Abstract

There has been a need for improved flow prediction methods for low pressure turbine (LPT) blades operating at high altitudes with a reduced inlet Reynolds number. These conditions present an increased amount of laminar-to-turbulent transitional flow within the boundary layers on the LPT blade surfaces. Also, boundary layer separation is more likely to occur within the flowfield of the LPT stages due to the lower freestream velocities in the regions of adverse pressure gradients on the suction surfaces. More accurate predictions of aerodynamic losses due to low Reynolds effects are needed for CFD to provide more accurate input to the design process for LPT stages operating at high altitudes. Steady flow CFD simulations of flow in a multistage LPT geometry were completed at nominal and high altitude conditions with the conventional Spalart-Allmaras turbulence model and a recentlydeveloped three-equation eddy-viscosity type transitional flow model. These models were used in combination with a mixing plane model for the simulation of flow through a three stage low pressure turbine. Flow visualizations were completed using surface flow and streamline calculations to help identify vortical structures present within the flowfield. Also, the total pressure loss coefficient was calculated for each blade row. Qualitative comparisons indicate the amount and the location of the flow separation differed significantly depending on the chosen turbulence model. Overall, the high altitude condition had an increased amount of separated flow compared to the nominal altitude condition resulting in an increase in the loss coefficient. The altitude effect on the laminar-to-turbulent transition location was studied using the three-equation model. The model provided a more detailed understanding of the aerodynamic loss mechanisms present in low Reynolds number flows, since it accounted for transitional boundary layer flow effects. Based on the these results, the CFD using the three-equation model has the potential to be a more effective method for turbine flow prediction at low Reynolds numbers compared to conventional RANS turbulence models.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call