Identification of the hydrogen diffusion parameters in bearing steel by Evolutionary Algorithm

Abstract A novel mathematical method for the identification of hydrogen diffusion parameters in metals has been implemented. This method is based on a clustering-based Evolutionary Algorithm which provides robust and efficient optimization of the diffusion parameters. It is assumed that hydrogen diffusion obeys the McNabb–Foster equations, and by using the Evolutionary Algorithm the 1D time-dependent solution is calibrated to the experimental curve measured by the two cells permeation test. The original McNabb–Foster diffusion model, including reversible and irreversible traps, was simplified to the models with only reversible or irreversible traps and the calibration quality was tested. Due to the random nature of Evolutionary Algorithms, multiple calibration sets were identified, which corresponds to the non-uniqueness of calibration by the McNabb–Foster diffusion model. The definition of the invariant diffusion parameters describing uniquely the material and the test is the main novelty of the proposed study.