Use of predictive quality analytics in body-in-white production

Body in White production in automotive

The topic of “Predictive Quality Analytics” plays an important role in the planning of many companies. When it comes to the concrete implementation of a solution in the customer environment, not only the technology but also a deep understanding of processes by all those involved is required.

Below we present a typical use case as an example:

Initial situation

In the past, individual inspection orders were generated on the basis of predefined static inspection plans. In the SPC data collection, either parts are acknowledged as “OK” or “nOK” in relation to the respective inspection order. In order to obtain an overview of the reported defects, all defect information including images are recorded and displayed. In the actual customer case, more than 10 employees are permanently engaged in ultrasonic testing.

AI Use Case

metal production

In a joint project from 2019, the basis for the realization of an AI-based process optimization was set up. One of the goals was to support anomaly detection with a defined algorithm and to reduce the inspection scope. For this purpose, the process data is compared with other welding information already in the system and thus it can be decided in real time whether the welded spot which was classified as “not OK” is relevant for a manual re-inspection and thus has to be re-inspected by an inspector. In this case, the body is directed to a designated inspection station and the inspector, using an overview, can recheck all the anomaly points collected.

The number and the effort of manual ultrasonic inspections could be reduced by more than half by using the “Predictive Quality Analytics” approach.

metal production

Improvement / Benefit

  • By integrating “Predictive Quality Analytics” into the existing inspection strategy, the quality of the ultrasonic inspection is significantly increased, since only conspicuous weld spots (anomaly spots) are inspected in addition to the existing equipment inspection.
  • By integrating “Predictive Quality Analytics”, the inspection rate increases to 100% of all welded spots.
  • The image database enables a pictorial representation of the weld spots to be inspected and feedback of the anomaly points.
  • As a result, the customer achieved a significant reduction in manual ultrasonic inspections.


For more information or comparable please contact us.

Get in touch!

Do you have specific questions about our solutions? Get in touch!

Ludmila Lebedev