Abstract
After the tragedy of the World Trade Center of New York City at September 11th 2000, the possibility of the attack of nuclear power plant by terrorists attracts public attention. The security of nuclear power plant becomes an ever more important issue than before. The risk of operating a nuclear power plant associated with the terrorist attack becomes a very interesting research topic. Theoretically, the risk can be quantified as the summation of the risk of each individual area of the plant. The risk each individual area can be viewed as the product of five factors. These factors are the frequency of terrorist attack, the probability that the terrorist can break into vital area of the plant, the probability of a specific area within the vital area becomes the target of the attack, the probability that terrorist can reach the area successfully, the conditional code damage probability (CCDP) of the specific area once the terrorists reach the area. In the present study, a mathematical model is developed to quantify the probability of a specific area within the vital area of the plant. The specific area on certain floor within certain building in the area is defined according to the fire zones used in the fire risk analysis of the probabilistic safety assessment (PSA) of nuclear power plant. It is assumed that the terrorists’ acts in the plant are purely random, i.e. their behavior can be simulated using Monte Carlo method with assumed probability distribution functions. It is also assumed that the invaders will change their direction of movement after travel certain distance. The distribution of the distance follows the exponential distribution. The same distribution that governs the distance traveled between two collisions of a particle in a medium. Each specific area has its own mean free path. The new direction is determined randomly based on a uniform distribution of 360 degrees. When there are doors within the travel distance, the invaders will make a decision whether they will leave the specific area and enter another area of the floor. The number of “stops” of invaders made in the specific area is recorded. The simulation starts when invaders enter the floor via certain entrance and stops when the invaders leave the floor. Three thousands simulations are performed for each floor area. The total number of “stops” within each specific area divides by the summation of the number stops of all the area in one floor is the probability of that specific area becomes the target of the terrorist attack. The Monte Carlo simulations are performed separately for every important floor of almost all the buildings within the vital area. The probability of invaders leave the floor through a particular entrance or exit can also be determined. Through the entrance and/or exit, the invaders can move to another floor. Therefore, the above mentioned probability is also the probability that invaders will enter a particular floor through the specified entrance and/or exit. Another set of Monte Carlo simulation based on these probabilities is performed to determine the probability that a particular floor and building will become the target of the attack. The surrogate plant used in the present study is Kuoshen Nuclear Power Station of Taiwan Power Company. The station employs a General Electric designed BWR VI (Boiling Water Reactor) reactor with Mark III containment. In the simulation, the mean free path of each specific area is correlated with the number of components and weight of cable in the area. The product of probability of a specific area becomes the target and the CCDP of a specific area represents the risk of the area. In the present study, the CCDP in the fire analyses of PSA is used. The model has identified the specific areas within the vital area of the plant that have higher risk and also the areas with higher probability that terrorist will appear. These areas are not necessary identical. The latter areas are also the areas that the security force can arrest the invaders. The results demonstrate that the risk of terrorist attack is dominated by the CCDP of the specific area. The results of the present study can used to enhance the security of the plant.