Just another WordPress site - Ruhr-Universität Bochum

The selection and weighting of experimental and simulated datasets is a necessary step in the development of thermodynamic property models in the calculation of phase diagrams (CALPHAD) approach. Currently, this step requires painstaking and complicated evaluation of the reliability of datasets and thermodynamic consistency between them. In this work, we present two novel and independently developed statistical approaches to aid in this process by addressing outliers and performing automated dataset weighting. The first method, presented here for the first time, applies classical statistical techniques and commonly available optimization algorithms. The second method employs Bayesian statistics via numerical sampling techniques. We compare the strengths and weaknesses of the two approaches through an assessment of the specific heat of aluminum and hafnium metal versus temperature for several experimental datasets. We then compare the weightings of each dataset versus a number of metrics employed by experts to evaluate the reliability of datasets.