ZEKISS
Research Project

Condition assessment of railway bridges and vehicles with AI methods for the evaluation of sensor data and structural dynamic models

The ZEKISS project, funded by the BMVI within the mFUND research initiative, deals with the interdisciplinary evaluation of sensor data in railroad traffic to record and evaluate the condition of the components vehicle and bridge. A predictive maintenance approach is being developed using methods of artificial intelligence.

The interdisciplinary evaluation of data from sensor systems in railway traffic opens up new possibilities for the recording and evaluation of the condition of the components vehicle and structure. So far, data from monitoring systems on structures (mainly bridges) have already been used to carry out structural assessments and maintenance measures. Similarly, sensor systems installed on vehicles are used to monitor vehicle components. Inherently, however, the respective sensor systems are used to collect data containing further information: An instrumented train crosses a large number of bridges on its journey and a large number of trains cross an instrumented bridge. If an instrumented train crosses an instrumented bridge, both sensor systems can be calibrated; in the further course of the process, information about other, non-instrumented trains can be collected from the instrumented bridge or information about other non-instrumented bridges can be collected from an instrumented train that it crosses.

For this purpose, a digital tool for the in-situ monitoring of railway bridge structures is to be developed, implemented and validated within the framework of a sensor-based predictive maintenance concept using a BIM-integrated digital twin and based on artificial intelligence. The core objective is the development of a template for a highly automated and improved statement for the condition assessment (resonance hazard, structural safety and remaining service life) of existing railway bridges. For this purpose, the condition data obtained by dynamic monitoring shall be considered by an automated and continuous update of the actual condition of the Digital Twin. Based on the status data, an artificial intelligence, taking into account mechanical relationships of the structure, then decides on a necessary adaptation of the structural models to the actual state.

The project, funded by the BMVI, is coordinated by the Institute of Structural Dynamics and Design (ISM+D) of FB 13, TU Darmstadt. Scientific partners are the Institute of Numerical Methods and Informatics in Civil Engineering (IIB) of TU Darmstadt, Wölfel Engineering GmbH + Co KG from Höchberg, DB Netz AG from Frankfurt a. Main and GMG Ingenieurgesellschaft mbH from Dresden.

  Name Contact
Jascha Brötzmann M.Sc.
+49 6151 16-21338

ZEKISS is funded by the Federal Ministry of Transport and Digital Infrastructure based on a resolution of the German Bundestag.

Funding reference number: 19F2123A
Duration: 05/2020 – 04/2023