DeDiFlow

Against the backdrop of market individualization and the accompanying diversification of product ranges, manufacturing companies are faced more than ever with the challenge of being able to economically reflect high variability in their production. The occurrence of dynamic bottlenecks in the resulting production systems is now the rule rather than the exception. Managing these bottlenecks requires real-time analysis, evaluation, and the derivation of measures, which is technically achieved through adapted data mining methods and software support. In practice, however, the widespread use of these methods is hampered by limited user acceptance, a lack of metrics for bottleneck diagnosis, and an insufficient database to map the required fine temporal granularity of the production process. The potential of dynamic bottleneck management thus remains largely untapped.

The DeDiFlow research project aims to counteract both practical restrictions of dynamic bottleneck management. On the one hand, the development of a system for retrofitting buffer level monitoring based on low-cost sensor technology for small and medium-sized enterprises (SMEs) will close the data gap and combine this novel data collection method with the further development of existing congestion detection methods. Second, the project supplements the missing procedures and metrics for bottleneck diagnosis in order to derive and justify intervention measures applicable to SMEs. In addition to evaluating the frequency and severity of occurrences, the development will therefore also include the derivation of suitable improvements, taking into account the operating conditions in manufacturing SMEs.

In order to achieve the research objectives, requirements and success criteria will first be defined in consultation with industrial users through expert interviews. For data collection, the technical infrastructure must be set up, including the implementation of a standardized supply from the sensor kit. The subsequent methodological work comprises the detection method and methods and metrics for diagnosis, which are developed and validated in interaction.

For the participating SMEs, however, the application must not end with detection and diagnosis. Based on the results of the methods, improvement measures are derived, tested, and evaluated, always with the aim of accelerating the cycle of dynamic bottleneck management. The application-oriented development in cooperation with industry partners is flanked by a demonstrator development that presents use cases in a presentable manner and can contribute significantly to the transfer of solutions.

The research project "DeDiFlow" is funded under the funding number 01IF23719N by the Bundesvereinigung Logistik e.V. by the Federal Ministry for Economic Affairs and Energy (BMWE) based on a resolution of the German Bundestag as part of the program to promote industrial collaborative research (IGF).