Machine Learning Based Critical Velocity Predictions in Slurry Flows

This paper examines the application of machine learning models for the prediction of critical velocity in Newtonian and non-Newtonian slurry flows. Experimental data and sensor measurements from established literature are used to train and test the models. Data cleaning, preprocessing, descriptive statistics, comparative analysis, and error evaluation have been conducted to make the models reliable. Several multi regression models have been trained and evaluated using original and synthetic data sets to improve generalization. Results look promising in their predictive ability, where different performances are realized for diverse slurry types and experimental conditions.