on possible BMS implementation in order to improve the safety and sustainability of future battery systems.

The project BATMAX sets out to pave the way for advanced nex- generation data-based and adaptable battery management systems capable of fulfilling the needs and requirements of various mobile and stationary applications and use cases. Given the role of battery systems as a key enabling technology within the green shift in transport, mobility and energy, it is evident that multiple combinations of requirements, use cases, duties, and businesses are placed upon battery systems.

This study presents an advanced strategy for State of Charge (SOC) and State of Health (SOH) estimation that has achieved errors of less than 1%. This strategy includes combined spectral and temporal characterisation of cells. It uses the Smooth Variable Structure Filter together with the Interacting Multiple Model concept for estimation.

Battery characterisation and aging tests span several months to years, posing significant challenges for manufacturers and OEMs seeking to accelerate testing and extract comprehensive insights, particularly on battery aging. This work addresses these challenges by integrating physical modelling with machine learning to analyze battery performance at the parameter level. Leveraging robotics and high-throughput testing platforms, we develop a framework that digitises and automates the testing process, enabling faster battery evaluation.

In this presenstation we introduce innovative AI-based approaches to enhance testing efficiency, including machine learning for anomaly detection and predictive models, which significantly reduce testing times and the number of required test specimens.

Quality control in electrode manufacturing is essential for producing high-performance batteries used in electric vehicles, renewable energy storage, and consumer electronics. This research explores advanced sensor technologies for monitoring wet film thickness and homogeneity. The study also develops sensors for inline defect detection, using AI-based neural networks for classification. The goal is to optimise coating parameters, enhance defect detection, and integrate adaptive control systems, advancing battery production technologies and performance.

The self-heating, the first phase of thermal events in batteries, is caused by an intricate balance of exothermic and endothermic reactions. Deciphering their interplay allows to engineer solutions to decrease the risk of a thermal event. The talk presents a model-based analysis of the impact of production conditions and cell state. High temperature operando electrochemical mass spectrometry reveals the degradation processes leading to self-heating and allows model-parameterisation.