Architecture for autonomous shape error compensation in tool grinding

Abstract

Process planning of tool grinding operations for individual cylindrical tools requires expert knowledge as well as adjustment tests in order to enable productive manufacturing according to the quality requirements. Static deflections of the cylindrical blank lead especially in the case of drilling tools to shape errors and core diameter deviations that vary with the axial workpiece position. This paper presents an architecture to compensate for shape errors autonomously in process planning by using a technological NC-Simulation. Based on a fast prediction of the elastic workpiece deflection, the initial NC code is modified by optimizing process parameters and adapting the tool path according to the bending line. A concept for data feedback ensures self-learning effects and enables model adaption. It is shown how the prediction can be adjusted for unknown grinding wheel specifications between the grain sizes D9 and D54. In experimental investigations, the shape error could be reduced in a range of 88 % to 99 % with a productivity increase of 47 %.