Structural Integrity and Risk Management based on Value of Information and Action Analysis

Arifian Agusta

Research output: Book/ReportPh.D. thesisResearch

126 Downloads (Pure)

Abstract

Structural integrity and risk management (SIRM) play an important role in ensuring safe, cost efficient and sustainable operation of structural systems. Operators and owners of structures can have different objectives and numerous decision alternatives when planning a SIRM strategy. During SIRM planning, uncertainties related to structural performance and structural health information modeling need to be accounted for along with the costs and benefits associated with the decision alternatives. Bayesian decision analysis facilitates the analysis of such decision problems to aid operators/owners (based on the costs, risks, and benefits of SIRM strategies) at the start or during the service life.
This thesis proposes a decision theoretical framework for planning an optimal SIRM strategy by utilizing extended Bayesian decision analysis. The value of a SIRM strategy is quantified using the value of information and action (VoIA) analysis, to determine the difference between the expected total costs with and without SIRM. Comparing the VoIA of several SIRM scenarios allows the optimal SIRM strategy to be identified by maximizing VoIA.
The thesis starts with a review of state-of-the-art inspection, structural health monitoring (SHM), and damage detection system (DDS) performance modeling and planning in the context of SIRM planning. A brief description of classical and extended Bayesian decision analysis is presented along with several examples. A study of inspection and repair planning for offshore structures using Bayesian decision analysis is performed to find the optimal inspection and repair time and to derive the optimal decision rule for inspection planning. Modeling approaches for inspection and DDS are presented along with a novel SHM information modeling approach, in which the outcome of SHM is modeled in conjunction with the realizations of the model uncertainty. By defining two thresholds, the outcomes can be distinguished into three events: high performance, as expected/normal performance, and low performance. The value of single and multiple information acquirement systems is quantified by utilizing VoIA analysis and the effects of measurement uncertainty and model uncertainty bias are discussed. A software tool called VoICalc was developed to perform decision analyses for moderately complex decision problems and to calculate the VoIA of inspection, SHM, and DDS. A decision theoretical framework for SIRM planning that facilitates multiple structural health information sources (inspection, SHM, and DDS) in a complex SIRM scenario is introduced and applied to a generic redundant structural system. The results show that information from SHM and DDS can have a significant impact on future inspection efforts and can enhance the VoIA of SIRM compared to a SIRM scenario with only inspections. The developed framework is also applied to plan an optimal SIRM strategy for a typical deepwater offshore structure, by utilizing a specific complex system modeling in conjunction with response surface for load modeling and nonlinear finite element model.
It is demonstrated that the developed decision theoretical framework can be used to quantify and optimize the value of SIRM. The framework facilitates complex SIRM scenarios and provides decision support for planning an optimal SIRM strategy before or during the service life. With several case studies, it is shown that the developed framework can be applied to generic redundant and complex structural systems. The developed framework also can be used to find the optimal information acquirement type for a specific structural performance model. Limitations and outlook for further studies are discussed in the final chapter of the thesis.
Original languageEnglish
PublisherTechnical University of Denmark, Department of Civil Engineering
Number of pages242
Publication statusPublished - 2020

Fingerprint

Dive into the research topics of 'Structural Integrity and Risk Management based on Value of Information and Action Analysis'. Together they form a unique fingerprint.

Cite this