Cloud-based Analysis Makes High-voltage Batteries More Efficient and Durable

Digital Twin, Machine Learning and Big Data Optimize the Electric Drive

Electromobility is not only changing the drive itself, but also the entire vehicle ecosystem – especially at a digital level. As the central component of the electric car, the battery is becoming the focus of new strategies for monitoring, maintenance and optimization. At the VDI Congress Dritev 2025 on July 9 and 10, 2025 in Baden-Baden, Dipl.-Ing. Stephan Schade, Head of Connected Services, Vehicle Health at Robert Bosch GmbH, will explain how cloud-based analysis methods help to make high-voltage batteries more efficient and extend their service life.
 

The battery has a significant influence on the price, range and everyday usability of electric vehicles. Despite falling costs per kilowatt hour, it remains a key cost factor. At the same time, expectations are rising in terms of charging speed, durability and secondary benefits such as sustainability and resale value. Various target variables must therefore be reconciled when operating high-voltage batteries. Battery management can make a significant contribution to this. With “Battery in the Cloud”, Bosch has developed a system based on digital twins, big data, machine learning and electrochemical models – opening up new possibilities for predictive battery maintenance.
 

Transparency About the Battery Condition

“The battery is the most expensive component of an electric vehicle,” says Stephan Schade. “It is therefore essential to understand how it ages, to continuously monitor its condition and to extend its service life through data-based optimization.” According to him, OEMs, fleet operators and end customers alike can benefit from the resulting advantages – right through to the used car market. “More and more used electric vehicles are coming onto the market, but until now there has been a lack of transparency about the condition of the battery,” explains Stephan Schade. Cloud-based analyses can reduce uncertainties when determining residual value and provide fact-based diagnoses.

At the heart of this is the digital twin, a dynamic and learning model that maps the real use of the battery. "We use a concept that combines physical or electrochemical models, AI-based models and big data approaches in a hybrid algorithm. Only this combination creates a precise and scalable system," emphasizes Schade. 
 

Predictive Analytics for Early Fault Detection

The strength of cloud-based systems lies in their ability to analyze large volumes of data and derive insights that outperform conventional systems. “We can now identify anomalies in battery behavior at an early stage – long before an actual fault occurs,” says Stephan Schade. These include, for example, thermal anomalies, cell imbalances or atypical discharge profiles. A two-digit number of such anomalies have already been defined in order to identify them at an early stage using AI-based pattern recognition.
 

Lifetime Optimization: Software Extends Service Life

One specific benefit of these analyses is the “Lifetime Optimization” service, which has been successfully in use since 2022. This involves continuously analyzing charging and driving data to derive recommendations for action. “In a benchmark project with around 3,000 vehicles over 18 months, we were initially able to prove that the service life of the batteries can be extended by up to 20 percent through software-supported optimization – without any changes to the hardware,” explains Stephan Schade. Vehicles with intensive fast charging in particular benefit from significantly reduced cell ageing and improved range. “We can provide specific charging profiles for individual vehicles – depending on how the vehicle is used,” explains Schade. This not only optimizes the user experience, but also extends battery life.
 

Real-time Feedback on Battery Condition

For fleet operators, Bosch provides real-time dashboards that visualize the health status of the batteries in real time. The fleet manager can decide whether and how to react to certain loads. “We don't give instructions, but provide a basis for decision-making,” says Schade. Feedback is also possible for end users – for example via manufacturer apps. “Data sovereignty remains with the OEM, we only provide the anonymized analyses as a basis for decision-making,” emphasizes Stephan Schade.
 

Battery Passport Creates Transparency for Resale

Another result of these analyses is the battery passport – a certificate that documents the state of health and remaining capacity and thus facilitates resale. This is relevant for both fleets and private customers when reselling. The intention: more transparency about the actual battery condition when selling can help to create more trust in the used market for e-vehicles.
 

From the Street to the Lab – and Back

The knowledge gained from real-life use flows directly back into product development. Models can already be used in the concept phase to simulate cell types or charging behavior. "Within a few days, we can make statements about the behavior of a battery even before it is actually used. This saves time, money and resources and supports the continuous optimization of future battery generations," concludes Stephan Schade.