Deep-Learning-Based Predictive Architectures for Self-Piercing Riveting Process

Hyungson Ki,Dong-Hyuck Kam,Sehyeok Oh,Taek-Eon Jeong,Hyun Kyung Kim

doi:10.1109/access.2020.3004337

Abstract

Deep-learning architectures were developed for the self-piercing riveting (SPR) process to predict the cross-sectional shape from the scalar input of the punch force. Traditionally, the SPR process is studied using a physic-based approach, including finite element modeling, but in this study, a data-driven approach consisting of two supervised deep-learning models was proposed. The first model was used for data transformation from an optical microscopic image to a material segmentation map, which characterizes the shape and location of the two sheets and the rivet by applying a convolutional neural network (CNN)-based deep-learning structure. To validate the developed models, two types of sheet combinations were tested, namely, carbon-fiber-reinforced plastic (CFRP) and galvanized dual-phase steel (GA590DP) sheets, and steel alloy (SPFC590DP) and aluminum alloy (Al5052-H32) sheets. The transformation was performed with a mean intersection-over-union of 98.50% and a mean pixel accuracy of 99.78%. The next model, which was a novel generative model based on a CNN and conditional generative adversarial network with residual blocks, was then trained to predict the cross-sectional shape from the input punch force. The predicted cross-sectional shapes were compared with the experimental results of SPR. The overall accuracy was 94.20% for CFRP-GA590DP and 96.31% for SPFC590DP-Al5052, with respect to three key geometrical indexes, namely, rivet head height, interlock length, and bottom thickness.

Highlights

Self-piercing riveting (SPR) is one of the most promising techniques for joining sheet materials
The deep-learning model was designed to predict the VOLUME 8, 2020 cross-sectional shape from an input punch force. This is a novel generative model based on the convolutional neural network (CNN), called the conditional generative adversarial network [17], [18]; it comprised residual blocks [19], [20], where predictive segmentation images were generated from a scalar value
The optical microscopic (OM) inputs, their ground truths, and artificial intelligence (AI) prediction results are presented in the first, second, and third columns, respectively, and the last three columns are the segmentation images separated by channel, comprising the ground truths and the predictions

Summary

INTRODUCTION

Self-piercing riveting (SPR) is one of the most promising techniques for joining sheet materials. A novel deep-learning framework was developed to predict the cross-sectional shape of the SPR joint from the scalar input punch force. The deep-learning model was designed to predict the VOLUME 8, 2020 cross-sectional shape (material segmentation map) from an input punch force (a scalar value). This is a novel generative model based on the CNN, called the conditional generative adversarial network (cGAN) [17], [18]; it comprised residual blocks [19], [20], where predictive segmentation images were generated from a scalar value (namely, a scalar-to-seg generator). A simple case study was performed to validate the designed generator

SPR EXPERIMENT

RESULTS AND DISCUSSION

CONCLUSIONS