In the neutron activation calculation, the inherent uncertainty of the input nuclear data will cause a certain impact on the calculation results. The uncertainty of the calculation results plays an important role in the source term analysis and radiation shielding design of nuclear facilities.

This study aims to analyze the uncertainty of neutron activation calculation based on direct derivative method.

Firstly, the direct derivation method and Gear algorithm for uncertainty analysis were investigated, and the activation coefficient matrix and sensitivity coefficient matrix were constructed. Then, the Gear algorithm was employed to solve the activation equation and sensitivity equation simultaneously, and the sensitivity coefficient of nuclide inventory to nuclear data was obtained. The relative uncertainty of nuclide inventory was obtained by combining the relative uncertainty of nuclear data. Finally, This method was integrated into the neutron activation program ABURN, and typical examples was selected to test and verify its performance.

The calculation results of the nuclide inventory and its sensitivity coefficient and relative uncertainty by the ABURN program have little deviation from the analytical solution and the numerical solution of the European activation program FISPACT, most of the deviations do not exceed 0.2%, and the maximum deviation does not exceed 1%.

Verification results show that the method and procedure developed in this paper have the ability to analyze the sensitivity and uncertainty of nuclide inventory with high precision, and can provide tools and data support for the radiation protection of nuclear facilities and the source term analysis.