Sebastian Feldmann^1*, Michael Schmiedt^1*, Julian Schlosser^1*, Wolfgang Rimkus^1*, Tobias Stempfle^1*

1 University of Applied Science - Aalen, Beethovenstr. 1, 73430 Aalen, Germany

* Corresponding authors emails: Sebastian.Feldmann@HS-Aalen.de, Michael.Schmiedt@HS-Aalen.de, Julian.Schlosser@HS-Aalen.de, Wolfgang.Rimkus@HS-Aalen.de, Tobias.Stempfle@Studimail.HS-Aalen.de

DOI10.24435/materialscloud:74-x1 [version v1]

Publication date: Jan 31, 2022

How to cite this record

Sebastian Feldmann, Michael Schmiedt, Julian Schlosser, Wolfgang Rimkus, Tobias Stempfle, Recursive quality optimization of a smart forming tool under use of perception based hybrid datasets for training of a deep neural network, Materials Cloud Archive 2022.16 (2022), https://doi.org/10.24435/materialscloud:74-x1

Description

In industrial metal forming processes, the generation of datasets for inline and optical quality assessment is expensive and time-consuming. Within the research project SimKI, conventional metal forming plants were digitalized under use of perception-based sensors in combination with a completely redesigned forming tool. The integration of optical quality observation methods connected with a retrofitting approach of the press tool provides the opportunity to generate an information-feedback loop that predicts part defects prior to their occurrence. The SimKI-method additionally combines conventional statistical measurement methods with AI-based defect detection algorithms that are trained by a) generic datasets of a finite-element simulation, b) real component images of a 3D imaging device, and c) a combination of both. The generated datasets are used to accelerate the training of a DNN-based algorithm in order to identify the position and deviation from the agreed quality. The high degree of innovation is based on obtaining real-time component quality information under use of AI-based optical quality assessment, which in turn provides information to the control algorithm of the smart forming tool.

Materials Cloud sections using this data

Files

File name	Size	Description
220122_SIMKI_Training_Datasets.zip MD5md5:a5233cc5e8b5d5fb201611175f5e4b61	17.0 MiB	This archive contains the training and validation datasets, including the trained neural network of the SIMKI-Project of Hochschule-Aalen.

License

Files and data are licensed under the terms of the following license: Creative Commons Attribution 4.0 International.
Metadata, except for email addresses, are licensed under the Creative Commons Attribution Share-Alike 4.0 International license.

External references

Keywords

Artificial Intelligence Gocator Defect Detection