Material removal rate prediction with force signals and machine learning in magnetically driven internal finishing

Magnetically driven internal finishing (MDIF) is a novel internal polishing technique that employs an external magnet to drive a magnetic polishing tool to achieve material removal (MR) in internal surfaces. It has localised finishing and corrective polishing abilities. Modelling the material removal rate (MRR) can provide better control of the polishing process. However, the complex interaction between the polishing tool and the workpiece in MDIF results in oscillating polishing forces which are significantly different from conventional polishing processes. This makes the bottom-to-up physical modelling of MRR very challenging. Herein, we developed a novel broad echo state learning system (BESLS) to predict MRR in MDIF based on the polishing force under various polishing conditions. The MRR is represented by the removal depth of the polishing footprint. The results show that the BESLS model can achieve a predicting error as low as 7.5%. The proposed method will boost the in-process control to achieve uniform removal in MDIF and may also be further applied to other polishing processes. © European Society for Precision Engineering and Nanotechnology, Conference Proceedings - 22nd International Conference and Exhibition, EUSPEN 2022. All rights reserved.