Reactor, criticality and transport simulations are widely used in the nuclear community to e.g. determine safety parameters, evaluate transients, and calculate the fuel inventory. These simulations use nuclear data (ND) as one of their most important inputs. ND is obtained by performing experiments and using theoretical nuclear models. Both experiments and theory are associated with uncertainties and consequently all ND are associated with uncertainties. There are also strong correlations in the uncertainties between different energies, reaction channels and isotopes, in both experiments and modelling. Many ND libraries (e.g. JEFF-3.2 and ENDF/B-VII.1) contain information on the ND uncertainties and their correlations in covariance files. However, this information relies on assumptions of normal distributions and is not complete. Furthermore, many reactor codes do not use ND uncertainties as input, and when they do, they rely on different assumptions, which tend to underestimate the propagated uncertainty. In order to address these issues the Total Monte Carlo (TMC) methodology has been developed . The basic principles of the TMC method are illustrated in Figure 1.
Figure 1. The TMC uncertainty propagation and TENDL production. In the TMC uncertainty propagation the final result is the spread in a macroscopic parameter. This spread is the systematic uncertainty in the calculation due to ND in the investigated parameter. (CS = cross section, FY = fission yield)
In the TMC method a large set of random files are derived by sampling nuclear model parameters in the nuclear model codes TALYS. The random files are subsequently compared to experimental values . Consequently, each random file is a complete nuclear data library containing one possible representation of the nuclear data given the uncertainties from theory and experiments.
By running a reactor simulation multiple times, each time with a new random file as input, distributions of the different nuclear engineering parameters (e.g. keff, temperature coefficient, inventory, fuel temperature) are derived. These distributions are interpreted as the uncertainties in the engineering parameters due to ND. The TMC method can also be used to produce nuclear data for all open reaction channels including covariances; the “TALYS Evaluated Nuclear Data Library” (TENDL) is an example . In order to select the best file to be included in the TENDL library, the random nuclear data files produced are compared against differential experimental data taken from the EXFOR database and a large set of integral measurements.
The TMC method can be used for any neutronic-system and in this contribution we present results from both thermal and fast neutron systems. The importance to include nuclear data uncertainty from angular distribution and thermal scattering are highlighted.
 A.J.Koning, and D.Rochman, Modern Nuclear Data Evaluation with the TALYS Code System. Nuclear Data Sheets 113 12 2841-2934 (2012)
 P. Helgesson, H.Sjöstrand, A.J.Koning, D.Rochman, E.Alhassan, S.Pomp
Nuclear Data Sheets, Volume 123, Pages 214-219, 2015.
 "TENDL-2014: TALYS-based evaluated nuclear data library", A.J. Koning, D. Rochman, S. van der Marck, J. Kopecky, J. Ch. Sublet, S. Pomp, H. Sjostrand, R. Forrest, E. Bauge, H. Henriksson, O. Cabellos, S. Goriely J. Leppanen, H. Leeb, A. Plompen and R. Mills, www.talys.eu/tendl-2014.html
uncertainty propagation, TMC, TENDL, nuclear data.
The 17th meeting on Reactor Physics in the Nordic Countries Chalmers University of Technology, Gothenburg, Sweden, May 11-12, 2015