[ZMS16] A New Approach to Estimate Uncertainty in Waste Characterization
Conférence Internationale avec comité de lecture :
Advances in Nuclear Nonproliferation Technology and Policy Conference ,
September 2016,
pp.pp,
Santa Fe,
USA,
Mots clés: radionuclear waste, bootstrap, linear models
Résumé:
The characterization of radioactive waste is a complex
task, especially when historical waste is involved. At the
European Organization for Nuclear Research (CERN),
gamma-ray spectrometry is used to estimate the specific activity of Easyto-
Measure (ETM) radionuclides [1]. Difficult-to-Measure
(DTM) radionuclides, which are beta and low-energy X-ray emitters,
are either measured by radiochemical techniques or evaluated
by calculations and Monte Carlo simulations .
The specific activities are then compared to the acceptance
limits of the national agencies for waste management. Waste
producers must ensure that these limits are respected and must
estimate the distribution of the quantities of interest together
with their uncertainties.
We selected weighted linear models for studying the relationship
between ETM and measured DTM radionuclides.
The use of the so-called bootstrap is described when calculating
average specific activities of DTM radionuclides and
their distributions. This last technique is also useful when
a limited number of samples is available or their collection
is made following a non-probabilistic model. Bootstrap can
furthermore be used for bias estimation.
We conclude by presenting a scheme to estimate the total
uncertainty of the waste characterization process.
Commentaires:
September 25-30, 2016