Figure 2. The first panel shows the first step of the assessment process. The experimental data on a system is collected by materials experts. The diagram in this panel illustrates the phase equilibria data for the Al-Mg system. Note that the dashed-line is only a guide to the eye. After the materials scientist identifies all phases and their crystal structures for a given system, a model is assigned for the Gibbs energy of each phase. This model is basically a polynomial function of composition and temperature, and terms in the function varies from phase to phase. The free model parameters are fitted to the experimental data. This step is performed for all phases together, in other words, a multi-phase model optimization because the model parameters of all phases are interconnected to each other. A plot of Gibbs energy functions with optimized model parameters for three phases that exist at 730 K for the Al-Mg system is given in the second panel. With such critically optimized model parameters, we can then map the phase diagram through the minimization of the Gibbs energy function as shown in the third panel.