Scale-resolving Simulations ?for Innovations in Turbomachinery Design (SciFiTurbo)

SCI-Fi-Turbo旨在通过将高阶尺度分辨模拟(SRS)和优化方法改进并集成到标准工业工作流程中，从而彻底改变航空发动机设计流程。SRS是开发超高效推进系统的关键推动因素，这些系统到2035年大幅减少温室气体排放，并实现欧盟到2050年实现气候中立的目标。这些改进将提高设计过程能力并缩短产品开发周期。未来的发动机概念需要开拓设计空间，并在所需的时间框架内解决当今标准工业设计流程无法解决的复杂设计问题。为了在发动机设计中实现必要的阶跃变化，设计方法也需要类似的阶跃变化。SCI-Fi-Turbo通过利用三项基础技术的机会来满足这一迫切需求：高性能计算、高阶数值方法和AI/ML。这一组合用于实现和演示两个关键的进步。首先，为现代CPU/GPU硬件建立了高度集成的高阶SRS设计流程，满足了健壮性、准确性和周转时间的要求。它将在工业层面上提供更多的功能和有效性，并为业界采用基于SRS的设计铺平道路。该方法的高精度还将减少低TRL测试的需要，并使新的概念和延长的运行条件成为可能。其次，开发了一种SRS辅助的多保真、数据驱动的优化框架，该框架嵌入和利用了高精度高阶SRS的优势，同时利用AI/ML方法来提高低保真模拟的预测能力，最大限度地提高整体过程的精度和速度。专门的实验支持技术进步，将使净零排放发动机的设计在适当的时候成为可能，并为航空业的数字化转型做出贡献。

Sci-Fi-Turbo aims to revolutionise the aero engine design process by advancing and integrating high-order scale-resolving simulations (SRS) and optimization methodologiesinto standard industrial workflows. SRS are a key enabler for developing ultra-efficient propulsion systems that drastically reduce GHG emissions by 2035 and achieve the EU's target to be climate-neutral by 2050. The advancements will boost design process capabilities and reduce product development cycles. Future engine concepts require opening up the design space and solving complex design problems out of reach for today's standard industrial design processes within the required timeframe. To achieve the necessary step change in engine design, a similar step change is needed for the design approach. Sci-Fi-Turbo fills this urgent need by exploiting opportunities in three foundation technologies: High-performance computing, high-order numerical methods, and AI/ML. The combination is used to implement and demonstrate two key advancements. First, a highly integrated high-order SRS design process is established for modern CPU/GPU hardware, meeting robustness, accuracy, and turnaround time requirements. It will provide increased functionality and effectivity at an industrial level and pave the way for the uptake of SRS-based design by the industry. The high accuracy of the methodology will also reduce the need for low-TRL testing and enable new concepts and extended operating conditions. Second, an SRS-assisted multi-fidelity, data-driven optimisation framework is developed, which embeds and exploits the advantages of highly accurate high-order SRS while leveraging AI/ML methods to increase the predictive capability of lower-fidelity simulations and maximize overall process accuracy and speed. Dedicated experiments support the technology advancement and will enable the design of net-zero-emission engines in due time and contribute to the digital transformation of the aviation industry.