Options
2024
Journal Article
Title
Evaluation of Offline Data Synchronization Approaches in data-intense Manufacturing
Abstract
In smart manufacturing, the usage of data is crucial to enable data-based insights, analytics, and optimizations. Quality and performance of those data-based services directly depend on the quality of the data from the production, making the management of data quality a critical step in the smart manufacturing pipeline. One data quality feature is the synchronicity of the time-series data sets linked with each other. While most approaches for manufacturing data synchronization target the synchronization of clocks in the data acquisition and sensing equipment, those are not applicable when adding and merging further data sets afterward or in case some devices do not support online synchronization. Therefore, offline synchronization has to be used for the synchronization after the actual production, which is often performed by manual labor. In this paper, we deduce metrics for evaluating the quality of offline synchronization methods as well as their applicability depending on the data formats and production setup provided. This also includes the aspect of scalability for the use in productions from low to high amounts of time series data sources. Using these metrics, we evaluate existing approaches for offline synchronization showing the most likely applicable methods depending on the scenario in comparison to the manual labor baseline. The findings show, that for the observed approaches for offline synchronization, the metrics can be used. These metrics show the differences between the approaches and support the decision on the approach to use for a specific use case. Still, the use of those shows a shortcoming of approaches for offline synchronization, as none of the observed approaches is capable of performing great for all the metrics.
Author(s)
Conference