Discussion on the Application of Deep Learning Algorithms and CAD Systems in Industrial Design

The traditional industrial product design process relies on manual modeling, which has some problems, such as high cost, and the knowledge of internal geometry and relationship of the model has not been reused. Based on this, this article discusses the application of DL (Deep Learning) algorithm and CAD system in industrial design, and proposes a 3D modeling algorithm based on DL model. In order to improve the training effect of the model, the training set is preprocessed, including algorithm detection and data enhancement. The depth model obtained by transfer learning training is used in industrial design. The test results show that DNN (Deep Neural Network) can meet the accuracy requirements after 19 iterations, and the network training time is relatively short. And the training value is in good agreement with the target value. Compared with the traditional modeling methods, the 3D modeling algorithm based on DL model learns the layout and dimension features of previous 3D models through graph automatic encoder, generates the layout features of parts of 3D objects, then synthesizes the details of parts, and gradually completes the 3D generation task from coarse to fine. The research in this article provides a new idea for the application of DL algorithm and CAD system in industrial design. © 2024, CAD Solutions, LLC. All rights reserved.