Through experience feedback of the Chernobyl and Fukushima nuclear accidents, source term determination including release rates and release curves, becomes a key issue in nuclear accident impact assessment. When faced with monitoring device failure, use of off-site monitoring data and atmospheric dispersion model simulation results to inverse the source term has been proven an effective means in many case studies.
In this paper, we have developed a powerful source term inverse system based on constrained least squares algorithm approach and Gaussian atmospheric diffusion model, and we carried out 15 SF6 field trace experiments at a coastal nuclear power plant site in China from August 22, 2007 to August 31,2007, to validate the source term inversion model. From the 15 simulation results, the 5th,8th and 9th inverse results are the best, and the relative error of inverse source term and true source term is between 10.12% and 16.42%, this is mainly because these three experiments are the most successful, and effective sample points of three experiments are above 25, its distribution is also uniform. It is shown that accuracy, number and distribution of the on-site monitoring will directly affect the inversion source term results.
In general, results of source term inversion model are close to the field experiment, within the same magnitude, which shows that constrained least squares algorithm method and Gaussian atmospheric diffusion model can be applied in nuclear accident source term determination. Results of this study show that constrained least squares method is simple and highly efficient and has a high application value.
The design of a class of prefect reconstruction two-channel FIR and wavelets filterbanks using constrained least squares method and semidefinite programming
