Holistic approach to machine tool data analytics

J. Lenz

doi:10.1016/j.jmsy.2018.03.003

2018-07-01

PRE

J. Lenz

Thorsten Wuest *

Engelbert Westkämper

DOI：10.1016/j.jmsy.2018.03.003

原文链接

原文求助

摘要

En 中文

Recent developments across all phases of the knowledge discovery process in the machine tool data analytics process call for a paradigm shift regarding how to combine the different analytics objectives. Several machine tool data analytics processes are carried out individually by different departments. They are highly dependent on the specific analytics objectives of the individual department. All these individual tasks make use of the data coming from the same source - the machine tool controller and connected sensors. One result of today's rather diverse machine tool data analytics landscape in many manufacturing companies is that we exhibit several pockets of expertise and large numbers of individual dedicated solutions. Hence, processes and structures tend be inefficient, e.g,, exhibit redundant processes, and the exchange between the different domains is difficult. Manufacturers face heavy competition for manufacturing experts, interested and qualified in data analytics. Therefore, it is in their best interest to utilizing this scarce resource as efficiently and effectively as possible, In this paper, we discuss the current situation exhibited in machine tool data analytics in manufacturing. Based on these insights, we propose a holistic approach to machine tool data analytics in order to tackle some of the identified shortcomings of current practices. We propose combining the tasks and bundling up analytics objectives across different departments and/or functions at the production line, factory or even the supply chain level. To evaluate our proposed approach, we provide selected implementation examples following the identified analytics objectives, including cross-domain analytics that focus on the interface between domains. Following, we critically discuss our proposed approach focused on the associated potential benefits, challenges and limitations. Lastly, we conclude the paper and provide an outlook on further research. (C) 2018 The Society of Manufacturing Engineers. Published by Elsevier Ltd. All rights reserved.

Keyword:

Industry 4.0

Smart manufacturing

Machine tool data

Data analytics

Machine learning

Machine tool controller