Searching for Sustainable Refrigerants by Bridging MolecularModeling with Machine Learning

Identificador:  imarina:9266829

Handle: https://hdl.handle.net/20.500.11797/imarina9266829

Autors:  Alkhatib, III; Albà, CG; Darwish, AS; Llovell, F; Vega, LF

Resum:
We present here a novel integrated approachemploying machine learning algorithms for predicting thermo-physical properties offluids. The approach allows obtainingmolecular parameters to be used in the polar soft-statisticalassociatingfluid theory (SAFT) equation of state using moleculardescriptors obtained from the conductor-like screening model forreal solvents (COSMO-RS). The procedure is used for modeling18 refrigerants including hydrofluorocarbons, hydrofluoroolefins,and hydrochlorofluoroolefins. The training dataset included sixinputs obtained from COSMO-RS andfive outputs from polar soft-SAFT parameters, with the accurate algorithm training ensured byits high statistical accuracy. The predicted molecular parameters were used in polar soft-SAFT for evaluating the thermophysicalproperties of the refrigerants such as density, vapor pressure, heat capacity, enthalpy of vaporization, and speed of sound. Predictionsprovided a good level of accuracy (AADs = 1.3-10.5%) compared to experimental data, and within a similar level of accuracy usingparameters obtained from standardfitting procedures. Moreover, the predicted parameters provided a comparable level of predictiveaccuracy to parameters obtained from standard procedure when extended to modeling selected binary mixtures. The proposedapproach enables bridging the gap in the data of thermodynamic properties of low global warming potential refrigerants, whichhinders their technical evaluation and hence theirfinal application