Session: 35-02 High-Fidelity CFD

Paper Number: 84281

84281 - Design Sensitivity of a 1-1/2 Stage Unshrouded High Work Turbine Using Very-Large Eddy Simulations 

The variability in gaps, fillets and geometry details in turbomachines can be higher than expected due to the wide range of operating conditions or the tolerances in manufacturing processes. Uncertainty predictions thanks to numerical methods are crucial to estimate correctly the performance robustness of a given design and prevent component failures.

The paper presents a comparison between experimental measurements and simulations of a 1-1/2 stage unshrouded high work turbine. The experimental investigations were conducted by the Turbomachinery Laboratory of ETH Zurich. The data was obtained from steady and unsteady probe measurements that were performed in four axial planes between stator and rotor rows.

Previous simulations performed by Gonzalez-Martino and Gautier (2016) compared averaged quantities along the turbine stage such as pressure drop, the degree of reaction, the loading coefficient, and the flow coefficient; averaged midspan inlet and exit angles for each turbine blade rows; and flow distribution at four axial planes between the rotor and stator rows. Simulations were performed using the commercial CFD solver PowerFLOW based on the Lattice Boltzmann (LB) method to compute unsteady flows ranging from low-subsonic to transonic speeds. The turbulent flow fluctuations are resolved up to a certain scale using a Very Large Eddy Simulation (VLES) approach. This work demonstrated the accuracy of the Lattice Boltzmann approach and its capability to tackle full 360º geometry in a cost efficient manner.

The current paper is a continuation of the previous work addressing some of the uncertainties in test conditions and deviations in the geometry. This will first provide a better understanding of the impact of uncertainties on operating conditions like inlet pressure and temperature profiles. Second, this study will study the boundary layer tripping to trigger resolved turbulence upstream the turbine and its impact on corner stall phenomena. Next, some geometry variants will be considered: variations in tip clearance and airfoil root fillet will impact the development of the secondary flow structures and thus impact on stage losses. These geometry variations could be due to operating clearance uncertainties or manufacturing process accuracy, for example. It should also be noted that ETH Zurich experiment did not specify the fillet geometry, where previous simulation did not include a fillet radius. Finally, the study will assess the effect of cavities under stator vanes on secondary flows. Cavity sizes and shapes will be as accurate as possible based on previous publications on LISA test rig.

All these different geometry modifications may provide a useful insight on the performance and unsteady flow sensitivities of the LISA turbine to design modifications and experimental uncertainties.

Presenting Author: Ignacio Gonzalez-Martino Dassault Systèmes

Presenting Author Biography: Ignacio holds a Ph.D. in numerical fluid dynamics from University Pierre et Marie Curie. He has been working at Exa Corporation and then Dassault Systèmes for more than 10 years in aerospace applications. He has developed an expertise on aerodynamics, aerothermal and aeroacoustics simulations of turbomachines.

Authors:

Ignacio Gonzalez-Martino Dassault Systèmes