The progressive adoption of IT systems and machine digitization lead to the creation of ever greater quantities of data in the manufacturing domain. Simultaneously, the complexity of shop-floor operation and control is increasing while product life cycles shorten, and customer expectations rise. Data-driven methods promise to mitigate the increased complexity while realizing potentials for service improvement, cost reduction and energy savings. The sophisticated application of such methods is becoming a decisive factor for the competitiveness of manufacturing firms.
The AI-based Process Analytics and Control on the Shop Floor (AIPACS) project aims to increase shop floor control by applying AI-based methods to make use of the heterogenous data sources of the shopfloor (i.e., production line sensors and IT-Systems). The concept of the research project is shown on the right. By integrating the necessary data into a data-lake the processability of the differing data sources is ensured. After a pre-processing step, AI- and other data-driven methods are applied in the production process analytics step. The resulting models and algorithms are then used in the monitoring step. If a process relevant deviation is predicted or detected, the deviation is either corrected by the system itself or an employee gets notified with a correction task. In both cases the correction measure will lead to adjusted parameters which are observable in the data sources. The observed adjustments can then be used as a feedback loop to enhance the continuous cycle of analysis and monitoring.
Researcher
Partner
REHAU Industries SE & Co. KG