Authors & Affiliations

Gay E.V.
Russian Federation State Scientific Centre — A.I. Leipunsky Institute of Physics and Power Engineering, Obninsk, Russia


The problem of ambiguity of covariance matrix of experimental data evaluation uncertainties is considered. The relationships between covariance matrix of experimental data and covariance matrix of least square method evaluation, true for any physical or mathematical model (regression function), used in data reprocessing, are obtained. It is proved that uncertainties of some weighted averages of evaluated values are independent from these models. The proximity of the uncertainties of the integral parameters of nuclear reactors to these invariant uncertainties has been assessed and then considered as argument for the proposed solution of the ambiguity problem: any correctly derived covariance matrix of the evaluated data uncertainties can be used to estimate integral parameters uncertainty.

covariance matrix of the errors, least square method, regression function, functional reactor

Article Text (in Russian, PDF)


UDC 539.17

Problems of Atomic Science and Technology. Series: Nuclear Constants", issue 1-2, 2007