Optimal experimental design methods for the parameterization of entropy-scaling transport property models

Accurate transport property models for fluids are critical in science and engineering. Both physical and computer experiments for determining transport coefficients are expensive, making careful experimental design essential. In this work, a new model-based optimal experimental design (MbOED) method is proposed that uses an entropy scaling (ES) model in combination with an equation of state. We demonstrate how the proposed MbOED + ES method can be employed for different transport properties of fluids, namely the shear viscosity, thermal conductivity, and self-diffusion coefficient. The MbOED + ES method is tested using substances from different chemical families, i.e. n[jls-end-space/]-butane, benzene, nitrogen, and 1-octanol. Also different equation of state models for describing the configurational entropy within ES are compared. Overall, the four substances show consistent trends: The most informative experiments are generally located at high configurational entropy, while notable differences between transport properties arise from their distinct entropy-scaling behavior. The coupled MbOED + ES method provides important new insights into the optimal design of experiment regarding the thermodynamic states favorable for ES model parameterization. The application of the new MbOED + ES method demonstrates how the experimental effort can be reduced.