Machine Learning Based Fluid-Transportation Monitoring and Controlling

Abstract

The discipline of fluid mechanics is developing quickly, propelled by previously unheard-of data volumes from experiments, field measurements, and expansive simulations at various spatiotemporal scales. The field of machine learning (ML) provides a plethora of methods for gleaning insights from data that can be used to inform our understanding of the fluid dynamics at play. As an added bonus, ML algorithms can be used to automate duties associated with flow control and optimization, while also enhancing domain expertise. This article provides a review of the background, current state, and potential future applications of ML in fluid mechanics. We provide an introduction to the most fundamental ML approaches and describe their applications to the study, modelling, optimization, and management of fluid flows. From the standpoint of scientific inquiry, which treats data as an integral aspect of modelling, experiments, and simulations, the benefits and drawbacks of these approaches are discussed. Since ML provides a robust information-processing framework, it can supplement and potentially revolutionize conventional approaches to fluid mechanics study and industrial applications.