Biorefinery processes are challenged to make effective use of commercial flowsheeting software. Challenges include the lack of property data, complexity of raw materials, and emerging non-conventional processes and technologies. Surrogate models could assist by combining data-based models (by means of surrogate models) with conventional models available from flowsheeting vendors. The challenge would be to translate experimental data into property and process parameters compatible with models used by commercial software. The paper introduces an iterative approach for the systematic evaluation of such parameters. Degrees of freedom include options for pseudo-components, thermodynamic methods, and process models. The approach is applied for the modeling of a real-life biorefinery. Three case studies demonstrate the potential of the proposed scheme. The first case study has fixed options for the property and process models, while the second one has them as a degree of freedom. The third case study extrapolates the resulted metamodels to different capacities and six different feedstock types (wheat and rice straw, wood, sugarcane bagasse, banana stem, and miscanthus). Degrees of freedom expand as the cases address limited options for surrogate models, gradually incorporating additional options and, apparently, better regression results. The proposed framework could embed additional levels of sophistication using machine learning and artificial intelligence technology. The emphasis of the paper stands on the methodological framework and its demonstration with real-life examples.
