Abstract
The aim of this study was to pinpoint uncertainties in solubility calculations in aqueous solutions with the use of sensitivity and uncertainty (s/u) analysesSolubility calculations are often used as a basis for other calculations in performance assessments of repositories for spent nuclear fuel. It is thus important to know how reliable the predicted solubilities are. A mean is to use a verified and validated code for the calculations, and to be able to express the importance of the uncertainty in the input data. Thus it may be more correct to give the solubility as an interval with a related.
frequency function.
This paper is focused on solubility calculations with a well-defined database. Uncertainty and sensitivity analyses are made with respect to uncertainties in the stability and solubility constants, pH, pe and temperature.
The solubility calculations are performed using the PHREEQE program and the statistical evaluation is made using the SENVAR package. The s/u analysis is based on a modified Monte Carlo sampling procedure followed by a stepwise regression in order to determine the relative importance of the parameters concerned.
The empirical frequency function may be displayed using this program together with some statistics, e.g. mean, variance, skewness and confidence intervals for both the mean and the solubility population. The last is valid for an arbitrary distribution, while a log-normal assumption is made for the mean.