Comparison of cost-reduced computational methods for nuclide transmutation problems

This technical material compares the performance of the Mini-Max Polynomial Approximation method (MMPA) and the Okamura Explicit Method (OEM) for nuclide transmutation calculations suitable for nuclear fuel cycle simulations which require low computational cost. The choice between these two methods strongly depends on the target accuracy; the performance of OEM is satisfactory and there is little incentive to replace OEM by the low-order MMPA under the criteria that the maximum error of nuclide number density prediction is less than 10%. The increase in the computational cost of the low-order MMPA from the 10% target accuracy to 1% is insignificant, and thus the low-order MMPA is promising if higher accuracy is required. Additional comparisons are made between the MMPA coefficient sets obtained through the original procedure to minimize the global maximum error and the alternative procedure to minimize the local error in the polynomial approximation. The alternative MMPA coefficients provide more accurate results than the original ones, and the divergence phenomenon appears only in the results of the original MMPA coefficients.