Related Experiment Video
Updated: Jul 2, 2025
Ultrasonic Welding of Thermoplastic Composite Coupons for Mechanical Characterization of Welded Joints through Single Lap Shear Testing
Published on: February 11, 2016
Application of Neural Network Models with Ultra-Small Samples to Optimize the Ultrasonic Consolidation Parameters for
Dmitry Y Stepanov1, Defang Tian2, Vladislav O Alexenko1
1Laboratory of Mechanics of Polymer Composite Materials, Institute of Strength Physics and Materials Science of Siberian Branch of Russian Academy of Sciences, 634055 Tomsk, Russia.
Abstract:
The aim of this study was to optimize the ultrasonic consolidation (USC) parameters for 'PEI adherend/Prepreg (CF-PEI fabric)/PEI adherend' lap joints. For this purpose, artificial neural network (ANN) simulation was carried out. Two ANNs were trained using an ultra-small data sample, which did not provide acceptable predictive accuracy for the applied simulation methods. To solve this issue, it was proposed to artificially increase the learning sample by including additional data synthesized according to the knowledge and experience of experts. As a result, a relationship between the USC parameters and the functional characteristics of the lap joints was determined. The results of ANN simulation were successfully verified; the developed USC procedures were able to form a laminate with an even regular structure characterized by a minimum number of discontinuities and minimal damage to the consolidated components.
