Generative AI meets CAD: enhancing engineering design to manufacturing processes with large language models
Amirmohammad Daareyni, Antti Martikkala, Hossein Mokhtarian, Iñigo Flores Ituarte
Abstract
Abstract As a key technology in the field of Artificial Intelligence (AI), large language models (LLMs) have gained substantial attention due to their wide range of capabilities in understanding and generating human-like speech. These advancements extend opportunities beyond text generation offering solutions for more complex engineering tasks. Despite extensive exploration of LLM capabilities in multiple fields, there is a noticeable lack of comprehensive studies within the engineering sector. This study aims to address this gap by investigating the potential of LLMs in solving engineering design problems. This study presents a framework for LLM-assisted CAD that employs advanced artificial intelligence systems, specifically GPT-4o model, to provide users with the ability to generate and manipulate 3D models through text commands, visual inputs such as blueprints, images, voice, and any other type of general or professional information about the desired model. By integrating LLMs with CAD software and incorporating image recognition and voice detection capabilities, we aim to lower the barrier to entry for 3D modeling, making it more user-friendly and accessible to non-professionals. The proposed framework employs LLM to process user inputs (text, images, and/or voice) and generate the code corresponding the CAD model. A closed feedback loop ensures that the generated models are refined iteratively until they meet the user’s specifications. This research demonstrates the potential of generative AI-assisted design (GAD) in simplifying complex design processes and providing valuable insights into areas where LLM enhanced processes can assist design to manufacturing workflows.