Modern design and analysis techniques for turbomachinery rely heavily on CFD backed up by experimental validation work. Validations have recently been performed by engineers at Fluent Europe for a pair of test cases defined by the Turbomachinery Special Group of the European Research Community on Flow Turbulence and Combustion (ERCOFTAC). The two test cases modeled were:

• steady flow in the Durham linear turbine cascade (ERCOFTAC F1)
• unsteady blade interaction in a 1.5 stage Aachen turbine (ERCOFTAC U1)

The Durham rotor test case was modeled as a steady state simulation using a combination of GAMBIT and FLUENT. A 3D hexahedral paved mesh of 300,000 elements with fine boundary layer mesh resolution was used. The RSM turbuence model with the Speziale quadratic pressure-strain model and nonequilibrium wall functions was chosen in the hope of achieving the most accurate results. When compared to the experiment, the CFD simulations were accurate to within 0.5 percent for static pressure coefficients at various spanwise positions.

Static pressure plot around the blade surfaces (13mm from the end wall) in the Durham turbine

(click image for enlarged view)

Experimental (left) and CFD prediction (right) for total pressure loss in the Durham turbine

The Aachen turbine CFD simulation in FLUENT used the time-dependent sliding mesh approach to pick up strong rotor-stator interaction effects. A hybrid hexahedral/tetrahedral mesh of nearly 900,000 elements was used for the rotor and stators in three adjacent fluid zones. When compared with experimental data, the CFD predictions for Mach number data agreed to within 2 percent. A Spalart-Allmaras RNG k-e turbulence model with a vorticity-based production option was used in the simulation.

Mach number contour plots on the Aachen turbine hub and blades

For these two very demanding ERCOFTAC test cases, FLUENT has shown good comparisons with experimental data. In addition, the speed and accuracy of hybrid unstruc- tured mesh generation and solution has been validated by these studies, with turnaround times of days rather than weeks.

References:

1. Gregory-Smith, D.G. "ERCOFTAC Seminar and Workshop on 3D Turbomachinery Flow Prediction", ERCOFTAC Bulletin, pp. 18-22, 1995.
2. Gregory-Smith, D.G. and Yan, J. "CFD Simulations of 3-Dimensional Flow in Turbomachinery Applications", ERCOFTAC Turbomachinery Flow Predictions VII Workshop in LaClusaz, France, March, 2000.
3. Walraevens, R.E. and Hallus, H.E. "Stator-Rotor-Stator Interaction in an Axial Flow Turbine and its Influence on Loss Mechanisms", AGARD PEP 85th Symposium on "Loss Mechanisms and Unsteady Flows in Turbomachinery", 1995.

FluentNEWS Supplement