Beschreibung
Grinding of helical flutes is an important step in the process chain of cylindrical tool manufacturing. The grinding process defines the dynamic performance of the manufactured tool. Moreover, the surface quality and flute shape are primarily determined. For this reason, finding optimum process parameters is essential, especially in process planning of individual tools. In the industry, manual experiments are carried out for the machine set up. This leads to high costs due to machine hours, labor and material costs. This paper presents a self-optimizing and adaptable process planning method to reduce costs for process planning and improve the process result. The developed method allows to identify optimum cutting speed and feed with respect to economic efficiency and quality for new tools without additional machining experiments. Geometric-kinematical cutting simulations in combination with empirical models are used to predict the process outcome. The empirical models are derived from machine learning and improved automatically with an increase of process data. Applying the presented method in a case study, the machining time could be reduced by up to 38.1% and the core diameter deviation by up to 73.7%. Moreover, it is shown that the presented methods allow a continuous improvement of the process models.
