Model based Fault Diagnosis and Supervision of the Drilling Process

Abstract The supervision of manufacturing processes is of primary interest in the course of progressive automation. This paper paper presents a new approach in fault diagnosis and supervision of the drilling process. The supervision of manufacturing processes with additional sensors is mostly expensive or not possible at all. Therefore mathematical process models were developed for the drilling process. By means of least squares estimation techniques the physical parameters are determined based on easily measurable signals of the drives. Changes of those process parameters during operation are used for fault diagnosis. Experimental results on two flexible machining centers show the detected parameter changes during operation in dependence on tool wear.