Hirschvogel is turning to AI in die forging
AI-assisted inspection instead of visual inspection: Hirschvogel, a specialist in mass forming, is replacing the conventional manual inspection of die-forged parts with a fully automated final inspection.
Together with Hirschvogel, one of the world’s largest automotive suppliers in the field of bulk forming, SMS is setting a technological milestone in the quality assurance of drop-forged parts. In collaboration with the image processing specialist Sightwise, a system for the automated final inspection of drop-forged components has been developed, which replaces an existing manual inspection station at the Denklingen site in Germany. As an integral part of iForge – the SMS group’s digitalisation solution for die forging – the system enables fully automated inspection for surface defects and geometric deviations. This replaces the time-consuming manual visual inspection with a scalable digital process, which sustainably boosts efficiency at the industrial site in Germany.
At the heart of the solution is a platform that efficiently trains analysis models using synthetic data. Instead of collecting real defect images from physical parts over a period of weeks, the system uses CAD data or 3D models. Material properties and surface textures are simulated digitally with such precision that they accurately reflect real-world conditions. An automated process generates photorealistic image data and virtually applies various defect types to the components. The result: a ready-to-use analysis model within just six to eight hours.
Classification error rates are minimised
The inspection process at Hirschvogel involves a fully automated feeding system: a robot removes the forged parts directly from the transport crate. High-resolution camera sensors capture the components, whilst the software analyses them in a fraction of a second. The system reliably identifies geometric deviations and surface defects and classifies them precisely according to the digital defect catalogue. This objective AI evaluation minimises the error rate in classification, thereby almost completely eliminating pseudo-rejects – that is, the erroneous sorting out of good parts.
According to SMS, the benefits for mass forming are measurable: the high degree of automation reduces reliance on staff, lowers labour costs and eliminates sources of human error. The high scalability also allows the system to be quickly adapted to new product variants.
Source: SMS group