Journal Published Online: 15 January 2021
Volume 5, Issue 1

Federated Learning as a Privacy-Providing Machine Learning for Defect Predictions in Smart Manufacturing

CODEN: SSMSCY

Abstract

In this work, the federated learning methodology is applied to predict defects in sheet metal forming processes exposed to sources of scatter in the material properties and process parameters. Numerical simulations of the U-channel forming process were performed to analyze springback for three types of sheet steel materials. The datasets of different clients are used to train a single machine learning model. With this approach, multiple parties would simultaneously train a machine learning model on their combined data by training the models locally on the client nodes and progressively improving the learning model through interaction with the central server. This way the industrial peers have no access to the others local data in a centralized server. The predictive performance achieved is similar to a standard centralized learning method, offering competitive results of collaborative machine learning in industrial environment.

Author Information

da Silveira Dib, Mario Alberto
University of Coimbra, Centre for Informatics and Systems of the University of Coimbra (CISUC), Coimbra, Portugal
Ribeiro, Bernardete
University of Coimbra, Centre for Informatics and Systems of the University of Coimbra (CISUC), Department of Informatics Engineering, Polo II – Pinhal de Marrocos, Coimbra, Portugal
Prates, Pedro
University of Coimbra, Centre for Mechanical Engineering, Materials and Processes (CEMMPRE), Department of Mechanical Engineering, Coimbra, Portugal
Pages: 17
Price: Free
Related
Reprints and Permissions
Reprints and copyright permissions can be requested through the
Copyright Clearance Center
Details
Stock #: SSMS20200029
ISSN: 2520-6478
DOI: 10.1520/SSMS20200029