Home

AI Surrogate Model Trial Platform

Reduce the time and cost of CAE¹ analysis via the easy introduction of surrogate models².

Challenges

It takes time to master CAE analysis and the computation is costly.

Building surrogate models requires learning costs at the time of introduction and specialized knowledge in machine learning for model construction and validation.

The benefits of AI Surrogate Model Trial Platform

Supporting CAE users to introduce and validate AI surrogate models (PiNNs³), and reducing barriers to adoption

• Assisting the efficient construction of optimal models, as well as the evaluation of learning and inference results by defining governing equations, handling boundary conditions and optimizing hyperparameters.

• The use of surrogate models to reduce simulation time and computational costs delivers tangible savings and accelerates development.

[Application Areas] Fluid analysis, structural analysis, electromagnetic wave analysis, etc.

Features of the technology

Support for the introduction of surrogate models using PiNNs (such as problem definition, visualization, and hyperparameter tuning).

Optimized for the energy-efficient processor FUJITSU-MONAKA⁴ for next-generation data centers.

Contact Us

For a demonstration or to test Fujitsu Kozuchi, please get in touch

---

1. CAE (Computer Aided Engineering): Performing technical calculations, simulations, and analyses on a computer ↩

2. Surrogate Model: An approximate model built using machine learning, which serves as an alternative to complex simulations, enabling significant reductions in computation time and cost ↩

3. PiNNs: Neural networks that integrate physical laws, capable of making physically plausible predictions even with limited data ↩

4. FUJITSU-MONAKA : Available in 2027 ↩