Limitations of explicit modeling of common cause failures within fault trees

The significant impact that common cause failures can have on reliability and safety of a system comprising redundant components is widely acknowledged. These common cause failures can significantly endanger the benefits of the redundancy which is seen as the main principle upon which safety systems design is based. Thus, consideration of common cause failures is one of the most challenging and critical issues in the probabilistic safety assessment (PSA). This is especially emphasized within PSA fault tree modeling of safety systems within nuclear power plants. This study presents a new method for explicit modeling of single component failure in different common cause component groups (CCCGs) as well as the associated limitations. These limitations arise from the principle of explicit modeling of common cause failures where each components failure space can be decomposed, in terms of causes, to independent portion and dependent, common cause failure affected portion. A method-complementary approach for acting upon these limitations, presented in terms of constraining the dependent portion of the total components failure space, is proposed within the paper. The method is applied on a selected case study system. The results and insights gained out of this application are presented and discussed. In general, the application of this method implicates improved and more detailed PSA models. These improved models consequently direct more realistic PSA results.