Design optimization methodology of reactor system for severe accident mitigation using the artificial neural neworks = 신경회로망을 이용한 중대사고완화를 위한 원자로계통설계 최적화 방법론에 관한 연구

Cited 0 time in webofscience Cited 0 time in scopus
  • Hit : 459
  • Download : 0
Recently, it is strongly recommended that severe accidents are to be considered in design stage of Nuclear Power Plants (NPP). However, there are no previous works on reactor system design to mitigate severe accidents in design stage. In this study, an op-timization methodology of design parameters, which consists of investigation of de-sign alternatives, analysis of plant characteristics, and suggestion of optimization direc-tion, has been developed to mitigate severe accidents in Nuclear Power Plant (NPP). Backpropagation neural network (BPN), one of Artificial Neural Networks (ANNs), was used to analyze plant characteristics under severe accident conditions. To investi-gate the effects of system design parameters on severe accident progression, an estima-tion model for system design parameters has been proposed using BPN. The strategy for design optimization was to increase safety design margins so that the NPP system may have a longer response time and be less sensitive to severe acci-dents investigated. Core uncovery time, vessel failure time, containment failure time, and radioactivity released to environment were selected as safety margin indicators, since they are important issues in severe accident management. The suggested methodology was applied to the system design optimization to pre-vent and mitigate core uncovery, vessel failure, containment failure and release of ra-dioactive materials. This work was limited to the thermal-hydraulic design of reactor coolant system (RCS) and engineered safety feature systems. Young Gwang nuclear power plant (YGN) 3&4 and Ulchin 3&4 Units in Korea were selected as reference plants to apply the methodology. Nine design parameters, which are important in ther-mal hydraulic design, were chosen as optimized parameters. Design alternatives, core uncovery time, vessel failure time, containment failure time, and release amount of ra-dioactive materials were generated using Latin Hypercube Sampling (LHS) and Modular Ac...
Advisors
Chang, Soon-Heungresearcher장순흥researcher
Description
한국과학기술원 : 원자력공학과,
Publisher
한국과학기술원
Issue Date
2001
Identifier
165875/325007 / 000955115
Language
eng
Description

학위논문(박사) - 한국과학기술원 : 원자력공학과, 2001.2, [ xi, 104 p. ]

Keywords

Accident Mitigation; Nuclear Power Plant; Severe Accident; Neural Network; Design Optimization; 설계최적화; 사고완화; 원자력발전소; 중대사고; 신경회로망

URI
http://hdl.handle.net/10203/48921
Link
http://library.kaist.ac.kr/search/detail/view.do?bibCtrlNo=165875&flag=dissertation
Appears in Collection
NE-Theses_Ph.D.(박사논문)
Files in This Item
There are no files associated with this item.

qr_code

  • mendeley

    citeulike


rss_1.0 rss_2.0 atom_1.0