Fission Yields evaluation for the main actinides involved in the nuclear fuel cycle

The operation of nuclear reactors dedicated to the production of electricity by chain reactions requires the control of a large number of physical quantities throughout the life of the reactor. This control is essential for the operation and safety of the installation and is partly based on nuclear data. Indeed, given the power of today's computers and the continuous improvement of our calculation tools, nuclear data are now the main source of bias and uncertainty in neutron simulations. The optimization of current and future nuclear systems therefore requires a better knowledge of nuclear data and their uncertainty.Among these data of interest for nuclear cycle studies are the fission yields which represent the production rate of the different nuclei produced by fission reaction. These data are particularly important in residual power and end-of-cycle inventory studies. Current estimates of fission yields, although compatible with current applications, remain incomplete. Moreover, the development of new generations of reactors and the potential new regulatory requirements require better data and better controlled uncertainties, in particular through the determination of variance-covariance matrices which allow to calculate rigorously the uncertainties associated with these neutron quantities.The present work is devoted to the development of a new method for estimating independent and cumulative fission yields, with their associated covariance matrices, for the ²³⁵U(nth,f) and ²³⁹Pu(nth,f) reactions. To achieve this, a careful analysis of the experimental data was conducted to produce a more accurate estimate of the uncertainties in the fission product yields. The compatibility of the experimental data was carefully examined and then corrected to provide complete estimates in mass and charge, consistent in uncertainties and correlations between the different fission observables. Following the statistical analysis of the experimental data, several evaluations were proposed depending on the method used to resolve the data compatibility issues. The first proposed solution, called 'conservative', consists in keeping all the data by applying a penalty to all the measurements. The second solution consists in discarding a part of the measurements in order to keep only a consistent set of data. A second important aspect of this study was the development of a new approach to model the nuclear charge distributions before and after prompt neutron emission, going beyond the parameterization proposed by A.C. Wahl (parameterization used in most of the evaluated files). This work provides a comprehensive evaluation of nuclear charge distributions of interest for both applications and fundamental physics studies of the fission process. For the first time, probability density distributions and variance-covariance matrices have been determined for the observables associated with fission fragments. A validation step was performed to evaluate the impact of these new evaluations on the integral parameters of a reactor, such as residual power, reactivity loss and number of delayed neutrons. This step shows the coherence of the new evaluations of the fission yields and allowed to improve the precision of the observables useful for the physics of the reactors. (author)