Seminar: Employing advanced methodologies for Uncertainty Quantification and Probabilistic Risk Assessment in an integrated fashion: RAVEN framework

Dr. Andrea Alfonsi, Idaho National Laboratory

All dates for this event occur in the past.

E141 Scott Lab
E141 Scott Lab
201 W. 19th Ave.
Columbus, OH 43210
United States

The development of high-fidelity codes, for thermal-hydraulic systems and integrated multi-physics, has undergone a significant acceleration in the last few years. Multi-physics codes involve simulations that treat multiple physical models or multiple simultaneous physical phenomena, in an integrated solving environment. Multi-physics application typically resolve in solving coupled systems of partial differential equations, generally characterized by different geometry and time scales.

The new multi-physics codes are characterized by remarkable improvements  in the approximation of physics (high approximation order and reduced use of  empirical correlations). This higher fidelity is generally accompanied by a  larger computational effort (calculation time increased). This peculiarity is an obstacle in the application of traditional computational techniques of  quantification of uncertainty and risk associated with the operation of  particular industrial plant (e.g., nuclear power plants).  In this context, innovative methods for the assessment of the risk and the impact of the uncertainties in a probabilistic environment  (Probabilistic Risk Assessment) are necessary.  Indeed, the current methodologies based on static fault and event tree approaches, result to be limited overall when the timing of events is crucial for an accurate risk evaluation associated with accident scenarios. In addition, these methodologies present several modeling issues when the status of the  system (e.g. Nuclear Power Plant) cannot be modeled in terms of Top Event success/failures only. For example, when complex scenarios need to be analyzed, which are generally affected by operator actions. In order to tackle these new challenges, a new tool has been developed at the Idaho National Laboratory: Risk Analysis Virtual ENviroment (RAVEN). RAVEN is a generic software framework designed to perform parametric and stochastic analysis based on the response of complex system codes. It has been developed with the mindset of incorporating classical and advanced methodologies in the same software platform. Indeed, it is capable of investigating the system response as well as the input space using classical sampling schemes, but its strength is focused toward system feature discovery, such as limit surfaces, separating regions of the input space leading to system failure, using dynamic supervised learning techniques. This seminar will provide an overview of the capabilities of the RAVEN framework providing examples on its application on real challenging problems.

About the Speaker

Dr. Andrea Alfonsi is a Senior Nuclear Engineer at the Idaho National Laboratory in the Nuclear Science & Technology Directorate, Nuclear Engineering Methods Development Department. Dr. Alfonsi serves as technical leader of the Risk Analysis Virtual ENviroment (RAVEN) software. RAVEN is a generic software framework designed to perform parametric and stochastic analysis based on the response of complex system codes. Dr. Alfonsi is also the main developer of the Parallel and Highly Innovative Simulation for the INL Code System (PHISICS) toolkit. PHISICS is intended to provide a modern analysis tool for reactor physics investigation. It is designed with the mindset to maximize accuracy for a given availability of computational resources and to give state of the art tools to the nuclear engineer. Dr. Alfonsi is involved is several other projects among which the Nuclear-Renewable Hybrid Energy Systems (N-R HES) program under the DOE-NE Crosscutting Technologies Program. N-R HES seeks to coordinate the use of multiple clean energy generation sources to meet both thermal and electrical energy needs.

Hosted by Professor Carol Smidts