04/01843 On-line neuro-expert monitoring system for Borssele nuclear power plant: Nabeshima, K. et al. Progress in Nuclear Energy, 2003, 43, (1–4), 397–404
- Book ID
- 104279635
- Publisher
- Elsevier Science
- Year
- 2004
- Weight
- 175 KB
- Volume
- 45
- Category
- Article
- ISSN
- 0140-6701
No coin nor oath required. For personal study only.
✦ Synopsis
crossover, mutation and random assessment of population for multicycle loading pattern (LP) optimization. A detailed description of the chromosomes in the genetic algorithm coded is presented. Artificial Neural Networks (ANNs) have been constructed and trained to accelerate the GA-based search during the optimization process. The whole package, called GAOPT, is linked to the reactor analysis code PANTHER, which performs fresh fuel loading, burn-up and power shaping calculations for each reactor cycle by imposing station-specific safety and operational constraints. GAOPT has been verified by performing a number of tests, which are applied to the Hinkley Point B and Hartlepool reactors. The test results giving loading pattern (LP) scenarios obtained from single and multi-cycle optimization calculations applied to realistic reactor states of the Hartlepool and Hinkley Point B reactors are discussed. The results have shown that the GA/ ANN algorithms developed can help the fuel engineer to optimize loading patterns in an efficient and more profitable way than currently available for multi-cycle refuelling of AGRs. Research leading to parallel GAs applied to LP optimization are outlined, which can be adapted to present day LWR fuel management problems.
04/01839 Goal-oriented flexible sensing for higher diagnostic performance of nuclear plant instrumentation Takahashi, M. et al. Progress in Nuclear Energy, 2003, 43, (1-4), 105-111. A framework of goal-oriented sensing has been proposed with the emphasis on the integration of an inference mechanism with an active sensing module, which is equipped with a driving mechanism and a set of sensors. The validity of the dynamic failure identification based on the proposed framework of goal-oriented sensing has been examined under realistic experimental conditions using a small-scale test loop. Though the loops are small in scale, the validity of introducing mobile sensing mechanism has been successfully shown through the experiments emulating realistic failure conditions.