HP is relying on artificial intelligence to further develop its 3D printing technology. In collaboration with NVIDIA, the company is using the AI tool Modulus to improve the metal sintering process in 3D printing. The result of this cooperation is Virtual Foundry Graphnet, an AI-based system for predicting and optimizing the sintering process.
The new technology uses physically informed neural networks. These integrate physical laws into machine learning models, thereby increasing their efficiency and accuracy. Virtual Foundry Graphnet enables precise prediction of the behavior of metal powders during the printing process. This enables manufacturers to optimize production, reduce errors and improve the quality of parts.
A decisive advantage of the system is its speed. The complex sintering process is simulated in a matter of seconds. Conventional methods require significantly more time. This increase in efficiency makes its use in industrial production particularly attractive.
HP shares its AI algorithms with NVIDIA, making them accessible to other manufacturers. This open source approach promotes collaboration in the industry and accelerates innovation. Smaller companies also benefit from the advanced AI tools.
The technology was described in detail in a study entitled “Virtual Foundry Graphnet for Metal Sintering Deformation Prediction“. The authors, including researchers from HP and NVIDIA, show how the model predicts deformations during sintering at the voxel level. The accuracy is 0.7 micrometers for a single sintering step and 0.3 millimeters for a complete cycle.
HP’s Virtual Foundry Graphnet is part of a broader Digital Twin approach. This enables process engineers to optimize both design and process control parameters. The goal is to improve part quality and manufacturing yield.
Subscribe to our Newsletter
3DPresso is a weekly newsletter that links to the most exciting global stories from the 3D printing and additive manufacturing industry.
