(711B) Machine Learning for Predicting the Viscosity of Binary Liquid Mixtures

Viscosity is a critical parameter in process engineering and a key design objective for application areas ranging from coatings and lubricants to the personal care and pharmaceutical industries. The lack of reliable and general methods for predicting the viscosities of mixtures creates a barrier for formulation design and process engineering. In this work, we developed a graph-based neural network architecture and applied it to the problem of predicting the viscosity of binary liquid mixtures as a function of composition and temperature. To obtain a high-quality training dataset, we developed an automated curation pipeline and applied it to a large dataset collected from the literature by the National Institute of Standards and Technology (NIST) to be used as training data. The resulting model accurately predicts viscosity, with an MAE of 0.043 and an RMSE of 0.080 in log cP units. To improve the dependability of the model in application settings, we also developed a “reliability evaluator” that determines whether or not a prediction is reliable based on the variance between an ensemble of models. For predictions considered “reliable” (80% of the test set), the model performs sufficiently well that its error falls within the experimental uncertainty of the training data, with an MAE of 0.029 and an RMSE of 0.047. Overall, this work provides 1) a large set of curated viscosity data that can be used for future machine learning efforts, 2) a new, graph-based deep learning approach for predicting the viscosity of binary mixtures that inf, and 3) an illustrative case study for understanding the role that deep learning can play in achieving accurate and reliable property prediction.