Requisite accuracy for hot spot factors in fast reactors

In the thermal design of a fast reactor, it should be most effective to reduce hot spot factors to the lowest possible level compatible with safety considerations, in order to minimize the design margin for the temperature prevailing in the core. Hot spot factors account for probabilistic and statistic deviations from nominal value of fuel element temperatures, due to uncertainties in the data adopted for estimating various factors including the physical properties. Such temperature deviations necessitate the provision of correspondingly large design margins for temperatures in order to keep within permissible limits the probability of exceeding the allowable temperatures. Evaluation of the desired accuracy for hot spot factors is performed by a method of optimization, which permits determination of the degree of accuracy that should minimize the design margins, to give realistic results with consideration given not only to sensitivity coefficients but also to the present-day uncertainty levels in the data adopted in the calculations. A concept of ''degree of difficulty'' is introduced for the purpose of determining the hot spot factors to be given higher priority for reduction. Application of this method to the core of a prototype fast reactor leads to the conclusion that the hot spot factors to be given the highest priority are those relevant to the power distribution, the flow distribution, the fuel enrichment, the fuel-cladding gap conductance and the fuel thermal conductivity. (auth.)