HUHN Pauli Deutschland GmbH specializes in high-speed cold forming of metals using progressive dies. It is economically crucial to detect quality defects in parts and tools as early as possible because each defective part results in waste and lost production capacity. Typically, only statistical quality inspections are performed, as in-situ inspections, e.g., using camera systems, are expensive and require complex mechanical setups.
aiXbrain shall analyze process data collected by special sensors embedded in the forming tool to identify emerging quality issues using artificial intelligence.
The joint goal is to support machine operators, who often oversee multiple production lines, with an AI-driven solution for continuous in-situ quality monitoring.
As part of the BMBF-funded research project GeMeKI, a progressive die was equipped with a special sensor package and the meastream Industrial Edge Toolbox.
The AI mainly analyzes pressure and vibration signals, which are spectrally transformed (FFT) in a near-process pre-processing step for each press stroke.
A specially designed production and measurement campaign generates training data that also includes rare wear effects.
We reduce the dataset by filtering spectral coefficients, enabling fast calculations even on cost-efficient hardware.
We train the AI using supervised deep learning and unsupervised random forest models.
The trained AI accurately detects bending angle quality and distinguishes between defects such as too small or too large bending angles, enabling cost-effective in-situ quality monitoring.
The combination of sensor package, edge computing, and AI creates an affordable solution for 100% quality control, reducing scrap and unplanned maintenance. At the same time, it relieves machine operators and quality assurance teams.
Looking ahead, this approach can achieve manufacturing cost savings of 5–10%.
aiXbrain offers the AI component through its Dataray® AI Framework either as a managed service or as an AI app in a SaaS model across various platforms.
"We experienced aiXbrain as a highly reliable partner, combining strong AI expertise with a deep understanding of our processes and specific needs."
"We have found aiXbrain to be a reliable project partner with great AI expertise and a deep understanding of our processes and requirements."