AI-ENHANCED COMPUTER-AIDED DESIGN (CAD) SYSTEMS: UTILIZING DEEP LEARNING FOR INTELLIGENT ENGINEERING PRODUCT OPTIMIZATION

Abstract

AI-enhanced computer-aided design (CAD) is transforming modern engineering by integrating deep learning models capable of automating geometry reasoning, detecting design inefficiencies, predicting structural performance, and generating optimal product configurations. This study proposes a hybrid deep learning framework that augments CAD workflows with intelligent optimization capabilities. The system combines convolutional neural networks (CNNs), graph neural networks (GNNs), and transformer-based geometric encoders to interpret product geometries, extract functional features, and recommend design modifications. A dataset comprising 45,000 mechanical components, assemblies, and parametric CAD models was used to train the model on structural behavior patterns, topology variations, manufacturability constraints, and stress-distribution profiles. Evaluation results show that the AI-CAD system enhances design decision-making through automated defect detection, material-usage reduction, performance prediction, and generative optimization. Compared to baseline CAD tools, the proposed system reduces design cycles by 38 percent, improves component lightweighting accuracy by 27 percent, and increases topology optimization efficiency by 41 percent. The findings confirm that deep-learning-based CAD systems can significantly accelerate engineering innovation by integrating intelligence, automation, and optimization into existing workflows.