Case Study
Software Sensors for Battery Management System
The challenge
Obtain Longer Driving Ranges by optimizing the Battery Management System
A quick adoption of electric vehicles is slowed down by their limited range compared to combustion engine cars. The range of electric vehicles depends on the capacity of the battery obtained after charging and the proposed strategy of extracting the energy. The battery management system (BMS) is responsible for both cases.
Our customer wants to better understand the optimal conditions for charging the battery to its maximum capacity and the optimal strategy of extracting the most energy out of the battery.
The solution
Charge & discharge behavioral analysis via AI to optimize real usage battery capacity
Endurance charging-discharging experiments with varying conditions were performed on same types of batteries. Different conditions of temperature, age, charging voltage, and charging and discharging current were tested in each of these experiments.
Key information of each of the experiments were extracted from which correlation structures and robust predictive models were created by using the self-service AI DataStories Platform. These predictive models gave understanding of the best condition for charging and discharging the battery. Software sensor models were created for timely predictions of key performance indicators of the battery management system like the State-Of-Charge (SOC).
The impact we created
Boosted battery capacity: more than 3% increase by novel charging strategy only
The new insights and predictive models combined with software sensors models (within 1% accuracy) boost capacity and help to provide timely predictions of BMS KPIs (key performance indicator) values. This will help the company to outperform the competition.
Results are being validated on track (June 2019).
Also read:
Belgium’s Agoria claims line honours in world solar electric car challenge