Research on the source-detector variance reduction method based on the AIS adjoint Monte Carlo method

•A coupling variance reduction method based on the AIS adjoint Monte Carlo method is proposed and implemented. Avoid the coupling problem of determinism and Monte Carlo method in the traditional CADIS method.•A simple example is used to verify the correctness of the coupling variance reduction method of AIS with Monte Carlo.•A commercial reactor shielding example is calculated and compared with the traditional variance reduction method. The maximum statistical error is less than 5%, and the relative deviation from the measured value is within 20%. The coupled variance reduction method based on adjoint calculation has a good variance reduction effect when solving the source-detector problem. This paper incorporates the Automatic Importance Sampling (AIS) method into the adjoint Monte Carlo method and then proposes a coupled variance reduction method based on the AIS adjoint Monte Carlo method. The result shows that the total flux deviation between the AIS adjoint Monte Carlo method and the original adjoint method is 0.13%, indicating good agreement and verifying the accuracy of the AIS adjoint method. In this paper, the AIS adjoint Monte Carlo method is applied to the calculation of a commercial reactor shielding example, and four sets of source bias and weight window parameters of the AIS adjoint calculation are used as calculation examples. The maximum statistical error is less than 5.00%, and the relative deviation from the measured value is within 20.00%. Generally, the calculation efficiency of AIS-CADIS is slightly better than that of CADIS.