Emergency Evacuation Management for Nuclear Power Plant Accidents Under Multiple Uncertainties

Abstract

Nuclear power accidents are one of the most dangerous disasters posing a lethal

threat to human health and have detrimental effects lasting for decades. Therefore,

emergency evacuation is important to minimize injuries and prevent lethal

consequences resulting from a nuclear power accident. An evacuation management

system in response to a nuclear power plant accident, involves a number of processes

with a variety of socio-economic and environmental implications. These processes

may be influenced by a number of factors, such as availability of traffic vehicles

employed, number of evacuees, destinations of evacuees, limitations of evacuation

times, budgets for evacuation, shelter and hospital capacities, and environmental

regulations of the areas involved. Extensive uncertainties exist in the evacuation

process and the associated factors. Hence, innovative inexact optimization approaches

may be proposed to account for various uncertainties existing in an evacuation

planning system.

In this dissertation, a series of inexact optimization approaches will be proposed

to reflect compound uncertainties existing in emergency evacuation planning systems.

In detail, (i) an interval-based evacuation management (IBEM) model will be

developed in response to a nuclear-power plant accident; (ii) an inexact

gradient-based fuzzy chance constrained programming (IGFCCP) method will be

proposed to balance a decision maker’s optimistic and pessimistic preferences and

applied to the planning of the evacuation scheme for the Qinshan Nuclear Power Site

(QNPS); (iii) an inexact fuzzy stochastic chance constrained programming (IFSCCP)

approach will be developed to address various uncertainties and optimize the planning

of the evacuation scheme for the QNPS; (iv) an inexact mixed-integer credibility-constrained de Novo programming (IMICDNP) method will be developed

to analyze the trade-offs between conflicting objectives under uncertainty.

The major accomplishments of this research are summarized as follows:

(i) A set of inexact optimization approaches are proposed to deal with

uncertainties in emergency evacuation planning systems. The uncertainties, expressed

in various forms such as interval, fuzzy, fuzz random variables, can be well reflected

through the developed IBEM, IGFCCP, IFSCCP, IMCDP methods. Optimal decision

alternatives for emergency evacuation planning are obtained through the proposed

approached under various uncertain conditions.

(ii) The applicability of the proposed approaches are demonstrated through a

real-case study of emergency evacuation management at the Qinshan Nulcear Power

Site (QNPS), which is the first nuclear power plant in China. Under consideration of

various system complexities and uncertainties, potential.