Heterogeneous data-driven hybrid machine learning for tool condition prognosis

Peng Wang, Ziye Liu, Robert X. Gao, Yuebin Guo

Research output: Contribution to journalArticlepeer-review

52 Scopus citations


Cutting tool condition prognosis is critical to process stability and quality assurance, but affected by complex material-process interactions. This paper presents a hybrid machine learning method that integrates heterogeneous data (structured process parameters and unstructured power profiles and tool wear images) for tool condition prognosis. Surface and wear images are first analyzed by a convolutional neural network to identify surface roughness and wear severity. The results are subsequently fed into a recurrent neural network to reveal the relationship between tool condition degradation and power profiles. The fidelity of the method is validated in milling of H13 steel and Inconel 718.

Original languageEnglish
Pages (from-to)455-458
Number of pages4
JournalCIRP Annals
Issue number1
StatePublished - 2019

Bibliographical note

Publisher Copyright:
© 2019


  • Energy
  • Machine learning
  • Surface

ASJC Scopus subject areas

  • Mechanical Engineering
  • Industrial and Manufacturing Engineering


Dive into the research topics of 'Heterogeneous data-driven hybrid machine learning for tool condition prognosis'. Together they form a unique fingerprint.

Cite this