Optimisation of internal contamination monitoring programme by integration of uncertainties.

Potential internal contamination of workers is monitored by periodic bioassay measurements interpreted in terms of intake and committed effective dose by the use of biokinetic and dosimetric models. After a prospective evaluation of exposure at a workplace, a suitable monitoring programme can be defined by choosing adequate measurement techniques and frequency. In this study, the sensitivity of a programme is evaluated by the minimum intake and dose, which may be detected with a given level of confidence by taking into account uncertainties on exposure conditions and measurements. This is made for programme optimisation, which is performed by comparing the sensitivities of different alternative programmes. These methods were applied at the AREVA NC reprocessing plant and support the current monitoring programme as the best compromise between the cost of the measurements and the sensitivity of the programme.

Integration of uncertainties into internal contamination monitoring.

Potential internal contaminations of workers are monitored by periodic bioassays interpreted in terms of intake and committed effective dose through biokinetic and dosimetric models. After a prospective evaluation of exposure at a workplace, a suitable monitoring program can be defined by the choice of measurement techniques and frequency of measurements. However, the actual conditions of exposure are usually not well defined and the measurements are subject to errors. In this study we took into consideration the uncertainties associated with a routine monitoring program in order to evaluate the minimum intake and dose detectable for a given level of confidence. Major sources of uncertainty are the contamination time, the size distribution and absorption into blood of the incorporated particles, and the measurement errors. Different assumptions may be applied to model uncertain knowledge, which lead to different statistical approaches. The available information is modeled here by classical or Bayesian probability distributions. These techniques are implemented in the OPSCI software under development. This methodology was applied to the monitoring program of workers in charge of plutonium purification at the AREVA NC reprocessing facility (La Hague, France). A sensitivity analysis was carried out to determine the important parameters for the minimum detectable dose. The methods presented here may be used for assessment of any other routine monitoring program through the comparison of the minimum detectable dose for a given confidence level with dose constraints.

A simple algorithm for solving the inverse problem of interpretation of uncertain individual measurements in internal dosimetry.

The individual monitoring of internal exposure of workers comprises two steps: measurement and measurement interpretation. The latter consists in reconstructing the intake of a radionuclide from the activity measurement and calculating the dose using a biokinetic model of the radionuclide behavior in the human body. Mathematically, reconstructing the intake is solving an inverse problem described by a measurement-model equation. The aim of this paper is to propose a solution to this inverse problem when the measurement-model parameters are considered as uncertain. For that, an analysis of the uncertainty on the intake calculation is performed taking into account the dispersion of the measured quantity and the uncertainties of the measurement-model parameters. It is shown that both frequentist and Bayesian approaches can be used to solve the problem according to the measurement-model formulation. A common calculation algorithm is proposed to support both approaches and applied to the examples of tritiated water intake and plutonium inhalation by a worker.

Modeling the imprecision in prospective dosimetry of internal exposure to uranium.

The dosimetry of internal exposure to radionuclides is performed on the basis of biokinetic and dosimetric models. For prospective purpose, the organ or effective dose resulting from potential conditions of exposure can be calculated by applying these models with dedicated software. However, it is acknowledged that a significant uncertainty is associated with such calculation due to the variability of individual cases and to the possible lack of knowledge about some factors influencing the dosimetry. This uncertainty has been studied in a range of situations by modeling the uncertainty on the model parameters by probability distributions and propagating this uncertainty onto the dose result by Monte Carlo calculation. However, while probability distributions are well adapted to model the known variability of a parameter, they may lead to an unrealistically low estimate of the uncertainty due to a lack of knowledge about some input parameters. Here we present a mathematical method, based on the Dempster-Shafer theory, to deal with such imprecise knowledge. We apply this method to the prospective dosimetry of inhaled uranium dust in the nuclear fuel cycle when its physico-chemical properties are not precisely known. The results show an increased estimation of the range of uncertainty as compared to the application of a probabilistic method. This Dempster-Shafer method may valuably be applied in future prospective dosimetry of internal exposure in order to more realistically estimate the uncertainty resulting from an imprecise knowledge of the parameters of the dose calculation.