Lifetime monitoring and anomaly detection

Smart analysis of battery data and identification of stress factors

A variety of factors are responsible for causing premature battery aging, though they are often related to usage behavior. Among the negative influencing factors are, for example, fast charging at low temperatures, hard acceleration, and battery states of charge that are too high (> 95% SOC) at high temperatures. Bosch’s service identifies these stress factors and can detect battery anomalies at the cell and module level early on. For this purpose, battery condition data is acquired in real time for each connected vehicle and is transmitted to the Bosch cloud.

Digital battery twin in the cloud

The data acquired in the connected vehicle is transmitted to the cloud where a digital twin exists for each battery. This virtual copy of the real battery includes electrochemical and data-driven models combined with battery data; it portrays the processes taking place inside the battery in the cloud. Any anomalies at the cell or module level are identified using hybrid models, which combine the electrochemical models with machine learning methods. This combination of electrochemistry and artificial intelligence enables an especially accurate analysis of the battery data with minimal uncertainty.