CONFLUENCE OF CALCULATIONAL AND EXPERIMENTAL INFORMATION FOR DETERMINATION OF POWER DISTRIBUTION AND BURNUP

A technique for the statistical confluence of any number of possibly c orrelated informational sources employed in reactor analysis can be us ed to improve the estimates of physical quantities given by the source s taken separately. The formulas of the presented technique being base d on multivariate Bayesian conditioning are general and can be employe d in different applications. Insight into the nature of the informatio nal source allows different types of data associated with the source t o be improved. Estimation of biases, variances and correlation coeffic ients for the systematic and statistical errors associated with the in formational sources is essential for reliable confluence, but pays off by providing optimal estimates. The technique of the calculational an d experimental information confluence is applied to the determination of the power distribution and burnup for the research reactor HOR of t he Delft University of Technology. The code system CONHOR carries out all the stages of the calculation for the HOR reactor, using an existi ng code for static core calculations and burnup calculations.