Currently, pre-made generative AI products such as OpenAI’s GPT-3 for text generation, NVIDIA’s StyleGAN2 for image generation, and OpenAI’s DALL-E 2 for video creation based on written descriptions are highly prevalent.
Within the context of engineering design, modeling, and simulations, generative AI can be used to improve data inputs, generate scenarios, optimize processes, and generate synthetic data.
- By analyzing and enhancing data inputs used in simulations, generative AI can improve the accuracy and overall quality of simulations.
- Generative AI can also generate new scenarios and variations in simulations, enabling organizations to test different scenarios, identify potential issues, and make informed decisions based on the simulated results.
- Additionally, generative AI can optimize processes within simulations by learning from simulation results and automatically adapting and improving the simulated processes.
- Finally, generative AI can generate synthetic data that closely resembles real-world data, which can be used to augment existing datasets in simulations.
For instance, generative AI has been used in the automotive industry to optimize the design of car parts and reduce their weight while maintaining their strength. Audi was able to reduce the cycle time of its assembly line by 30% in 2023 by using generative design and simulation in the manufacturing industry.
Another German car manufacturer, BMW, has combined generative AI with additive manufacturing to create a new innovation. BMW used generative adversarial networks (GANs) to create a new version of its 3D-printed water pump pulley, resulting in a 48% weight reduction and a 25% increase in efficiency.
References:
- Antonucci, G., Bottino, A., & Culmone, R. (2021). A Review of Artificial Intelligence Techniques Applied to Simulation and Modelling for Manufacturing.
- Yu, L., Zhang, W., & Zhang, L. (2021). A Survey on Generative Adversarial Networks and Their Applications in Computer Vision.