Remaining Useful Life Estimation of Cutting Tools Using Bayesian Augmented Lagrangian Algorithm

Cutting tools are vital components of computer numerical control (CNC) milling machines, which have high failure rates. The failure of cutting tools will lead to a complete production stoppage, which in turn will result in significant financial losses. As a result, an effective remaining useful life (RUL) estimation technique is urgently requested to monitor cutting tool states in order to prevent negative impacts on the product due to damage to the cutting tools. At present, a variety of RUL methods have been attempted trying to estimate the tool wear levels. However, they consume significant computational resources. To overcome this issue, in the present study, a novel Bayesian augmented Lagrangian (BAL) algorithm is applied to estimate the cutting tool wear of a CNC milling machine. The characteristic of BAL is that it transforms the original optimization problem into several sub-optimization problems which can be solved separately under the Bayesian framework. This process can greatly increase the speed of computation. A case study for estimating CNC milling machine cutting tool wear based on the BAL method is presented, and the results validate the effectiveness and reliability of the method.