CNC machine tool health assessment using Dynamic Bayesian Networks

Abstract

AbstractThe failure of critical components in physical systems may have negative consequences on the availability, the productivity, their security and on the environment. Thus, the assessment of the critical component's health condition, which can be done in the diagnostic and prognostic framework, should be constantly ensured. In this paper, a contribution on the assessment of the health condition of the cutting tool from a Computer Numerical Control (CNC) machine tool and the prediction of its remaining useful life before its complete failure is addressed. The proposed method is based on the use of monitoring data and relies on two main phases: an off-line phase and an on-line phase. During the first phase, the raw data provided by the sensors are processed to extract reliable features. These latter are then fed as inputs to the learning algorithms in order to generate relevant models that best represent the behavior of the cutting tool. The second phase is an assessment one, which uses the constructed models to identify the current health state and to compute the remaining useful life and the associated confidence value. The method is applied on monitoring data gathered during several cuts of the CNC tool and simulation results are given and discussed.