Solids transport models are used to predict the fluid velocity required to transport solid particles in hydraulic and pneumatic systems. It is important that the processes in these applications are designed and operated at a sufficient fluid velocity to avoid solid deposition. Mechanistic models are used to provide a reasonable estimate for the minimum fluid velocity needed to transport the particles. However, those models are commonly applicable in their respective ranges of data fitting; and are limited by the applicability of the empirically based closure relations that are a part of such models. Artificial Intelligence (AI) and Machine learning (ML) offer a wealth of techniques to extract information from data that can be translated into knowledge about the underlying fluid mechanics. The purpose of this work is to investigate the use of several ML models to predict the critical velocities of various single-phase carrier fluids in horizontal and inclined flow conditions.

A frame of work is developed to predict critical velocities in pipes via machine learning, using accessible parameters as inputs, namely, liquid and particle properties and inclination angles. The ML algorithms are trained on a large dataset (more than 1400 data points) of critical velocities in single-phase carrier fluid that is collected from open source: articles and dissertations. The trained algorithms are Elastic Net, Support Vector Machine, Random Forest, and Extreme Gradient Boosting Decision Trees. Moreover, the influence of key features in critical velocity prediction was identified by the applied algorithms. Finally, the predictive abilities of the models are cross-compared and were further validated by comparing its performance with well-established mechanistic models based on empirical correlations. The ML approach is observed to have superior performance to other models across a wide range of flow conditions and inclination angles.

This content is only available via PDF.
You can access this article if you purchase or spend a download.