Abstract
A deep learning model was applied for predicting a cross-sectional bead image from laser welding process parameters. The proposed model consists of two successive generators. The first generator produces a weld bead segmentation map from laser intensity and interaction time, which is subsequently translated into an optical microscopic (OM) image by the second generator. Both generators exhibit an encoder-decoder structure based on a convolutional neural network (CNN). In the second generator, a conditional generative adversarial network (cGAN) was additionally employed with multiscale discriminators and residual blocks, considering the size of the OM image. For a training dataset, laser welding experiments with AISI 1020 steel were conducted on a large process window using a 2 KW fiber laser, and a total of 39 process conditions were used for the training. High-resolution OM images were successfully generated, and the predicted bead shapes were reasonably accurate (R-Squared: 89.0% for penetration depth, 93.6% for weld bead area).
Highlights
In recent decades, laser welding has been actively applied to the high-precision joining of metal parts in the automotive, electronics, aerospace, shipbuilding, and medical industries
For the first time, a novel deep learning framework was proposed for predicting optical microscopic (OM) image of the cross-sectional laser weld bead, from only two laser processing parameters
- A novel deep learning framework was presented for predicting high-resolution cross-section weld bead images in laser welding, from two laser processing parameters
Summary
Laser welding has been actively applied to the high-precision joining of metal parts in the automotive, electronics, aerospace, shipbuilding, and medical industries. For the first time, a novel deep learning framework was proposed for predicting optical microscopic (OM) image of the cross-sectional laser weld bead, from only two laser processing parameters (laser intensity I0 and beam interaction time ti). Our deep learning model can predict the weld bead in real image, i.e. including keyhole, heat affected zone, substrate, microstructures, porosity as well as the geometrical bead shape, which synthetically determine the mechanical properties and weld quality It can instantly generate multiple predictive bead images in a few seconds from the given input laser process conditions once training ends, so is very handy as well as practical (one can share the trained model online, using the open source deep learning libraries such as TensorFlow and PyTorch on GitHub). - A novel deep learning framework was presented for predicting high-resolution cross-section weld bead images in laser welding, from two laser processing parameters (laser intensity and beam interaction time). Inside of the generators was presented to get a better understanding about the predicting process
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
