eduzhai > Applied Sciences > Engineering >

StressGAN A Generative Deep Learning Model for 2D Stress Distribution Prediction

  • Save

... pages left unread,continue reading

Document pages: 14 pages

Abstract: Using deep learning to analyze mechanical stress distributions has beengaining interest with the demand for fast stress analysis methods. Deeplearning approaches have achieved excellent outcomes when utilized to speed upstress computation and learn the physics without prior knowledge of underlyingequations. However, most studies restrict the variation of geometry or boundaryconditions, making these methods difficult to be generalized to unseenconfigurations. We propose a conditional generative adversarial network (cGAN)model for predicting 2D von Mises stress distributions in solid structures. ThecGAN learns to generate stress distributions conditioned by geometries, load,and boundary conditions through a two-player minimax game between two neuralnetworks with no prior knowledge. By evaluating the generative network on twostress distribution datasets under multiple metrics, we demonstrate that ourmodel can predict more accurate high-resolution stress distributions than abaseline convolutional neural network model, given various and complex cases ofgeometry, load and boundary conditions.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×