doi:10.3808/jei.201500315
Copyright © 2017 ISEIS. All rights reserved

A Fuzzy Gradient Chance-Constrained Evacuation Model for Managing Risks of Nuclear Power Plants under Multiple Uncertainties

Z. Li1, G. H. Huang1,2*, L. Guo1, Y. R. Fan1,2 and J. P. Chen1

  1. Institute for Energy, Environment and Sustainability Research, UR-NCEPU, University of Regina, Regina, Saskatchewan, Canada S4S 0A2
  2. Institute for Energy, Environment and Sustainability Research, UR-NCEPU, North China Electric Power University, Beijing 102206, China

*Corresponding author. Tel: Fax: Email: huang@iseis.org

Abstract


Emergency evacuation is one of the most important risk management measures for nuclear accidents. The evacuation planning system contains great complexities, which have posed many challenges for decision makers. In the study, an inexact fuzzy gradient chance-constrained evacuation model (IFGCCEM) is proposed to address various uncertainties. The IFGCCEM is developed by incorporating fuzzy gradient chance-constrained programming into an inexact optimization framework. It can reflect uncertainties expressed as discrete intervals, and it is also capable of balancing decision makers’ optimism and pessimism using a combination of possibility and necessity, named fuzzy gradient measure. The proposed model has been applied to the evacuation planning of the Qinshan Nuclear Power Site, which is the largest nuclear site in China. It can generate optimal evacuation routes to maximize the evacuated population within limited time under various accident scenarios. Meanwhile, evacuation schemes with varied levels of system-failure risk can be obtained through post-optimization analysis. The results can provide an insight into the complex relationships in evacuation planning systems, as well as valuable decision support for effective risk management in response to nuclear emergencies.

Keywords: Risk management; evacuation planning; fuzzy programming; optimization; uncertainty; Qinshan Nuclear Power Site


Supplementary Files:

Refbacks

  • There are currently no refbacks.