Skip to Main content Skip to Navigation

Multi-Hazards Risk Aggregation Considering the Trustworthiness of the Risk Assessment

Abstract : This PhD thesis addresses the problem of Multi-Hazards Risk Aggregation (MHRA), which aims at aggregating the risk estimates from Probabilistic Risk Assessment (PRA) models for the different contributors. The current practice of MHRA is based on a simple arithmetic summation of the risk estimates. However, the risk estimates are obtained from PRA models that have different degrees of trustworthiness, because of the different background knowledge they are based on. Ignoring this difference in MHRA could lead to misleading results for Decision-Making (DM). In this thesis, a structured framework is proposed to assess the level of trustworthiness, which risk assessment results are based on and to integrate it in the process of MHRA. The original scientific contributions are: (i) Factors contributing to the trustworthiness of risk assessment outcomes are identified and their criticalities are analyzed under different frameworks, to understand their influence on the risk results; (ii) An integrated hierarchical framework is developed for assessing the trustworthiness of risk analysis, based on the identified factors and related attributes; (iii) A reduced order model-based method is proposed to efficiently evaluate the trustworthiness of risk assessment in practice. Through the reduced-order model, the proposed method can limit the number of elements considered in the original risk assessment; (iv) A technique that combines Dempster Shafer Theory and the Analytic Hierarchy Process (namely, DSTAHP) is applied to the developed framework to assess the trustworthiness by a weighted average of the attributes in the framework: the AHP method is used to derive the weights of the attributes and the DST is used to account for the subjective uncertainty in the experts’ judgments for the evaluation of the weights; (v) A MHRA technique is developed based on Bayesian model averaging, to overcome the limitations of the current practice of risk aggregation that neglects the trustworthiness of the risk assessment of individual hazard groups; (vi) The developed framework is applied to real case studies from the Nuclear Power Plants (NPP) industry.
Document type :
Complete list of metadatas

Cited literature [291 references]  Display  Hide  Download
Contributor : Abes Star :  Contact
Submitted on : Saturday, May 2, 2020 - 1:05:42 AM
Last modification on : Tuesday, October 20, 2020 - 10:20:37 AM


Version validated by the jury (STAR)


  • HAL Id : tel-02560550, version 1


Tasneem Bani-Mustafa. Multi-Hazards Risk Aggregation Considering the Trustworthiness of the Risk Assessment. Risques. Université Paris Saclay (COmUE), 2019. English. ⟨NNT : 2019SACLC096⟩. ⟨tel-02560550⟩



Record views


Files downloads