Experimental study of the process failure diagnosis in additive manufacturing based on acoustic emission

•A data-driven monitoring and diagnostic method with acoustic emission sensor is developed.•Acoustic emission features are extracted to recognize process failures.•Unsupervised self-organizing map is applied to formalize the diagnostic procedure. Currently additive manufacturing (AM) faces many chal...

Full description

Saved in:
Bibliographic Details
Published in:Measurement : journal of the International Measurement Confederation 2019-03, Vol.136, p.445-453
Main Authors: Wu, Haixi, Yu, Zhonghua, Wang, Yan
Format: Article
Language:English
Subjects:
Acoustic emission
Acoustic emission testing
Additive manufacturing
Diagnostic software
Diagnostic systems
Failure analysis
Failure diagnosis
Failure modes
Fused deposition modeling
Fused filament fabrication
Machine learning
Monitoring
Monitoring systems
Printing
Self organizing maps
Self-organizing map
Signal processing
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:•A data-driven monitoring and diagnostic method with acoustic emission sensor is developed.•Acoustic emission features are extracted to recognize process failures.•Unsupervised self-organizing map is applied to formalize the diagnostic procedure. Currently additive manufacturing (AM) faces many challenges related to product quality, robustness, and reliability, which will hinder its industry-scale applications. Studies of the typical AM process failures and the corresponding effective sensor-based monitoring methods are needed to overcome these challenges. In this paper, a data-driven monitoring method for online AM process failure diagnosis based on acoustic emission (AE) is proposed, and its application to the fused filament fabrication (FFF) is demonstrated. Several typical FFF process failures are also investigated. Experimental study of a failed FFF printing process is benchmarked against the analysis of a normal printing process. AE signals from both normal and failed printing processes are recorded and processed. An unsupervised machine learning method, self-organizing map (SOM), is applied to formalize the diagnosis procedure of different failure modes. The experimental results show that it is feasible to use the proposed method to diagnose the typical process failures, including both detection and identification. This new method could serve as a non-intrusive diagnostic tool for FFF process monitoring, and has the potential to be applied to other AM processes.
ISSN:0263-2241
1873-412X
DOI:10.1016/j.measurement.2018.12.067