【摘 要】
：
In recent years,artificial neural networks (ANNs) and deep learning have become increasingly popular across a wide range of scientific and technical fields,including fluid mechanics.While it will take time to fully grasp the potentialities as well as the
【机 构】
：
Department of Mathematics,University of Oslo,Oslo,Norway;Research Center of Fluid Structure Interact
论文部分内容阅读
In recent years,artificial neural networks (ANNs) and deep learning have become increasingly popular across a wide range of scientific and technical fields,including fluid mechanics.While it will take time to fully grasp the potentialities as well as the limitations of these methods,evidence is starting to accumulate that point to their potential in helping solve problems for which no theoretically optimal solution method is known.This is particularly true in fluid mechanics,where problems involving optimal control and optimal design are involved.Indeed,such problems are famously difficult to solve effectively with traditional methods due to the combination of non linearity,non convexity,and high dimensionality they involve.By contrast,deep reinforcement learning (DRL),a method of optimization based on teaching empirical strategies to an ANN through trial and error,is well adapted to solving such problems.In this short review,we offer an insight into the current state of the art of the use of DRL within fluid mechanics,focusing on control and optimal design problems.
其他文献
An elastic model that combines the dead-end and the orifice at the downstream end is proposed in order to study the hydraulic transients during the rapid filling in small-scale pipelines with submerged outlet.In this model,the unsteady water flow is solve
Based on the computational fluid dynamics (CFD) and the experiment technology, this paper presents a new type of the twin-screw pump for the water-supply and studies its hydraulic performance with the hydraulic performance test. The internal flow characte
The wind turbines in the world\'s highest wind farm with an altitude of over 4 000 m in Naqu county of China were put into operation in 2014.Compared with the wind turbines at lower altitudes,the rated wind speed of these wind turbines are much larger,a
To study the drag reduction of a jet surface,an experiment is conducted with a gyrorotor model.The aim of this study is to analyze the influence of the jet aperture,the jet angle and the rotational velocity coupled with the jet velocity on the friction to
The random walk particle tracking (RWPT) method is compared with the Eulerian methods in investigating pollutant transport in depth-averaged flows.As a typical representative of the Eulerian model with high performance,the MacCormack scheme with Total Var
The paper presents a direct numerical simulation (DNS) for the drag-reducing channel flow using the Giesekus model with variable parameters.It is assumed that the relaxation time in the constitutive equation is varied depending on the local shear rate.The
The unsteady bearing forces of a propeller can be significantly affected by external transverse excitations.This paper numerically studies the unsteady variation characteristics of the bearing forces of a three-blade conventional propeller with external t
The compressible mixing layer is an important physical model to describe the mixing enhancement in scramjet combustors.The downstream coherent structures are normally regarded as the main contribution of the entrainment in the compressible mixing layer.In
The cavitation is directly related to the safety of many large water conservancy projects.There are two types of cavitation bubbles,one is generated by the gas nucleus,called the non gas-beating cavitation bubble,and another is generated by the gas bubble
Due to its multiscale and multi-layer natures,the coherent structures of turbulent in the open channel flow is complex and difficult to be visualized for understanding its evolution.In this paper,five types of methods for the vortical structure in the flu