Loading...

Development of an Effective Method to Support Severe Accident Management in Bushehr Nuclear Power Plant

Saghafi, Mahdi | 2016

550 Viewed
  1. Type of Document: Ph.D. Dissertation
  2. Language: Farsi
  3. Document No: 48940 (46)
  4. University: Sharif University of Technology
  5. Department: Energy Engineering
  6. Advisor(s): Ghofrani, Mohammad Bagher
  7. Abstract:
  8. Following Three Mile Island (TMI) accident in 1979, first severe accident (SA) in Nuclear Power Plants (NPPs), Accident Management Support Tools (AMSTs) were developed and installed in a number of NPPs. Lessons learned from Fukushima accident highlighted importance of Accident Management (AM) in mitigation severe radiological consequences after a SA and suggested reconsiderations of AM program which in turn created the need for AMSTs adaption and modernization. An efficient AMSTs should have the following principal capabilities: (1) Identification of accidents and diagnosis of the plant damage state (PDS), (2) Prediction of accident progress path and (3) Source term analysis and prediction of radioactive material release. AMSTs can provide vital information about the plant states, e.g. timing of critical events, severe AM entry time and quantitative estimation of important parameters. Such information cannot be provided by typical severe accident management guidelines due to their dependency to the plant initial and boundary conditions. Early prediction of the plant parameters can provide wider time window for selection of Candidate High Level Actions (CHLAs) while enabling NPP operators better understanding of the accident progress path. In this thesis, a comprehensive literature review is conducted in the light of the lessons learned from Fukushima accident, in order to summarize and categorize available modern methods for AMST design, including comparative assessment of the advantages and limitations of each method. Consequently, an innovative design for an efficient AMST for Bushehr Nuclear Power Plant (BNPP) is proposed. AMST structure includes the following parts: (1) Tracker, (2) Predictor and (3) Decision Support. Identification of accident initiator and the plant state are the tasks defined for ‘Tracker’. Modular neuro Fuzzy networks, a kind of soft computing techniques, is employed for accident identification. First, accidents are categorized by five modular Artificial Neural Networks (ANN) into five accident categories. Then, five modular fuzzy systems are used to determine intiating event of the accidents in each category consisting location of failures and breaks. In case of postulated breaks, NARX ANN is used for break size estimation based on pressure variation in primary system of Bushehr NPP. NARX is able to directly deal with time-depedent signals and to be applied in real-time applications. The Predictor is designed for prediction of accident progress path. MELCOR code is used as a computational engine in Predictor, to provide with most probable phenomena, events timing and source term evaluation. A nodalization is developed and qualified using experimental data of PSB-VVER integral test facility. Original and improved FFTBM are applied to determine the accuracy of the developed nodalization and MELCOR code. Finally, model of BNPP is developed based on the qualified nodalization of PSB-VVER. MELCOR results are in good agreement with the final safety assessment reports (FSAR) and beyond desing basis accidents (BDBA) control manual of Bushehr NPP.The last part of AMST, Decision support, is conceived to support decision making in selection of AM measures and their implementation times. It has interactive interface with the plant operators for effectiveness assessment of CHLAs. A rule-based Expert System with a mixed chaining reasoning algorithm is employed in the structure of ‘Decision support’ to compare positive and negative effects of the selected CHLAs. This thesis provides conceptual design of an AMST for BNPP which can be applied to other PWR-type NPPs. AMSTs reduce ordinary and mental work load, and increase situation awareness of operators. Using AMSTs, operators are able to timely assess the effectiveness of CHLAs
  9. Keywords:
  10. Decision Making Support System ; Expert System ; Crash Severity ; Accident Prediction ; Accidents Management Support Tool ; Accidents Identification ; Accident Development

 Digital Object List

 Bookmark

No TOC