Assessing EV batteries’ state-of-health with a Battery Quick Check.
Recent studies by RTWH Aachen revealed that the failure rate of electric vehicles (EVs) is only between 0.9 and 1.2 per 10,000 vehicles, compared to 7.8 fire accidents per 10,000 with conventional combustion engines.
However, reports of alleged EV battery fires have remained a critical safety concern for manufacturers and a potential deterrent for some end users.
This is where artificial intelligence (AI) offers new opportunities. If ChatGPT comes to mind, know that AI plays a much larger role in the auto industry. In the case of EVs, it can support battery optimization, addressing critical challenges in battery development, reliability, and safety.
Vehicles are increasingly networked to collect large amounts of data, and an EV battery is no different. It’s accompanied by various sensors and control units — the battery management system (BMS) — designed to ensure safe operation and generate large amounts of data.
By using AI, more powerful battery management systems can be developed with the aim of fully negating battery overloads and failures.
BMS: The benefits and limitations
Engineers are racing to develop an EV battery with the greatest range, fastest charging speed, and lowest cost. In doing so, they’re often limited in accessing data because the battery cells are too new or there’s too much manual data to sort through.
The BMS is a first step toward greater safety, as it controls the battery and protects it from harmful operating conditions such as charging or discharging too quickly. A BMS is designed to prevent catastrophic failures by regulating the battery operation. It does so by using data on electrical parameters (current and voltage) and thermal conditions. This information can be accessed and transferred to the cloud, allowing manufacturers and operators to monitor the battery’s condition remotely.
EV battery testing. This device generates randomized data sets for AI training and testing.
However, this system also has limits. Typically, the primary task of the BMS is to control the battery rather than to recognize long-term trends. It reacts more to faults that occur at a specific time, much like how a sprinkler system reacts to a fire that’s already burning but cannot warn of an outbreak. This means that future safety-critical effects might go unnoticed.
This is where Cloud-supported AI analytics found in battery analytics software can offer additional protection, detecting and qualifying battery anomalies — at the cell level, in individual EVs, or across an entire fleet of vehicles. Artificial intelligence can provide a timely warning of battery failure or, in the worst-case scenario, a fire long before either occurs.
Although such fires are extremely rare, as the RTWH Aachen report proved, they’re still possible. The chemical composition of an EV battery also makes such fires difficult or impossible to extinguish quickly.
Effects of battery aging
In addition to safety risks, an EV battery is also subject to continuous wear and tear — known as battery aging. Aging depends on time, usage, charging practices, and environmental conditions.
For example, frequent fast charging at low outside temperatures can lead to battery aging or damage, especially if the battery is not sufficiently preheated. The good news is that data from the vehicle can allow AI-supported systems to continuously monitor these effects and make usage recommendations or technical interventions (e.g., better charge control). The result is an AI system that protects the battery’s longevity over time, saving costly battery repairs or replacements.
AI-supported systems ensure safer, longer-lasting batteries and promote greater sustainability. According to a recent study, EVs already produce up to 89% less CO2 emissions than combustion engines. If EVs are charged with green electricity, their balance sheet improves further.
The TWAICE Battery Research Center, located in Munich, Germany.
Two additional factors positively impact sustainability: maximizing the battery’s service life and reusing it in the energy sector. Long-term service life can be achieved primarily through optimized design and use. These are mainly determined by insights into the battery’s aging and best practices (or activities to avoid, such as fast charging in cold environments without pre-conditioning the battery).
Transparency about the battery condition is also a prerequisite for reuse after EV life, such as for buffer storage at home or at fast-charging stations. But this can only happen if the battery is still suitable for operation after removing it from the vehicle and converting it. And this is where AI can provide a diagnosis of the current condition and a service life forecast.
AI is the key
Transportation is transforming, becoming increasingly electrified. Auto manufacturing is increasingly relying on data to optimize its products. If used correctly, it’s possible to learn from this data, ensuring safer vehicles and more sustainable mobility — all without sacrificing comfort. Artificial intelligence is providing the key. Its systematic use can extend the service life of EV batteries, reduce vehicle costs, and (most importantly) increase safety and sustainability.
You may also like:
Filed Under: Batteries, FAQs