Speaker
Description
During the last decades, the development of new powerful computers and high performance analytical tools, along with the reduction of the approximations due to new methods implemented in the algorithms for the solution of the transport equation, pushed nuclear cross-sections data as the main source of uncertainty in neutronic calculations. This points out the importance in quantifying nuclear uncertainties and highlights the need for additional efforts to reduce them.
To accomplish this task, more advanced experimental facilities should be designed and built, more accurate measurement devices developed, and nuclear models necessarily refined, so requiring considerable time and investments both from an economic and human perspective.
However, the heritage supplied by tens of fast spectrum systems which were built and operated in the past, provides a considerable database which can be leveraged in order to avoid performing new experiments through the nuclear data adjustment technique.
ENEA, through the participation to the FALCON Consortium, is pursuing all activities required to support the construction of ALFRED – the European demonstrator of the LFR technology – in Romania, notably to what concerns the core design, of which ENEA is responsible.
In this context, in order to refine target accuracies for complying with the ambitious safety goals and/or setting new optimized safety margins for the ALFRED reactor, an extended version of the old, proprietary AMARA code, which was developed for adjustment purposes, has been developed.
This new version, named AMARA+, uses covariances opportunely generated from the most updated nuclear data evaluations, and extends the capabilities of the old code by flexibly allowing any user-defined selection of integral experiments conducted on systems/facilities representative of the system of interest. The approach chosen for the adjustment procedure in AMARA+ is still based on the maximization of the so-called likelihood function, which correlates the original nuclear data to the selected experimental integral parameters, and the optimization problem is solved via Lagrange multipliers technique. Furthermore, a chi squared-type test has been added for better substantiating the statistical soundness of the procedure.
In the present case, the reference nuclear data are taken from the ENDF/B-VIII.0 library, while a number of integral experiments, taken from the OECD/NEA’s IRPhE database, were used for performing a preliminary adjustment for the ALFRED reactor.
A comprehensive description of AMARA+ is presented in this work, stressing the aspects related to the algorithm chosen, together with the preliminary results obtained using the restricted set of integral experiments chosen.
| Speaker's title | Mr |
|---|---|
| Speaker's email address | donato.castelluccio@enea.it |
| Country/Int. organization | Italy |
| Affiliation/Organization | ENEA |
