Using multivariate models to monitor end-mill wear and predict tool failure

John T. Roth, Sudhakar M. Pandit

Research output: Contribution to journalArticlepeer-review

2 Scopus citations

Abstract

By using multivariate autoregressive models fit to acceleration signals, instead of univariate models, an earlier indication of approaching failure is obtained. From end-milling life tests data, it is demonstrated that the multivariate models can identify the indications of impending failure earlier than is possible using univariate models; for the case presented, 30 inches prior to failure, as opposed to the 6.5 inches obtained with the univariate models. This extra warning time, allows for preventive action to be taken and allows the possibility of finishing the current cut so that the tool change can be made between cuts.

Original languageEnglish (US)
Pages (from-to)1-6
Number of pages6
JournalTechnical Paper - Society of Manufacturing Engineers. MR
Issue numberMR99-138
StatePublished - 1999

All Science Journal Classification (ASJC) codes

  • Engineering(all)

Fingerprint

Dive into the research topics of 'Using multivariate models to monitor end-mill wear and predict tool failure'. Together they form a unique fingerprint.

Cite this