|Project head||Prof. Dr.-Ing. R.
associate(s)||Dr.-Ing. E. Matta, A. Hassan, MSc
assistant||C. Scheer, BSc, Y. Ma,
|Project period||September 2017 - December
|Funding||Federal Ministry of
Transport and Digital Infrastructure (BMVI), mFund research
GmbH, BearingPoint Technologie GmbH, Federal Institute of
Hydrology (BfG, Koblenz)|
|Link to the project||www.bmvi.de/SharedDocs/DE/Artikel/DG/mfund-projekte/entwicklung-digitaler-schifffahrtsassist-dsa.html
The current German Federal Transport Infrastructure Plan (Federal Ministry of Transport and Digital Infrastructure - BMVI) expects a growth of about 23% of inland navigation traffic between 2010 and 2030. With the infrastructures at the state of art, a more efficient use of the existing waterways is required. For this purpose, new approaches should be proposed especially to counteract undesired congestions on rivers and canals.
Project objective and implementation
- © Q. Zhang
The focus of the research project is the
development of an optimized water level prediction model, whose routes
and load limits can be determined in a fast and efficient way. The
model aims to allow reliable multi-day forecasts of water levels for
the expected journey time. The Digital Skipper Assistant (DSA) should
be demand-oriented in respect to the requirements for route- and
cargo-planning within inland waterways.
While the project partners are predicting water levels based on weather forecasts and hydrological model chains, it is the goal of the TU Berlin to produce such predictions using artificial neural networks (ANN). The basic idea consists of the determination of a downstream water level based on one or more measured upstream levels and possibly other variables such as precipitation measurements, provided by the partner Federal Institute of Hydrology (BfG). The concept of ANN is based on the creation of a main network with several sub-networks. The main network begins with a gauge PA, with a long-term data series of water levels available, and ends at the last investigated gauge station in the Rhine (PX). The sub-network for a target gauge station PD is constructed on the data basis of the previous water levels in PC and possibly further upstream level stations and/or further variables, such as precipitation measurements. The ANNs are usually trained with approximately 70-80% of the available data and validated with the remaining 20-30% of the data that was not used for training. In a future phase, the ANN and the hydrological models can be also combined. The focus of this specific work conducted by TU Berlin is on the Rhine, being a river of great importance for inland navigation in Germany.
The research project is funded by the Federal Ministry of Transport and Digital Infrastructure (BMVI) and is led by BearingPoint GmbH, in cooperation with its partners Federal Institute of Hydrology and BearingPoint Technologie GmbH. Other players in inland navigation, such as inland shippers, port operators, freight forwarders, as well as water and shipping administrative authorities (e.g. WSV) are also involved.
Further information about the project .
Catchment area of the Rhine up to gauge Emmerich, linking the hydrological and hydro-dynamic model components of the traffic-related prediction system of the Federal Institute of Hydrology (BfG) for the Rhine waterways.
- © Federal Institute of Hydrology - BfG