An approximate solution corresponding to a quasi-stationary state during nitriding

This work presents the computing of diffusion coefficients through an inverse problem of coefficient identification in a diffusion model, which considers several layers and Stefan type conditions. An approximate solution is obtained for the model, using qualitative and quantitative information from experimental results. To study the inverse problem we assume that we have information about the nitrogen concentration at different depths in the solid at the same time during the post-discharge nitriding process. To develop a stable numerical algorithm for the identification of the diffusion coefficients, a functional is built. This functional measures the deviation between the data theoretically obtained from the approximate solution of the model and the experimental data.