Comparison of Harmonic and Time Marching Unsteady Computational Fluid Dynamics Solutions With Measurements for a Single Stage High Pressure Turbine

Abstract

The unsteady aerodynamics of a singlestage highpressure turbine has been the subject of a study involving detailed measurements and computations. Data and predictions for this experiment have been presented previously, but the current study compares predictions obtained using the nonlinear harmonic simulation method to results obtained using a timemarching simulation with phaselag boundary conditions. The experimental configuration consisted of a singlestage highpressure turbine (HPT) and the adjacent, downstream, lowpressure turbine nozzle row (LPV) with an aerodynamic design that is typical to that of a commercial highpressure ratio HPT and LPV. The flow path geometry was equivalent to engine hardware and operated at the proper designcorrected conditions to match cruise conditions. The highpressure vane and blade were uncooled for these comparisons. All three blade rows are instrumented with flushmounted, highfrequency response pressure transducers on the airfoil surfaces and the inner and outer flow path surfaces, which include the rotating blade platform and the stationary shroud above the rotating blade. Predictions of the timedependent flow field for the turbine flow path were obtained using a threedimensional, Reynoldsaveraged Navier–Stokes computational fluid dynamics (CFD) code. Using a two blade row computational model of the turbine flow path, the unsteady surface pressure for the highpressure vane and rotor was calculated using both unsteady methods. The two sets of predictions are then compared to the measurements looking at both timeaveraged and timeaccurate results, which show good correlation between the two methods and the measurements. This paper concentrates on the similarities and differences between the two unsteady methods, and how the predictions compare with the measurements since the faster harmonic solution could allow turbomachinery designers to incorporate unsteady calculations in the design process without sacrificing accuracy when compared to the phaselag method.