Building upon our multi-objective optimization framework, the 2026.1 release focuses on scalability and universal integration. A key highlight is the exploitation of inverse design-based blade parametrization to provide the most effective means of achieving physics-enhanced machine learning for turbomachinery.
Unlike traditional CAD-based methods that require massive datasets to "learn" physics, our 3D inverse design approach uses aerodynamic loading parameters to inherently embed physics into the machine learning process. This enables the generation of highly accurate surrogate models via the Reactive Response Surface (RRS) with a fraction of the traditional simulation overhead.
TURBOdesign Suite 2026.1 release extends Fluent CAE integration to all applications using Fluent meshing and makes the RRS technology compatible with any customer CAE platform. To further accelerate the design cycle, we have introduced native support for submitting CAE jobs to remote clusters, allowing for massive parallelization of optimization tasks.
Finally, we introduce critical enhancements to TURBOdesign Pre. Designers of air compressors, chillers, and heat pumps can now use a specialized Inlet Guide Vane (IGV) model for centrifugal compressors. Additionally, we have expanded Design of Experiment (DoE) support to centrifugal pumps, ensuring that the earliest stages of design are more robust and data-driven than ever before.
