Development and Evaluation of a Machine Learning Model for the Prediction of Failures in an Injection Moulding Process

A. Rojas-Rodríguez, F. S. Chiwo, H. Arcos-Gutiérrez, C. Ovando-Vázquez, I. E. Garduño

Research output: Chapter in Book/Report/Conference proceedingChapter

1 Scopus citations

Abstract

Introduction: Developing and evaluating a machine learning model to predict failures in an injection moulding process offers significant potential to advance manufacturing competitiveness. Objectives: Failure prediction is considered one of the industry’s main applications of machine learning since it is crucial in process cost avoidance. The study of control parameters of the injection moulding (IM) process is a direct application of artificial intelligence in the industry. Still, there is the problem of the human acquisition of data. This research proposes an approach for failure detection and prediction in a dataset from an IM machine. Results: Results indicate that the approach helps find the optimal operating values for an IM process and reveals an identification scheme of failures and their relationship with the steps of the IM process. Conclusion: By performing a data clustering based on unlabelled data, information about productivity and quality of the IM process are obtained.

Original languageEnglish (US)
Title of host publicationEAI/Springer Innovations in Communication and Computing
PublisherSpringer Science and Business Media Deutschland GmbH
Pages101-122
Number of pages22
DOIs
StatePublished - 2023
Externally publishedYes

Publication series

NameEAI/Springer Innovations in Communication and Computing
VolumePart F673
ISSN (Print)2522-8595
ISSN (Electronic)2522-8609

Keywords

  • Data analytics for applications in Industry 4.0
  • DBSCAN
  • Failure detection
  • Industry 4.0
  • Injection moulding
  • Machine learning
  • Manufacturing processes
  • Parameters optimisation
  • Unsupervised learning

ASJC Scopus subject areas

  • Information Systems
  • Health Informatics
  • Computer Networks and Communications
  • Electrical and Electronic Engineering

Fingerprint

Dive into the research topics of 'Development and Evaluation of a Machine Learning Model for the Prediction of Failures in an Injection Moulding Process'. Together they form a unique fingerprint.

Cite this