Battery Safety and Engineering

A data-driven framework is introduced for predicting the energy yield and classifying the trigger mechanism of thermal runaway in lithium-ion batteries. The core of this approach is a 2D uniform data augmentation method that corrects for severe data imbalances across multiple feature dimensions simultaneously. A stacked ensemble regressor trained on this balanced data demonstrates high predictive accuracy across all failure modes, providing a valuable tool for designing safer battery systems.

Emergency discharge has been considered as an effective strategy to mitigate thermal runaway by reducing the state of charge (SOC), whereas substantial ambiguities remain due to the wide range of discharge rates beyond normal operation, complex coupling with electrochemistry, and competing effects from external heating. This talk focuses on recent experimental investigations of thermal runaway mitigation in a wide range of discharge and heating conditions.

Cell defects can trigger thermal runaway (TR), presenting significant safety risks for lithium-ion batteries. This presentation examines how metallized polymer current collectors can reduce these risks by limiting the energy released during internal short circuits. It will explore the mechanisms behind the reported tenfold reduction in thermal runaway risk and discuss the implications for battery safety, cell design, and safer large-scale energy storage systems.

This talk presents a spatially resolved single-particle model that captures variations in solid stoichiometry, reaction current, and overpotential across electrode thickness while retaining low complexity. This formulation adds only two additional states per electrode. Simulation results show that our “SPMe+” reduces voltage prediction error by 65 to 90% (depending on profile) compared with a conventional SPMe, making it a valuable contribution to next-generation battery controls.

Battery electric trucks place high demands on the safety and durability of lithium-ion batteries, which requires a detailed understanding of the electrochemical and mechanical processes inside the cell. Sensors integrated into the cell provide direct access to internal states that cannot be derived from terminal voltage, current, and temperature alone. Reference electrodes allow the anode potential to be monitored during charging, so that conditions favoring lithium plating can be identified before they occur, while swelling sensors capture the additional and irreversible cell expansion caused by plated lithium. Together, both measurements allow safe operating limits to be determined more precisely and provide a basis for safer charging and operating strategies for lithium-ion batteries in heavy-duty applications.

As electric vehicle (EV) fleets and heavy-industrial battery deployments scale, managing battery reliability, warranty exposure, and safety risks requires moving beyond isolated anomaly detection toward systematic, explainable root-cause diagnosis. We demonstrate how an evidence-led diagnostic framework evaluates sensor coverage, signal persistence, and conflicting explanations to deliver fully traceable diagnostic outputs. By mapping inferred physical states to targeted mitigation strategies, the approach enables operators and automotive OEMs to prioritize high-risk assets, prevent catastrophic field failures, and optimize warranty and second-life decisions.

With the breadth of applications for batteries expanding to demanding sectors like electric aviation, data centers, and behind the meter storage, alternative chemistries to conventional Li-ion cells are being adopted. These include alkali metal anodes, solid-state electrolytes, sodium-ion, potassium-ion, and more. The safety of next-generation chemistries is still not well understood. This talk will cover a comprehensive safety evaluation and comparative safety of next-generation chemistries.

Evaluating the relative safety of new battery chemistries can be a challenge, particularly when TRL is at the laboratory scale. Ideally, battery safety should be evaluated using a variety of quantitative methodologies. This presentation will compare the safety performance of commercially available lithium-ion and sodium-ion cells using multiple methodologies with the goal of providing a framework for “criteria-based” safety assessments that consider multiple facets of battery safety.

Sodium-ion cells have entered the commercial market but there is not much data available in the literature, on either the performance or safety of cells and batteries of this chemistry. At UL Research Institutes, commercial and custom-designed sodium-ion cells were tested for performance and safety. Tests were carried out on single cells, as well as 12V and 48V modules. 

As sodium-ion batteries gain attention as an alternative to lithium-ion technology, understanding their thermal safety characteristics is essential for reliable deployment. This presentation examines the thermochemical behavior of sodium-ion battery chemistries, including heat generation, thermal stability, and potential failure mechanisms. The findings will help inform safety assessment, cell design, and risk mitigation strategies for sodium-ion batteries in energy storage and other applications.

This presentation examines heat release and gas evolution in potassium-ion batteries using graphite anodes, Prussian white cathodes, and a low-flammability phosphate electrolyte. Results identify anode reactions below 100 °C and toxic gas generation from charged cathodes during heating. The findings show why low flammability alone cannot establish battery safety and provide a basis for improving materials screening, identifying gases to monitor, and designing thermal tests for this emerging chemistry.

Battery safety is often viewed as a chemistry, cell, or hardware challenge, yet many safety issues stem from system-level factors such as requirements, interfaces, diagnostics, validation gaps, and overlooked use cases. Effective safety requires a systems engineering approach across the product lifecycle, combining prevention, mitigation, validation, and fleet learning. Hardware is essential, but battery safety is ultimately achieved through integrated risk management across the battery, vehicle, and external factors.

Battery compliance programs usually pass discipline by discipline—classification, transport, storage, response, retirement—and still fail at the seams between them. This session walks the six points in a battery's handling life where compliant programs typically break, from classification through end of life, layering in how condition (new, in-service, damaged, or retired) changes the answer at each one. The goal is diagnostic: where does your program actually break?