Integral analysis and optimization of monitoring and inspection in aging structures: a Bayesian network approach
Applied Mechanics, Statics and Dynamics
Final Report Abstract
Deterioration processes in engineering structures lead to a reduction of service life and can affect the safety of the structure, the environment and people. To reduce the uncertainty in the deterioration process, inspection and monitoring are common practice in most engineering systems. The maintenance costs can however become high if not properly planned. This project focused first on developing a framework for following the evolution of a system reliability, as information is collected during its lifetime. A hierarchical Bayesian network (BN) model forms the basis of that framework. It models the probabilistic relationship between the components of a system, their deterioration process, and the performance of that system. Exact Bayesian inference is computationally powerful tool to include information from the components and reflect their impact on the system. In a second phase, the framework incorporates the optimization of inspection and maintenance strategies. The optimization was first carried out at component level without considering the component interactions. The component interdependence was then taken into account through the DBN framework, and the optimization performed exhaustively over a small number of strategies. Finally, the formulation of the problem in a restricted and parametrized strategy space allows for the implementation of powerful stochastic optimization methods, such as the cross entropy method combined with a noisy function approximator, such as kriging. This heuristic parametrization was outlined and illustrated. The choice of heuristics has been investigated at different levels. First, a component prioritization parameter was introduced, as a proxy to the value of information (VoI) of inspecting a certain component in a multi-component system. It demonstrated that the heuristic approach can model complex decision strategies and capture local effects in a set parameter. Secondly, the efficiency of the heuristic approach was analysed formally in a sequential decision problem for which the exact solution was available. Finally, incorporating continuous monitoring data has been the focus. Notably, the framework developed was utilized to model the information propagation from the components’ monitoring data to the system’s reliability, and evaluate the VoI of a monitoring system using the heuristic approach.
Publications
- (2013). Algorithms for optimal risk-based planning of inspections using influence diagrams. Proc. 11th International Probabilistic Workshop IPW11, Brno University of Technology, Czech Republic
Luque J., Straub D.
- (2015). Probabilistic Modeling of System Deterioration with Inspection and Monitoring Data using Bayesian Networks. Proc. 12th International Conference on Applications of Statistics and Probability in Civil Engineering, ICASP12, Vancouver
Luque J, Straub D
(See online at https://doi.org/10.14288/1.0076249) - (2016). Reliability analysis and updating of deteriorating systems with dynamic Bayesian networks. Structural Safety, 62, 34-46
Luque J., Straub D.
(See online at https://doi.org/10.1016/j.strusafe.2016.03.004) - (2016). Spatial Probabilistic Modeling of Corrosion in Ship Structures. ASCE-ASME Journal of Risk and Uncertainty in Engineering Systems, Part B: Mechanical Engineering, 3(3), 031001
Luque J., Hamann R., Straub D.
(See online at https://doi.org/10.1115/1.4035399) - (2017). Optimal prioritization of inspections in structural systems considering component interactions and interdependence. Proc. 12th International Conference on Structural Safety & Reliability, ICOSSAR 2017, Vienna
Bismut E., Luque J., Straub D.
- (2018). Adaptive direct policy search for inspection and maintenance planning in structural systems. Proc. Sixth International symposium on Life-Cycle Civil Engineering 2018, IALCCE 2018
Bismut, E., Straub, D.
- (2019). Direct policy search as an alternative to POMDP for sequential decision problems in infrastructure planning. Proc. 13th International Conference on Applications of Statistics and Probability in Civil Engineering, ICASP13
Bismut, E., Straub, D.
(See online at https://doi.org/10.22725/ICASP13.212) - (2019). Risk-based optimal inspection strategies for structural systems using dynamic Bayesian networks. Structural Safety, 76, 68-80
Luque J., Straub D.
(See online at https://doi.org/10.1016/j.strusafe.2018.08.002) - (2020). Reliability analysis of deteriorating structures, Structural Safety, 82: 101877
Straub D., Schneider R., Bismut E., Kim H.-J.
(See online at https://doi.org/10.1016/j.strusafe.2019.101877)
