CAMD for entrainer screening of extractive distillation process based on new thermodynamic criteria

Abstract

for entrainer screening of extractive distillation process based on new thermodynamic criteria. (2019) Chemical Engineering Research and Design, 147. 721-733.

Uncertainty estimation in deep learning‐based property models: Graph neural networks applied to the critical properties

“…Going forward for process systems engineering applications in general and property estimation in particular, the choice of which modeling approach, be it a traditional GC approach or a more sophisticated deep learning model, should be determined by the purpose of the engineering application. For screening/feasibility assessment of chemicals such as the work presented by Frütiger et al 1 and Cignitti et al, 12 a model with larger coverage is to be preferred such as the MPNN model with the bootstrap method for uncertainty estimation or the AFP model with LL‐dropout (as uncertainty may be large but we are assured the predictions are aligned with experimental data collected thus far). However, for applications requiring detailed design and optimization and where the model uncertainty and accuracy are more important, 56 a property model that can provide higher accuracy/lower uncertainty with affordable computational cost, is preferred such as the case with a GC model.…”

“…4 These models have been used to predict various properties such as the normal boiling point and critical point properties. 4,11 These models have also been applied to address computer-aided molecular design problems such as screening for entrainers in extractive distillation 12 and refrigerants. 13 One disadvantage of these models is their inability to extrapolate outside their domain of applicability since the models' ability to cover the chemical design space is highly dependent on the training sample.…”

Decrease in CO2 Emissions in Obtaining Polymer‐Grade Propylene by Extractive Distillation

“…It should have a high polarity and high boiling temperature compared to the components of the mixture to be separated, have a low solubility capacity, as well as a low environmental and toxicological impact [19,20]. Since the correct selection of the solvent plays an important role, computer-aided molecular design (CAMD) has gained importance, which allows finding the appropriate mass separating agent for a certain mixture that requires separating [21][22][23].…”

Evaluation of ionic liquid separation ability for the benzene–methanol mixture by extractive distillation

“…Special distillations, such as pressure swing distillation, azeotropic distillation and extractive distillation, are usually used to separate azeotropes . Extractive distillation is widely applied to separate azeotropes because of its advantages such as strong separation ability, simple operation and flexible selection of entrainer . Selecting entrainer is a key factor for extractive distillation .…”

“…[3][4][5][6][7][8] Extractive distillation is widely applied to separate azeotropes because of its advantages such as strong separation ability, simple operation and flexible selection of entrainer. [9][10][11][12][13] Selecting entrainer is a key factor for extractive distillation. [14][15][16][17][18][19][20][21] In the reports on the separation of benzene-methanol azeptrope by traditional organic entrainer, the azeotropic mixture can be completely separated by aniline, propyl butyrate, anisole, etc.…”