A New Adjustment Code Based on the Bayes’ Theory Combined with the Monte-Carlo Technique

Abstract

A new adjustment code using the Monte-Carlo technique, based on the Bayes’ theory, has been developed to estimate the non-negative neutron flux and spectra from multiple activation detectors.

The Monte-Carlo technique is used to generate the initial multi-group neutron spectrum as the random number series from a priori probability density function (p.d.f.). And final neutron spectra and its variance-covariance matrix are obtained through statistical estimations of these random spectrum considering each conditional probability.

An emphasis is placed on the ability to estimate a posteriori p.d.f. rigorously for every differential flux and integral quantities, which are thought to be a little different from the standard normal distribution.

Using a lognormal distribution with J-log type unfolding method[l], this new adjustment code has been satisfactorily applied to a typical test problem to which ordinary algorithms are sometimes to give unphysical negative answers due to its large a priori uncertainty.