
Prediction of Aerodynamic Flow Fields Using Convolutional Neural Networks
An approximation model based on convolutional neural networks (CNNs) is ...
read it

UNetBased Surrogate Model For Evaluation of Microfluidic Channels
Microfluidics have shown great promise in multiple applications, especia...
read it

Predicting Porosity, Permeability, and Tortuosity of Porous Media from Images by Deep Learning
Convolutional neural networks (CNN) are utilized to encode the relation ...
read it

Some limitations of norm based generalization bounds in deep neural networks
Deep convolutional neural networks have been shown to be able to fit a l...
read it

Visualizing 2D Flows with Animated Arrow Plots
Flow fields are often represented by a set of static arrows to illustrat...
read it

Towards highaccuracy deep learning inference of compressible turbulent flows over aerofoils
The present study investigates the accurate inference of Reynoldsaverag...
read it

Rapid feasibility assessment of components formed through hot stamping: A deep learning approach
The novel nonisothermal Hot Forming and cold die Quenching (HFQ) proces...
read it
Predicting the flow field in a Ubend with deep neural networks
This paper describes a study based on computational fluid dynamics (CFD) and deep neural networks that focusing on predicting the flow field in differently distorted Ushaped pipes. The main motivation of this work was to get an insight about the justification of the deep learning paradigm in hydrodynamic hull optimisation processes that heavily depend on computing turbulent flow fields and that could be accelerated with models like the one presented. The speedup can be even several orders of magnitude by surrogating the CFD model with a deep convolutional neural network. An automated geometry creation and evaluation process was set up to generate differently shaped twodimensional Ubends and to carry out CFD simulation on them. This process resulted in a database with different geometries and the corresponding flow fields (2dimensional velocity distribution), both represented on 128x128 equidistant grids. This database was used to train an encoderdecoder style deep convolutional neural network to predict the velocity distribution from the geometry. The effect of two different representations of the geometry (binary image and signed distance function) on the predictions was examined, both models gave acceptable predictions with a speedup of two orders of magnitude.
READ FULL TEXT
Comments
There are no comments yet.