In last decades maintenance optimisation of civil structures has gained increasing attention since the number of ageing structures is becoming large while available budget dedicated to maintenance is limited. Fatigue is one of the main degradation processes on steel structures that causes structural failure before the designed lifetime. Therefore, the main goal of our study is to provide an optimal maintenance planning to extend the lifetime of structures against fatigue. Structural health monitoring can help to have a better understanding of the structural condition for a better maintenance funding allocation. To reach our goal we might face different challenges in A: fatigue reliability assessment, B: application of monitoring information, and C: optimisation of maintenance strategies. Addressing these challenges can define different steps of this study.
Within the first step, fracture mechanism and S-N curves are two regular approaches that are used to evaluate fatigue damage and provide a proper limit state for reliability analysis. Fatigue reliability can be performed within a time-independent or time-dependent framework depending on the choice of fatigue model. The challenge here is related to how to evaluate low probabilities of failure when computationally expensive performance functions are involved. Hence, we have developed an efficient methodology to perform time-dependent reliability analysis. Kriging meta-modelling is used to replace the computationally expensive performance functions while using Monte Carlo simulation.
The second step is related to how to get benefit from monitoring information. Structural health monitoring provides us with valuable information about the current situation of civil structures. This information can be about crack development, actual loading conditions, etc. The challenge here is related to the way we employ the information coming from monitoring data. With this respect, a study has been developed on long-term monitoring data available at EPFL. Time series methods are employed to prepare a loading model for long-term monitoring data. The aim is to capture seasonality effect in traffic loading and to provide a model that gives more detail about structural fatigue loading. This load model can be used within S-N approach or fracture mechanism with some adjustments.
Finally, methods and approaches in previous steps will be combined to have a proper and updated indicator to apply maintenance actions. The desired indicator would be fatigue reliability index. The challenge in the final step can be related to the cost models for maintenance and monitoring. Therefore, an appropriate cost model will be chosen for the optimisation framework, and overall methodology will be applied on a shared object or another case study.
Thank you very much for reading this post.