Refine
Year of publication
- 2020 (4) (remove)
Document Type
- Journal article (4)
Language
- English (4)
Has full text
- yes (4)
Is part of the Bibliography
- yes (4)
Institute
Publisher
- Elsevier (4) (remove)
Driven by digital transformation, manufacturing systems are heading towards autonomy. The implementation of autonomous elements in manufacturing systems is still a big challenge. Especially small and medium sized enterprises (SME) often lack experience to assess the degree of Autonomous Production. Therefore, a description model for the assessment of stages for Autonomous Production has been identified as a core element to support such a transformation process. In contrast to existing models, the developed SME-tailored model comprises different levels within a manufacturing system, from single manufacturing cells to the factory level. Furthermore, the model has been validated in several case studies.
Process quality has reached a high level on mass production, utilizing well known methods like the DoE. The drawback of the unterlying statistical methods is the need for tests under real production conditions, which cause high costs due to the lost output. Research over the last decade let to methods for correcting a process by using in-situ data to correct the process parameters, but still a lot of pre-production is necessary to get this working. This paper presents a new approach in improving the product quality in process chains by using context data - which in part are gathered by using Industry 4.0 devices - to reduce the necessary pre-production.
In recent years, machine learning algorithms have made a huge development in performance and applicability in industry and especially maintenance. Their application enables predictive maintenance and thus offers efficiency increases. However, a successful implementation of such solutions still requires high effort in data preparation to obtain the right information, interdisciplinarity in teams as well as a good communication to employees. Here, small and medium sized enterprises (SME) often lack in experience, competence and capacity. This paper presents a systematic and practice-oriented method for an implementation of machine learning solutions for predictive maintenance in SME, which has already been validated.
The use of learning factories for education in maintenance concepts is limited, despite the important role maintenance plays in the effective operation of organizational assets. A training programme in a learning factory environment is presented where a combination of gamification, classroom training and learning factory applications is used to introduce students to the concepts of maintenance plan development, asset failure characteristics and the costs associated with maintenance decision-making. The programme included a practical task to develop a maintenance plan for different advanced manufacturing machines in a learning factory setting. The programme stretched over a four-day period and demonstrated how learning factories can be effectively utilized to teach management related concepts in an interdisciplinary team context, where participants had no, or very limited, previous exposure to these concepts.