Electric Vehicles: The Perilous State of Battery Safety
This story is contributed by Sam Kanakamedala and Dania Ghantous
- EV OEMs are rethinking lithium-ion battery safety and learning from historical failures in smartphones and laptop computers.
- The EV industry must invest in software intelligence to predict, prevent, and alert the user to unsafe batteries before they lead to fire.
- Achieving zero EV battery fires in the field is not a technology limitation anymore. It is a business priority.
When you read these recent press releases about General Motors and Hyundai recalls, the first question you ask is, “Didn’t they test these vehicles for safety?”
GM recalls 68,000 Chevrolet Bolt vehicles for the potential of an unattended fire in the high-voltage battery pack
· Hyundai recalls 82,000 Kona EVs due to the risk of short circuits possibly caused by faulty manufacturing of its high-voltage battery cells
Both vehicles received a 5-star NHTSA overall safety rating. So, why did their batteries catch fire?
Electric Vehicles (EVs) are very different from Internal Combustion Engine (ICEs) Vehicles. They are essentially big battery packs on wheels. Hundreds of battery cells are linked together into a package to store a considerable amount of energy up to 100 kWh or even more. Each cell is susceptible to failures that can lead to a fire that spreads to adjacent cells resulting in significant damage or loss of life. Furthermore, each cell is unique — even when cells are made on the same day and on the same equipment, there will be variances. Therefore, testing a few cells during a controlled safety certification process is insufficient to guarantee overall safety.
This article discusses key causes of battery fires and how to predict and prevent them.
Modern Applications Are Stressing the Batteries
For broader adoption, EVs must deliver a driver experience similar or better than conventional vehicles: >300 miles with a full tank; less than 20-minutes to fill the tank; >100,000-mile warranty, and >500,000-mile for commercial fleets; utmost safety; and affordability. These increasing demands for extended range, shorter charge times, safety, reliability, and longevity can add significant challenges to the current battery capabilities.
Vehicle range is a product differentiation. The range is a function of the total energy capacity of the battery. To pack more energy without a proportionate increase in size or weight, the energy density, defined as the amount of electrical energy per unit volume (or per unit weight) that a rechargeable battery can store, must increase.
There is a natural design tension between increasing energy density, fast charging, and safety. Increasing energy density requires a more tightly packed battery structure and may precipitate safety risks when subjected to fast charging. Over several charge and discharge cycles, dendrites form and may create internal shorts resulting in catastrophic failure. Extreme operating conditions can further accelerate this process.
The rapid adoption and broader usage of EVs may lead the battery industry to make more tradeoffs between reliability, performance, and safety.
Battery Defects Are One Source of Potential Safety Hazards
Battery failures occur when latent problems cause the battery to swell, lose its ability to hold a charge, or, worst-case, ignite.
Hidden defects are of particular concern. For example, a minute defect may lead to the formation of dendrites between the electrodes. A separator failure may also lead to an electric short. Furthermore, mismatched anode/cathode capacity ratios as well as uneven compression in the cell structure can introduce sources of defects that are difficult to detect early on. These types of defects manifest themselves much later during the life of the battery and can lead to safety risks. In 2016, minute latent defects in the battery led to the expensive recall of the Samsung Note 7 smartphone.
EV batteries undergo tremendous stress during their long life. The two leading stressors are fast charging and extreme temperature. The battery’s capacity to accept charge varies significantly with temperature. Traditional battery management systems do not adjust the degree and rate of charge based on health. Temperature gradients across the pack cause uneven degradation, especially during fast charging. Therefore, any one of the hundreds of cells in the battery pack can develop an internal short and ignite the whole pack.
A more intelligent approach would be to control the degree and rate of charge in real-time based on detailed diagnostics of the cells and the pack.
Automotive Safety Standards and Regulations Are Catching up to EVs
The automotive safety standards mainly focus on design and operational safety. The standard tests simulate real road conditions to certify the reliability and performance of equipment in the vehicle. These tests focus on passive safety equipment like thermal barriers that protect passengers when accidents occur. But passive safety is not sufficient to keep EVs safe. EVs require an active safety system besides thermal barriers (passive systems) to prevent battery failures and achieve zero battery fires.
ISO 26262 is the international standard for functional safety of electrical and electronic systems in vehicles. This standard covers only the hardware and software malfunction of the battery management system that causes failure of preventing excess temperature, charge, current, and voltage leading to a battery fire. The non-electric and electronic system malfunctions such as chemical or mechanical defects in batteries are out of these standards’ scope.
