A Digital Perspective on Machine Tool Calibration

Abstract

Machine tool calibration and subsequent controller-based compensation are industrially established and research-intensive techniques used to monitor and increase the volumetric performance in high-precision manufacturing. Moreover, a variety of interim performance checks and integrated sensor approaches have been developed to predict volumetric performance degradation and avoid an economically undesirable downtime. However, the fragmentation of data acquisition and management limits the potential for additional insights with respect to the value creation based on existing methods in the field of machine tool calibration. The authors reviewed the former from the perspective of data sources according to the frequency of their contribution to the digital twin of a machine tool, adopting a digital view regarding machine tool calibration within the Internet of Production concept. The latter proposes a semantic separation of cyber physical production systems into four layers: data sources, data access and provisioning, storage and analytics, and user respective agent feedback. To achieve a common representation across different layers, devices, and industrial Internet protocols, a model-based abstraction layer is required, which must be compatible with existing standards within the field, e.g., the ISO 230 series. Utilizing different Internet of Production architectures and platforms, a multitude of parallel analytic applications and an evaluation of complex models are enabled owing to the availability of ample computing resources, among which the machine tool’s numerical controller takes the role of an edge-device injecting the feedback into the production process. A proof-of-concept of a digital approach to machine tool calibration data storage and processing was established based on the software prototype VoluSoft, which implements an ISO 230-1:2012 based abstraction layer in JavaScript Object Notation format, and an evaluation of the kinematic models to estimate the volumetric performance at the functional point. Apart from generating compensation tables, the results are used to project the expected deviation at the tool tip to the computer-aided design-model of a work piece, correlate the error motions using the temperature data acquired by integrated sensors, and estimate the contribution of the volumetric performance limitation to the uncertainty budget of on-machine measurements.