Antriebsbasierte Zustandsdiagnose von Vorschubantrieben

Abstract

Diese Arbeit befasst sich mit antriebsbasierter Diagnose für Werkzeugmaschinen. Die Verknüpfung ist nach dem Ursache-Wirkungsprinzip von Phänomenen in antriebsinternen Signalen mit schadhaften Veränderungen in der Mechanik gelungen. Als Grundsatz der Arbeit ist festzuhalten: Der Zustand des Antriebsstranges lässt sich durch die spektrale Zusammensetzung der übertragenen Leistung bestimmen und diagnostizieren. Die Ergebnisse werden zu einer allgemeingültigen, skalaren Kenngröße abstrahiert.

The main aim of further developing machine tools is the improvement of accuracy and reducing the consumption of resources. Maintenance resources and system downtimes have a major impact on the total cost of the equipment. In order to be competitive, it is necessary to reduce the system downtime by applying sensorless automated condition monitoring (SACM). An effective and low-cost condition monitoring can be realized by avoiding additional sensors. The principal purpose is to accomplish an early recognition brought about by deviations of the normal action of a machine tool. Merely taking a look at the apparent conditions of the components does not meet the requirements. The feed drive examined is based on the typical mechanical structures in the machine tool. This thesis provides an outline on SACM-algorithms based only on signals which are available in position-controlled feed drives such as position, speed and motor current. The cause-and-effect principle is used to link phenomena in the available signals with defective changes in the mechanics. The theoretical considerations and derivations that are shown, are used for the interpretation and analysis of experiments which were performed. Phenomena in the drive signals were caused by intentionally provoking various damages. A causal connection between phenomena and damage is demonstrated by these experiments. This interconnection is in turn examined and substantiated by the cause-and-effect principle. It was achieved to set up a generally accepted categorization of the causes of the phenomena in the form damage process and damage types. The damage types can be classified by their effects and phenomena in drive signals. They are divided into periodical and non-periodical damages. Due to the analyses made, it is possible to hold as a principle: The condition of the feed drive can be determined and explained by the spectral composition of the transmitted power. There is a direct correlation between a deteriorating condition of the feed drive and increasing vibrations. Once the damage has approached the final stage, the mechanical transmission behavior is massively changed by increasing vibration. Furthermore, it was possible to exactly assign to a cause of damage those individual frequencies in the spectrum of oscillations which were sustained by damage. With the findings thus gained, it is possible to describe the condition of a feed drive. The results are abstracted to a scalar parameter. With the established “vibration energy parameter”, an automated condition monitoring function for a single feed drive is provided.