Application of probabilistic risk assessment (PRA) methods has become a powerful tool in making decisions relative to safety of various plant components. In the nuclear industry, the use of this concept has been endorsed by the USNRC in Regulatory Guide 1.174. As the emphasis on efficiency while maintaining safety continues, methods beyond the standard deterministic approaches are required to obtain the right information needed to make optimum decisions. SI has been extensively involved in development and application of probabilistic methods to address complex engineering issues in various industries including the nuclear and fossil power industries.

SI has the capability to perform full power and shutdown PRA for both BWRs and PWRs in the areas of Level 1 and Level 2.  In the Level 1 analyses, core damage is generally defined as fuel cladding temperatures reaching a value of 2200°F, above which point water-metal reaction begins to increase and cladding failure may occur.  Analysis of the containment response to scenarios postulated in the Level 1 assessment and determination of the source terms are defined as the Level 2 phase of a PRA project.  The magnitude, energy content, composition, and timing of radionuclide release are defined as the source term.  PRA capabilities include fault trees development, core damage and containment event trees development, Core Damage Frequency (CDF) calculations following postulated accidents during full power and shutdown conditions using CAFTA and other similar codes.  SI is capable of using severe accident analysis codes such as MAAP to determine the thermal-hydraulic behavior of the reactor vessel, reactor core, drywell, and containment and to calculate the radionulide releases into the containment and the environment.

Another key part of PRAs is Probabilistic Fracture Mechanics (PFM) in which many of the key variables are represented using appropriate statistical distributions. The essence of the PFM approach is to set up an engineering problem and assign statistical distributions to each key variable affecting the problem. Solution algorithms are set up as if each variable were a known, deterministic parameter. The algorithms are exercised repeatedly, randomly selecting a different value for each variable for each iteration. In PFM, each iteration results in a failure or non-failure. Probability of failure is simply the total failures divided by the total iterations.

SI has used PRA plus an understanding of component specific degradation mechanisms to establish effective plant integrity management programs which maintain safety while at the same time reducing the industry and regulatory burden associated with current in-service inspection requirements. This has been applied to inspection of vessels, piping and other plant components. SI has also used PRA to help plants make run/repair/replacement decisions on key components. SI has developed several computer programs for PRA application. A sampling of these programs include:

Program List:

SI’s expert capabilities using our sophisticated probabilistic based software can be applied to many types of complex problems in various industries where key parameters can be expressed as a distribution. SI has been successful in application of PRA type evaluations to help solve critical issues for our clients in a cost-effective manner.

Related Technical Literature·

• Operating Experience Regarding Thermal Fatigue of Unisolable Piping
• Thermal Stratification in Piping Systems
• Vibration Fatigue Testing of Socket Welds, Phase II
  

For more information, contact SI.