The project focused on improving the efficiency and reliability of inland shipping transport chains through a systematic analysis of processes, faults, data, and predictive modelling. In the process and fault analysis work package, the team first defined a suitable methodology and produced documentation to capture operational and information flows. Expert interviews with project partners and other stakeholders were conducted to gather detailed insights, which were then structured into a qualitative overall model of the transport chain using BPMN. Concurrently, the influence of potential disturbances was examined through a Failure Mode and Effects Analysis (FMEA). Additional interviews assessed the impact of these disturbances, and the findings were documented and integrated into a comprehensive systematics using reliability engineering tools such as Ishikawa diagrams and influence matrices. The team prioritized disturbances by strength and frequency, establishing clear definitions for those factors that required special attention in subsequent solution development. This work laid a solid foundation for further optimisation of the transport chain.
In the data analysis work package, the consortium identified suitable data types and sources in collaboration with data workshops. Confidentiality requirements were defined jointly with TU Berlin, and measures were implemented to ensure compliance with data protection regulations. The team provided data extracts and, where necessary, live interfaces to identified data sources. A key enhancement was the integration of additional ship movement records alongside AIS data, giving partners more comprehensive and accurate information for their projects. These efforts ensured that partner data needs were met and that the necessary resources were available for successful project execution.
The model development – forecasting work package aimed to optimise process time predictions for inland shipping. The team defined clear sub‑forecasting problems, verified their completeness against real transport processes, and identified special cases and limits of the chosen modelling approach. Machine learning techniques were evaluated for suitability in a future live deployment, with a critical assessment of their practical applicability. The forecasting models were designed to be integrated into a live system, providing real‑time predictions that could inform operational decisions.
Decision support modelling was addressed in a separate work package, where the team developed tools to assist stakeholders in evaluating alternative actions based on the forecasted outcomes. The live implementation concept was then formulated, outlining how the developed models and interfaces would be deployed in a production environment. Dissemination activities ensured that findings were communicated to the wider community through reports, workshops, and publications, thereby extending the impact beyond the consortium.
Collaboration was organised around eight work packages, with TU Berlin taking a leading role in project management, documentation, and coordination of the consortium’s activities. The Deutsche Binnenreederei GmbH contributed by providing progress reports, relevant information, and reviewing the produced documents, as well as facilitating communication with interested parties. Other partners, though not named explicitly in the excerpt, supported specific work packages such as data provision, model development, and live implementation. The project was carried out over its scheduled duration, with regular reporting to the project sponsor, which is implied to be a German federal funding body. The consortium’s joint effort resulted in a comprehensive set of analytical tools, data interfaces, and predictive models that collectively aim to enhance the operational performance of inland shipping transport chains.