Nicht-invasive Maschinenstatus- und Prozesserkennung

Presentation of a Novel Approach. Modern machining centers usually have numerous sensors, which can be used to determine characteristic values for evaluating productivity. Depending on the manufacturer and machine type, however, the amount and quality of the available data vary. This paper presents a non-invasive system that circumvents this heterogeneity. Both the machine status and processes executed can be recognized and classified. By segmenting the signals and extracting characteristic features, the executed process is recognized. In a subsequent step, this should enable the determination of productivity parameters.