Parameter Inversion of Tritium Migration Based on Parallel Genetic Algorithm

To calibrate model parameters of tritium migration in a test site of China, an intelligent parameter inversion model based on parallel genetic algorithm is built, a forward and inverse coupling program of radionuclide migration is designed, and the values of key parameters like hydraulic conductivity, dispersity and porosity are inverted automatically on a mainframe computer, by means of abundant observation data of tritium concentration. The inversion results accord with observation data well on the whole. Compared to manual adjustment method, this method has better overall convergence, higher calculated precision and efficiency, and less manpower cost. The results show that parallel genetic algorithm is feasible and valid in application to parameter inversion of tritium migration.