Feature extraction methods for quality attributes prediction in injection moulding

Issue title: Selected papers from the ISCA International Conference on Software Engineering and Data Engineering, and the ISCA International Conference on Computer Applications in Industry and Engineering, 2015, and Invited Papers

Guest editors: Takaaki Goto and Narayan C. Debnath

Article type: Research Article

Authors: Wunck, Christoph

Affiliations: College of Management, Information, Technology, Jade University of Applied Sciences, Wilhelmshaven 26389, Germany. Tel.: +49 4421 985 2466; Fax: +49 4421 985 2412; E-mail: wunck@jade-hs.de

Abstract: A method to extract essential features from time domain process signals of injection moulding machines based on discrete wavelet transform is presented and compared to existing methods. The performance of the new method is assessed by comparing the goodness of fit of linear regression models to estimate final part quality. The results of the experiments show that the wavelet-based feature extraction method performs best regarding model performance and size.

Keywords: Injection moulding, , machine learning, , feature extraction, , quality prediction

DOI: 10.3233/JCM-160681

Journal: Journal of Computational Methods in Sciences and Engineering, vol. 17, no. S1, pp. S71-S83, 2017