WP2 - Structural performance prediction by a multifunctional beam and plate FE model
Main researcher: ir. Andreas DeKeyser
Affiliation: Ghent University
Supervisors: Prof. dr. ir. Robby Caspeele, dr. ir. Roman Wan-Wendner, prof. dr. ir.-arch. Els Verstrynge
Publicatons: link
Work package description
When designing new structures, common beam and plate finite element models are used to calculate the structure’s deformations and internal forces. For existing structures however, such models are mostly absent. Hence, in this part of the research, a new multifunctional beam and plate FE model is developed for the assessment of existing concrete structures. This model will start from an existing FE model (the Direct Stiffness Method), and will extend it to account for time-dependent damage and degradation effects. Like this, the structural performance of existing corroded structures can be predicted in time.
In order to make these predictions, incorporating information about the condition of the structure is crucial. Hence, the model will be able to use all types of information from inspections and monitoring to predict the current condition of the structure as accurately as possible. Bayesian techniques will be used to deal with this, including the spatial and temporal variation of this information. The framework will then be further extended towards a probabilistic calculation method, enabling reliability-based performance predictions.
Lastly, the developed model will be validated both numerically and experimentally.
Research results
The Direct Stiffness Method (DSM), specifically for reinforced and prestressed concrete beams and reinforced concrete slabs, was extended to take into account the effects of corrosion of the steel. Moreover, damage effects such as steel cross-section reduction, steel ductility reduction, bond deterioration, and corrosion-induced concrete cracking can be accounted for in order to predict the structural response of existing elements.
The DSM-tool was also extended to incorporate additional information through Bayesian updating. This way, results from WP1 and WP3 are incorporated, together with additional information from result experimental data (e.g. visual inspection reports, compressive strength tests on drilled cores, strain measurements during a static loading test). Furthermore, the spatial character of material properties and corrosion parameters and hence also the influence of the location of a measurement on the posterior state of knowledge may also be accounted for in the DSM-tool. This enables the prediction of the structural performance of an existing element with significantly reduced uncertainty.
The DSM-tool is applied to specific real-life cases. An existing prestressed girder of a bridge in the Netherlands was structurally evaluated using chloride concentration measurements over the depth of the profile. These measurements were used to update the chloride initiation characteristics through Bayesian updating. Afterwards, the reliability index of the bending moment capacity over time was determined to evaluate the remaining lifetime with a prior and posterior state of knowledge.
The predictive ability of the DSM-tool for corroded reinforced concrete slabs is evaluated using experimental results from the literature. The load-displacement curves and the obtained crack patterns are compared and it is shown that the DSM-tool agrees well with the experiments. Future work includes the failure mode of punching shear, which is not predicted in the current model yet.
main objectives
This workpackage aims for the following Deliverables (D) and Milestones (M):
- D2.1 – Algorithm for matrix-based extension of the direct stiffness algorithm to account for the time-dependent and spatial character of cross-sectional properties
- D2.2 – Algorithm for Bayesian updating of Direct Stiffness Method framework
- D2.3 – Validation report of developed beam and plate FEM by numerical nonlinear FEM
- D2.4 – Algorithm for structural reliability calculations on the basis of the developed Direct Stiffness Method