Effects of multiple incident shock waves on the flow in a transonic turbine cascade

Turbine aerodynamic designers are currently focusing on unsteady passage flow to increase turbine performance. In particular, for high pressure turbine stages the effects of wakes and shocks shed from an upstream blade row on the downstream blade row need to be understood. Also, experimental data is needed for comparison with unsteady three-dimensional turbine stage calculations.

Previous simulations of the unsteady shock/wake inlet flow field for a turbine rotor or stator used a rotating disk with radial bars upstream of a linear cascade. An alternate method of shock generation is developed here using a capped shock tube with multiple outlets to get a traveling system of three shock waves. Different lengths of tubing are used to get time delays between the shocks, which are then introduced at the top of a linear cascade of turbine blades and travel downwards (tangentially) along the leading edge. Advantages of this method include the absence of wakes and excellent two-dimensionality of the inlet shock waves. The period of the incoming shocks is easily adjustable to simulate different Strouhal numbers.

Unsteady measurements of upstream total pressure, blade static pressures, and uncorrected downstream total pressure are made for a transonic mean flow with introduction of traveling shocks at M=1.3. An analytical solution (Bach and Lee, 1970) for the decay of cylindrical shock waves is used to estimate the behavior of flow variables other than pressure at the cascade inlet. The unsteady total pressure loss of the blade passage and the unsteady blade forces are measured with one shock passing and with three shocks passing at periods of 0.055 and 0.200 milliseconds. Loss is estimated as the normalized difference in unsteady total pressures and blade forces are integrated from seventeen unsteady surface pressure measurements.

The Strouhal number for the 0.200 msec case is 2.9, which is typical of a high-pressure turbine nozzle or rotor. Periodic behavior in blade force and loss are observed for this case. Blade lift shows peak-to-peak variation of 6% and the estimated loss fluctuates by 100%. No change is observed in the average level of loss due to the incident shock waves.