Fusion Integral Experiments and Analysis and the Determination of Design Safety Factors — I: Methodology

The role of the neutronics experimentation and analysis infusion neutronics research and development programs is discussed. A new methodology was developed to arrive at estimates to design safety factors based on the experimental and analytical results from design-oriented integral experiments. In this methodology, and for a particular nuclear response, R, a normalized density function (NDF) is constructed from the prediction uncertainties, u_i's, and their associated standard deviations, ±δ_i's, as found in the various integral experiments where that response, R, is measured. Important statistical parameters are derived from the NDF, such as the global mean prediction uncertainty, ū, and the possible spread, ±δ_u, around it. The method of deriving safety factors from many possible NDFs based on various calculational and measuring methods (among other variants) is also described. Associated with each safety factor is a confidence level, designers may choose to have, that the calculated response, R, will not exceed (or will not fall below) the actual measured value. An illustrative example is given on how to construct the NDFs. The methodology is applied in two areas, namely the line-integrated tritium production rate and bulk shielding integral experiments. Conditions under which these factors could be derived and the validity of the method are discussed. The described methodology could be applied to the integral experiments proposed for the International Thermonuclear Experimental Reactor (ITER) neutronics research and development, particularly, in deriving the required safety factors for ITER shielding.