Project Details
Projekt
Structural Health Monitoring with model based damage detection using nonlinear model adaption and Artificial Intelligence methods
Subject Area
Structural Engineering, Building Informatics and Construction Operation
Term
since 2022
Project identifier
Deutsche Forschungsgemeinschaft (DFG) - Project number 501496870
Condition evaluation and predictive maintenance are essential for extending the lifespan of highly loaded structures. In addition to manual inspections, sensor-based monitoring can be integrated into data-driven monitoring concepts, enabling continuous and objective detection of the structural conditions. Based on the sensor data, damage detection, a reliable forecast of the load-bearing behavior and possible recommendations for prescriptive maintenance can be provided. For highly stressed massive structures such as bridges, sensor-based monitoring is associated with considerable challenges. In addition to the complex load-bearing behavior, changing loads due to heavy traffic and temperature stress must also be considered for automated long-term monitoring. The application of artificial intelligence (AI) methods in a model-based long-term monitoring provides a promising approach for analyzing sensor data and adapting the load-bearing of the structure. The objective of the research project is to develop an approach for automated damage detection and prediction of material behavior for open-air reinforced concrete structures. In addition to system and load identification by adapting a numerical model to the structure based on the static response, the developed approach also includes the prediction of material degradation and the automated output of recommendations for maintenance strategies. The numerical model is adapted in an optimization process based on Evolutionary Algorithms (EA), considering the complex material behavior of the material within nonlinear finite element (FE) simulations. To increase the reliability of the adaptation and considering uncertainty of the model parameters, identified models are then assigned to uncertain clusters via cluster analysis to characterize possible damage mechanisms. Damage detection is carried out by comparing identified system parameters at different measurement time-steps. By interpreting the identified system parameters and implementing degradation models, a prognosis of the load-bearing behavior is provided. Based on the prognosis, damage detection and analysis of the sensor data, recommendations for action and maintenance strategies for extending the service life are proposed. The developed methods are implemented for the assessment of the conditions of bridge structures. Model-based analyses of a prestressed concrete bridge are carried out to evaluate the model adaptation of the developed approach. Damage detection and the prognosis model as well as the method for issuing recommendations for predictive maintenance are evaluated as part of destructive load tests of a research bridge.
DFG Programme
Priority Programmes
