Modeling of Electrical Cable Failure in a Dynamic Assessment of Fire Risk

摘要

Fires at a nuclear power plant are a safety concern because of their potential to defeat the redundant safety features that provide a high level of assurance of the ability to safely shutdown the plant. One of the added complexities of providing protection against fires is the need to determine the likelihood of electrical cable failure which can lead to the loss of the ability to control or spurious actuation of equipment that is required for safe shutdown. A number of plants are now transitioning from their deterministic fire protection programs to a risk-informed, performance based fire protection program according to the requirements of National Fire Protection Association (NFPA) 805. Within a risk-informed framework, credit can be taken for the analysis of fire progression within a fire zone that was not permissible within the deterministic framework of a 10 CFR 50.48 Appendix R safe shutdown analysis. To perform the analyses required for the transition, plants need to be able to demonstrate with some level of assurance that cables related to safe shutdown equipment will not be compromised during postulated fire scenarios.;This research contains the development of new cable failure models that have the potential to more accurately predict electrical cable failure in common cable bundle configurations. Methods to determine the thermal properties of the new models from empirical data are presented along with comparisons between the new models and existing techniques used in the nuclear industry today. A Dynamic Event Tree (DET) methodology is also presented which allows for the proper treatment of uncertainties associated with fire brigade intervention and its effects on cable failure analysis. Finally a shielding analysis is performed to determine the effects on the temperature response of a cable bundle that is shielded from a fire source by an intervening object such as another cable tray.;The results from the analyses demonstrate that models of similar complexity to existing cable failure techniques and tuned to empirical data can better approximate the temperature response of a cables located in tightly packed cable bundles. The new models also provide a way to determine the conditions insides a cable bundle which allows for separate treatment of cables on the interior of the bundle from cables on the exterior of the bundle. The results from the DET analysis show that the overall assessed probability of cable failure can be significantly reduced by more realistically accounting for the influence that the fire brigade has on a fire progression scenario. The shielding analysis results demonstrate a significant reduction in the temperature response of a shielded versus a non-shielded cable bundle; however the computational cost of using a fire progression model that can capture these effects may be prohibitive for performing DET analyses with currently available computational fluid dynamics models and computational resources.