Development of Smart Component Based Framework for Dynamic Reliability Analysis of Nuclear Safety Systems

Dynamic reliability methodologies account for the safety system's time dependent characteristics while estimating the reliability. Time dependence can arise due to interdependence of process conditions and hardware states and action of control element. Though static reliability models often capture the average behavior and try to make conservative estimates, it is inadequate from a number of perspectives. First, this requires that the analyst needs to establish that the model is conservative. Second, such modeling requires more expertise and experience in the appropriate domain of the problem, rather than in the reliability methods. Third, approximate methods may be inadequate to establish reliability enhancements or degradations due to subtle alterations in the system design. In spite of the significant effort in the reliability community to establish dynamic reliability analysis methods, there are no general purpose tools similar to that available for fault tree event tree modeling. In this regard Smart Component based method is identified as a suitable candidate for general purpose dynamic reliability assessment and developed for implementation. Smart Component based dynamic method uses elements of object oriented architecture and Monte Carlo simulation and, is suitable for being developed into a general purpose tool. The paper demonstrates the capability of the method to evaluate reliability of systems having various types of time dependence, interaction between hardware failure and process evolution and complexity by means of few case studies. The method is found to be promising for accurate modeling of dynamic as well as static scenarios. (author)