Abstract

In modern data-centric production, multiple sensors, machines, equipment, and systems are connected providing a high variety and amount of data to enable insights into ongoing production. By adding more of those data sources to production, the Time Synchronization Problem becomes more intense when aggregating the data streams for analysis. To tackle the Time Synchronization Problem and perform the actual synchronization of the data, an estimation of the potential uncertainties is required. In this paper, we present a methodology to identify and describe the chain of time series uncertainties from data sources to aggregation. This calculation takes into account the uncertainties from sensory and signal processing requirements (like signal processing cycle times) to network communication latencies and jitter in the production network. In the end, each data point gets augmented with information regarding the uncertainty of its time information in terms of the minimum and maximum added processing times in the data pipeline. Using this information, the data aggregation can be adapted accordingly, and the final analysis tailored down to the actual result uncertainties. To show this method and its applicability, a use-case in battery cell manufacturing is presented including a description of how this method can be applied and the data pipeline adapted.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call