While ISO 26262 covers automotive safety, a dedicated EV safety regulation body was formed in 2012 under the United Nations with China, Japan, the European Union, and the United States co-sponsoring two working groups to address EV environmental and safety issues. The EV Safety (EVS) working group specifies the vehicle’s in-use and post-crash safety. Vehicles are required to warn of a hazardous situation inside the passenger compartment that will allow egress within 5 minutes. Furthermore, vehicles remain safe and do not catch fire or explode for up to one hour after a crash. These requirements align with the China GB standards and address occupant safety for thermal events leading to fire, explosion or smoke. Its latest regulation (GTR) focuses on future standards for active safety systems that use software intelligence to perform prognostics and diagnostics to prevent battery fires.
There are two complementary ways to improve battery safety for safer EVs:
1) Develop safer cell chemistries. Considerable research and investments are pursuing new technologies.
2) Deploy software intelligence to predict the rare presence of defects and ensure the batteries operate safely.
We believe that intelligent software with predictive capabilities is essential to provide additional safety safeguards. Moreover, the technology is available today and can deliver the required safety at a fraction of the cost.
Computation Meets Chemistry: Intelligent Battery Management Software
Intelligent Battery Management Software must offer three layers of protection from battery hazards:
1) Mitigate conditions that could cause battery failures;
2) Predict battery defects;
3) Provide early warnings for possible safety hazards;
The first step is to complete detailed battery diagnostics of batteries in the vehicle. Electrochemical Impedance Spectroscopy (EIS) is a widely known and well-established diagnostic method used in laboratories worldwide to measure and separate the contribution of the different electrochemical processes taking place as the cell ages. These processes occur at different frequencies and can be isolated to determine which electrochemical process dominates the degradation. The EIS spectrum can be implemented in electric vehicles to diagnose the unique signatures of the battery’s underlying electrochemical processes and generate a chemical state of health in real-time. This can be implemented in software and should run a continuous assessment of the impact of the operating conditions such as voltage, current, temperature, state-of-charge, and depth of discharge on the battery’s health. By combining chemistry, control systems, and software, this historical diagnostic is critical to accurately assess and predict the state of health of the cell.
Second, accurate chemical models are necessary to normalize cell variations and provide a benchmark for expected behavior for a given cell type. These models describe the dependencies of material properties on temperature, state-of-charge, voltage, current, depth-of-discharge, cell design, chemistry, and vendors. They should also measure the impact of manufacturing variations, isolate embedded defects, compensate for non-uniformity of the material properties and assess the effect of the operating conditions in real-time to adapt the charging conditions to the battery’s health. These models must be dynamic and learn from the field data to improve their accuracy over time. This marriage of chemistry, data, and software can generate an accurate predictive state of health.
Lastly, the ability to adapt the charging conditions in real-time to optimize the operation of the battery based on the predictive state of health is critical to improving safety. This is where closed-loop algorithms come into play — being able to diagnose, measure, predict and adjust the charging conditions in real-time to ensure safety at all times. Current open-loop systems such as CCCV and step-charging use fixed charging profiles regardless of battery SOH, further contributing to potential safety issues, especially in the presence of defects.
Not all safety issues can be avoided. In rare conditions, batteries fail. These include manufacturing inconsistencies, mechanical defects, and collisions. While these inconsistencies may pass initial quality checks and remain latent for many charge-discharge cycles, these defects change internal chemical processes, leaving distinct early signatures. A database of defect signatures to identify and assess the failure under wide operating conditions can help predict the safety failures weeks before they lead to a fire.
Early warning systems are essential to identify and remove unsafe batteries from circulation long before they become safety hazards. With the growing size of EV batteries, this is one area where regulation or enforcement may help. The GTR 21 and GB standards are pushing the industry to adopt newer technologies for better safety.
The predictive safety software stack to identify the rare battery defect before they lead to fire comes standard with Qnovo’s intelligent battery management software. Our models, algorithms, and software are deployed across 100+ million devices worldwide and proven to work under various operating conditions.
Conclusion: Safety First
Failures can appear suddenly and violently, with the potential of severe injuries, product recalls, and loss of reputation. It is crucial to invest in software intelligence to monitor battery diagnostics, predict the presence of defects, and be prepared to remove hazardous batteries from circulation before they become catastrophic incidents.
Sam Kanakamedala is Director of Product Management at Qnovo. He is passionate about building products and bringing great technology to market. His past experience includes product management, technology implementation, and marketing strategy.
Dania, a Qnovo co-founder, is a battery scientist at heart. She brings a deep base of experience in Lithium-Ion technology and materials from her work at Imara, Nanogram, and Greatbatch. Her expertise enables her to understand the theoretical promise of new materials and technology, yet distill it into practical implementations that can be readily commercialized. Dania’s team develops the core battery science behind Qnovo’s battery management algorithms.
The views and opinions expressed in this article are those of the author and do not constitute an endorsement or recommendation by the BatteryBits publication of the products described therein.
Interested in publishing in Battery Bits? Apply at this link to become a contributor. Join our discussion with other battery professionals at the community slack space Battery Street!
