Smarter Simulation for More Efficient Compressor Design

What is it about?

What is this work about? This study shows how high-fidelity CFD simulation and automated design optimization can improve the aerodynamic performance of a compressor stator, a key component in turbomachinery and jet engine compressors. The work focuses on the TU Berlin TurboLab stator, where complex three-dimensional flow effects near the hub create corner separation and pressure losses. First, the baseline CFD model was validated against available experimental and published numerical data to confirm that the simulation could capture the important flow physics. Then, a fully automated optimization workflow was built for geometry parameterization and Flow360 for CFD simulation. The stator vane geometry was described using 55 design variables, allowing the optimizer to explore a wide range of realistic blade shapes. A surrogate-assisted multi-objective genetic algorithm was used to search for designs that reduce both total pressure loss and exit-flow-angle deviation. The final optimized designs showed clear aerodynamic improvements compared with the baseline geometry.

Featured Image

Photo by Anthony Maw on Unsplash

Photo by Anthony Maw on Unsplash

Why is it important?

Improving turbomachinery efficiency is important for reducing fuel consumption, emissions, and operating cost in aviation and energy systems. Even small aerodynamic losses inside compressors can affect the overall performance of gas turbines and aircraft engines. This work demonstrates that automated CFD-based design optimization can be used not only to evaluate a design, but to actively discover better designs. By combining a flexible CAD parameterization, fast CFD simulations, and multi-objective optimization, the workflow explored hundreds of design variants and identified improved stator geometries within a practical engineering timeframe. The results show that passive geometric optimization can reduce pressure loss, improve outlet flow direction, and reduce corner separation without adding active flow-control systems. This makes the approach valuable for industrial turbomachinery design, where performance gains must be achieved while respecting manufacturing and assembly constraints. Key takeaways 1. The baseline Flow360 simulation captured the stator corner separation with good agreement against experimental oil-flow visualization and published CFD data. 2. A 55-variable CAD-based stator parameterization was used to explore a broad design space. 3. More than 383 CFD simulations were completed through an automated optimization workflow. 4. The optimized Pareto-front designs reduced total pressure loss by more than 7%. 5. The optimized designs also reduced exit-flow-angle deviation by more than 16%. 6. The best designs reduced the size of the separated corner-flow region compared with the baseline geometry. 7. The study shows that CFD-driven multi-objective optimization can be a practical tool for improving turbomachinery components.

Perspectives