Dürr expands first paint shop AI application

Previously, AI could only be used to identify sources of defects, develop optimal maintenance plans and improve production processes in paint shops equipped with state-of-the-art robots. Now, however, Dürr is significantly expanding the use of AI by applying the analysis software of the DXQ product family to sealing robots. In addition, for the first time, a unique interface solution enables the connection of data from robots in existing paint shops.

Fig. 1: Using artificial intelligence, Advanced Analytics can detect the source of defects at an early stage of the paint application, and the technology can also be applied to high-viscosity materials.

Automotive manufacturing plants have a vast amount of potential data about manufacturing processes, raw materials and products. The key to utilising this data is data connectivity - i.e. having the right data interfaces at the control level to get first-hand data from robots, bakehouses, electrophoresis systems or conveyor technology.

Only by recording the relevant machine data in real time (e.g. axis position and temperature, or events such as alarms and programme start and end times) and uploading it to a database in a timely manner can IT be used to improve coating quality and plant availability. Jens H?cker, Vice President Control Systems at Dürr, explains: "Only by realising this basic prerequisite can the DXQ series software determine the current status of the plant equipment. Therefore, we use a combination of data acquisition, historical data and machine learning to detect previously unknown sources of defects or to precisely schedule maintenance."

Connecting data to existing plant equipment

Previously, data connectivity was not possible, whether it was with Dürr's earlier robots, robots from other manufacturers, or technology outside of painting applications. Even with Dürr robots, only the new generation of robots was equipped with the correct data acquisition interfaces. As a result, although the need for digital applications was high, their use in existing plants was limited. Nevertheless, Dürr has found a way to enable data connectivity between almost all common robots and behaviours.

Collecting and analysing detailed information on all process steps

The solution is an adapter consisting of hardware and software that interoperates with all current bus technologies and provides data with the necessary high time resolution of a few milliseconds. The adapter was developed by Dürr in cooperation with Techno-Step, a company specialising in the field of process data analysis and diagnostic systems and, since 2020. a subsidiary of the Dürr Group. "Operators can therefore read the data available in an existing plant from sensors and actuators, and then integrate all the rules, from pre-processing to application to conveyor technology, into a single analysis software. With DXQequipment.analytics, they have a detailed view of the individual process steps along the entire value chain and all the system information involved," says Jens H?cker.

Figure 2: Advanced Analytics from the Dürr DXQ software product family is the first AI application marketed for paint shops, and the technology is also available for sealing robots.

Expertise in mechanical engineering and IT

With its built-in advanced data analysis module, DXQequipment.analytics is the first market-oriented solution to date that uses AI technology to improve the overall efficiency of paint shop equipment. Dürr has adapted the AI model for analysing robot and process data to the specific requirements of the sealing sector, thus using the advanced data analysis module for the sealing sector. Dürr has comprehensive expertise in production technology and manufacturing processes for the automotive industry, as well as a high level of digitalisation. To meet the challenges of applying the Data Analytics Module to seals, Dürr is combining these two areas of expertise, so that in the future artificial intelligence can be used to accurately detect the source of defects early on in the application of highly viscous materials and to determine the optimal maintenance schedule. Take the detection of clogged nozzles as an example. Sealing materials can cause coating nozzles to become partially clogged, which affects material injection and results in coating quality defects that lead to rework. Unlike conventional control techniques, DXQ software detects this defect and enables early intervention.