Availability‐guaranteeing maintenance of series machine tools. Issue 7 (14th September 2021)

• Record Type: Journal Article
• Authors: Praedicow, Michael; Apitzsch, René; Richter, Anja; Wunderlich, Tim; Klimant, Philipp
• Other Names: Gude Maik guestEditor.; Klimant Philipp guestEditor.; Putz Matthias guestEditor.; Rafaja David guestEditor.; Weck Danie guestEditor.; Wuestefeld Christina guestEditor.
• Is Part Of: Engineering reports. Volume 4:Issue 7/8(2022)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2021-09-14
• Subjects: augmented reality -- condition monitoring -- condition‐based maintenance -- linked data -- machine tool -- predictive maintenance; Engineering -- Periodicals; Computer science -- Periodicals; 620.005
• DOI: 10.1002/eng2.12456
• Languages: English
• ISSNs: 2577-8196

Abstract

Abstract: Condition monitoring enables transparency, as for example irregularities are detected automatically and reported. A condition forecast, however, requires more. In contrast to AI black box methods, frequently used in this context, a combination of existing expert knowledge and classical statistics is used as a method for a reliable determination of the remaining component‐lifetime. This works, if meaningful historical data are available in a sufficient quantity and quality. And this in turn requires a corresponding number of machines that are as identical in construction as possible and which must also be subject to a defined test regime, temporally closely monitored outside the production process. However, the quantity can be significantly smaller than the number of cases required for a prescient analysis of the correlations between a condition parameter and the wear condition of a specific component. The main target audience of the strategy presented here is therefore in particular manufacturers of series machines who wish to offer maintenance packages with corresponding availability guarantees and on‐site support for maintenance personnel. Abstract : The negative effects of purely reactive or preventive maintenance can be substantially reduced by objectively evaluating the current wear condition based on control and sensor data, combined with statistical or model‐based prediction of future failure probability. In addition, the cost reduction is supported by front‐end visualization of the assembly conditions, including derivations of appropriate instructions for immediate action and maintenance.