RAVEN: a GUI and an Artificial Intelligence Engine in a Dynamic PRA Framework
Increases in computational power and pressure for more accurate simulations and estimations of accident scenario consequences are driving the need for Dynamic Probabilistic Risk Assessment (PRA) [1] of very complex models. While more sophisticated algorithms and computational power address the back end of this challenge, the front end is still handled by engineers that need to extract meaningful information from the large amount of data and build these complex models. Compounding this problem is the difficulty in knowledge transfer and retention, and the increasing speed of software development. The above-described issues would have negatively impacted deployment of the new high fidelity plant simulator RELAP-7 (Reactor Excursion and Leak Analysis Program) at Idaho National Laboratory. Therefore, RAVEN that was initially focused to be the plant controller for RELAP-7 will help mitigate future RELAP-7 software engineering risks. In order to accomplish this task, Reactor Analysis and Virtual Control Environment (RAVEN) has been designed to provide an easy to use Graphical User Interface (GUI) for building plant models and to leverage artificial intelligence algorithms in order to reduce computational time, improve results, and help the user to identify the behavioral pattern of the Nuclear Power Plants (NPPs). In this paper we will present the GUI implementation and its current capability status. We will also introduce the support vector machine algorithms and show our evaluation of their potentiality in increasing the accuracy and reducing the computational costs of PRA analysis. In this evaluation we will refer to preliminary studies performed under the Risk Informed Safety Margins Characterization (RISMC) project of the Light Water Reactors Sustainability (LWRS) campaign [3]. RISMC simulation needs and algorithm testing are currently used as a guidance to prioritize RAVEN developments relevant to PRA.
- Research Organization:
- Idaho National Lab. (INL), Idaho Falls, ID (United States)
- Sponsoring Organization:
- DOE - NE
- DOE Contract Number:
- DE-AC07-05ID14517
- OSTI ID:
- 1097158
- Report Number(s):
- INL/CON-13-28360
- Resource Relation:
- Conference: ANS,Atlanta, GA,06/16/2013,06/20/2013
- Country of Publication:
- United States
- Language:
- English
Similar Records
RAVEN Theory Manual
Methodology for the Incorporation of Passive Component Aging Modeling into the RAVEN/ RELAP-7 Environment