Nuclear systems, composed of reactors with varied fuel and cycle facilities (enrichment plant, fabrication plant, reprocessing plant, etc.), are complex and in constant evolution. Since 1985, the CEA has been developing the simulation software COSI to study different trajectories of nuclear fleet evolution and provide technical elements to decision makers. The principle of COSI, including the typical composition of the data set, is exposed. To evaluate as accurately as possible the isotopic compositions of materials, several physical models are implemented in COSI. The main ones, the evolution calculation and the equivalence models, are described in detail. An exercise of validation of COSI carried out on the French pressurized water reactor (PWR) historical nuclear fleet until 2010 is also presented, as well as a methodology for propagation of input uncertainties on COSI results.

To illustrate the possibilities of COSI, the results of different scenarios studied in the framework of the French Act for Waste Management are discussed. The objective of these scenarios is to evaluate the feasibility of sodium-cooled fast reactor (SFR) deployment to renew the French PWR fleet on different timescales and to analyze the costs and the benefits of different options of minor actinide (MA) partitioning and transmutation. The impacts of SFR deployment on cycle facilities such as the fabrication plant, the spent fuel storage, and the reprocessing plant are minimized. The SFR deployment appears to be feasible with regard to fissile material availability, with an adaptation of fuel cooling time before reprocessing or of SFR breeding gain.

Minor actinide transmutation in homogeneous mode MA diluted in core) and transmutation in heterogeneous mode (in MA-bearing blankets) are compared not only according to their impacts on cycle facilities and on ultimate waste but also according to the reduction of their inventories. The increases in fresh fuel thermal power and spent fuel decay heat due to the addition of MAs in fuels are quantified. The cases of transmutation of all MAs (americium, neptunium, and curium) and of americium only are distinguished. Alternative scenarios are explored to overcome the challenges associated with each option: reduction of the maximal MA content in fresh SFR fuel in the case of homogeneous transmutation and reduction of interim MA storage in the case of heterogeneous transmutation.