Solution of a Nuclear Reactor Parameter Identification Problem
Abstract
A continuous identification of parameters is performed on a simulated fast breeder nuclear reactor system using hybrid computation and applying techniques of statistical regression analysis and exponentially-mapped-past functions. Output states which are not directly measurable are estimated by use of a Kalman filter. The method developed in this study is applied to a numerical example which demonstrates that unknown parameters can be identified within 3% of their actual value, with signal noise ratios as low as 10:1 in the measured states. The example also demonstrates that convergence occurs in a reasonably short time.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 1971
- Accession Number
- AD0738874
Entities
People
- Oscar E. Brain Canepa
Organizations
- Naval Postgraduate School