Remaining Useful Life Prediction Based on Spindle Load and Cutting Process Parameters in Machining

Abstract

Abstract Tool wear increases machining resistance, part dimensional inaccuracy and machining vibration. Tool wear monitoring and Remaining Useful Life (RUL) prediction of the tool during machining operation will assist a machine operator to provide tool wear compensation at the right time and plan the tool change activity. These aspects become significantly important for economical and quality production. This work focuses on a physics and data-based approach for monitoring cutting tool wear state and Remaining Useful Life (RUL) during a machining operation by adapting a well-known empirical wear-rate equation. The constants in the model are estimated based on machine heuristics which depends on the tool-machine-workpiece combination. The proposed model takes real-time spindle power and machining process parameters as inputs, which are obtained directly through querying the CNC controller. Therefore, it does not require the mounting of any external sensors on the CNC machine tool. Hence, the proposed method is a more economical and convenient way to predict tool wear and RUL in a machining shop floor. The model is validated from experimental data and it can capture the progression of tool wear and RUL of the tool at any point of time during a machining operation. Since the model captures the physics of tool wear and machining heuristics, it is more robust than a purely data-based model.