The system, which uses artificial intelligence (AI), analyses data from conveying equipment to identify potential faults early and avoid disruptions.
The system has been successful in avoiding an average of around 500 minutes of disruption per year at the Regensburg plant, located 124km north-east from its headquarters here.
The system does not require additional sensors or hardware, but instead evaluates existing data from installed components and conveyor element control. An alarm is triggered if anomalies are found.
The data from the load carriers used in assembly is transmitted to the BMW Group's predictive maintenance cloud platform, where the analysis takes place.
The system constantly searches for irregularities such as fluctuations in power consumption or abnormalities in conveyor movements.
If anomalies are detected, a warning message is sent to the maintenance control centre, which can then take the affected vehicle out of the assembly line for repair.
The system has been standardised and implemented in collaboration with other BMW Group plant sites, making it cost-effective and easy to roll out to other locations.
Machine-learning models have been developed in-house to improve the system's accuracy.
The team is continuously working on refining the algorithms and integrating recommended actions into fault messages. The goal is to further exploit the possibilities of AI and improve predictability.
The system has already been successful in avoiding downtime and is being evaluated for use in other areas of the plant, such as filling vehicles with brake fluid and coolant.
The integrated learning system developed by BMW Group Plant Regensburg is the first of its kind and has already been patented by the company.
The Regensburg plant has been in operation since 1986 and is one of more than 30 BMW Group production locations worldwide.
A total of up to 1,000 vehicles of the 1 Series, X1 and X2 models come off the production line every workday – destined for customers all over the world.
