An Analytical Expression for the True-value Approach in Extreme-value Statistics Applied to Channel Powers

Abstract

A method for evaluating the true value of the maximum channel power, based on a set of measured channel powers, has been proposed in a Canadian Nuclear Society (CNS) conference paper by Paul Sermer et al. [1]. This method makes use of the Monte Carlo technique to calculate the extreme value distribution of channel powers and obtain the compliance probability of the true maximum. This paper describes an analytical expression for the same extreme value distribution, which leads to the same channel power compliance probabilities as are obtained with the Monte Carlo technique. Though the analytical expression must be evaluated numerically to obtain the probability, there is a potential for a significant gain in computing time when compared to the Monte Carlo technique, especially when the method is embedded in complex calculations such as the determination of trip probabilities in CANDU Regional Overpower Protection systems.