Fusion Integral Experiments and Analysis and the Determination of Design Safety Factors — II: Application to the Prediction Uncertainty of Tritium Production Rate from the U.S. DOE/JAERI Collaborative Program on Fusion Blanket Neutronics

Many fusion integral experiments were performed during the last decade within a well-established collaboration between the United States and Japan on fusion breeder neutronics. These experiments started in 1983 and aimed at verifying the prediction accuracy of key neutronics parameters based on the state-of-the-art neutron transport codes and basic nuclear databases. The tritium production rate (TPR) has the prime focus among other reactions. The experimental and calculational data sets of local TPR in each experiment were interpolated to give an estimate of the prediction uncertainty, u_i, and the standard deviation, δ_i of the line-integrated TPR, a quantity that is closely related to the total breeding ratio (TBR) in the test assembly. A novel methodology developed during the collaboration was applied to arrive at estimates to design safety factors that fusion blanket designers can use to ensure that the achievable TBR in a blanket does not fall below a minimum required value. Associated with each safety factor is a confidence level, designers may choose to have, that calculated TPR will not exceed the actual measured value. Higher confidence levels require larger safety factors. Tabular and graphical forms for these factors are given, as derived independently for TPR from Li-6 (T₆), Li-7 (T₇), and natural lithium (T_n). Furthermore, distinction was made between safety factors based on the technique applied, discrete ordinates methods, and Monte Carlo methods in the U.S. calculations, JAERI's calculations, and in both calculations considered simultaneously. The derived factors are applicable to TPR in Li₂O breeding material; nevertheless, the results can be used as initial guidance to assist in resolving the tritium self-sufficiency issue in other breeding media.