Artificial intelligence increases paint shop quality
Artificial intelligence can control highly sensitive systems in automobile production even more precisely. This is demonstrated by a pilot project in the paint shop at the BMW Group plant in Munich.
Despite state-of-the-art filter technology, the proportion of very fine dust particles in paint lines fluctuates depending on the ambient air drawn in. If the dust load is above the limit value, the still moist paint can trap particles and thus visually impair the painted surface. Artificial intelligence (AI) specialists from the planning department and the Munich plant have now found a way to avoid this situation. Every freshly painted body passes through an automatic surface inspection in the paint shop. The data from this inspection serves as the basis for building a comprehensive database for dust particle analysis. To compare live data from dust particle sensors in the paint booths and dryers with this database, the specialists are now using AI algorithms.
“Data-based solutions help us to secure and further expand the high quality requirements in the interest of our customers. Smart Data Analytics and AI provide our employees with important decision-making aids for process improvements. We have filed several patent applications for the innovative technology of dust particle analysis,” says Albin Dirndorfer, Head of Painted Body, Finish and Surface at the BMW Group. Two concrete examples underline the benefits of this new AI application: If dust levels rise due to the season or during prolonged dry periods, the algorithm recognizes this trend in good time and calculates, for example, an earlier time for filter changes. In combination with other analysis tools, additional patterns can be detected. Another analysis result could be that the system that uses ostrich feathers to remove dust particles from the car bodies needs to be fine-tuned. The BMW Group’s AI specialists see great potential in dust particle analysis. Fed by information from numerous sensors and data from surface inspection, the algorithm monitors over 160 characteristics of the car body and can predict the quality of paint application with great accuracy. A prerequisite for later series uses of this AI application is an even broader data basis for the algorithm, for which additional measuring points and even more precise sensor data are required in particular when cleaning the car bodies. The AI experts are confident that the dust particle analysis can also be used in other vehicle plants once the pilot phase at the Munich main plant has been completed.
Further information: www.bmwgroup.com