DATA-DRIVEN APPROACHES FOR GUN-DRILLING TOOL CONDITION MONITORING

Abstract

The main objective of data-driven tool condition monitoring (TCM) research is to estimate and predict operating machine tool health condition based on signal data processing and decision-making algorithms. Well-performed TCM system has promising research value due to its importance in achieving smart manufacturing, improving industry efficiency and bringing economic benefits. Although various research work has been conducted to develop effective decision-making algorithms for tool condition fault diagnosis and prognosis, there still exist some open problems to improve algorithms performance. In this thesis, a few artificial intelligence based methods are proposed to address the gaps in data-driven tool wear condition regression, classification, and remaining useful life prediction problems.