Self-optimizing tool path generation for 5-axis machining processes

Abstract This paper presents a self-optimizing process planning approach for 5-axis milling that allows an automatic compensation for tool deflection. For this purpose, process conditions are obtained from a process-parallel material removal simulation and merged with shape error measurements. Using machine learning methods, the resulting shape error is predicted and the tool path adapted automatically. The system has been implemented on a 5-axis CNC machine centre. It is shown that the resulting shape error can be reduced by 50%. Moreover, the article highlights the behaviour of the learning process and the transferability to other workpiece geometries.