Numerical Investigation on the Pressure Perturbation of a Transonic Aeroelastic Compressor Rig

A linear aeroelastic rig was designed to serve as a validation facility for forced response research in axial compressors. This rig focuses on the study of aerodynamic damping at high-reduced frequencies when the central blade oscillates in a controlled manner. The steady and unsteady flow field of the rig has been numerically investigated previously. This paper aims to understand the pressure perturbation throughout the domain, which will give more insight into the actual transonic linear cascade rig and will improve the design of the test rig in some aspects. Such aspects are the tip gaps of the side blades that with the previously optimized upper inclined wall and tailboards are of importance in keeping flow periodicity. The design process has shown that these components do influence the unsteady pressure distribution, which might reduce the accuracy of the experiments. Thus, the aeroelastic cascade itself needs detailed investigations on the systematic accuracy losses in the unsteady testing due to the current design layout. This paper presents the impacts and mechanisms of tip gaps, upper inclined wall, and bottom tailboard on the test rig's aeroelastic performance. It provides a new validation reference for cascade test results with a perspective on the numerical studies, which can introduce a correction procedure on the existing test rig or on future cascade designs. INTRODUCTION Since the developed in 1947, the transonic compressor generally becomes the most significant component within the gas turbine due to its high mass flow and high-pressure ratio per stage [1]. Modern transonic compressors create new challenges. The new higher loaded and thinner blade designs have led to the compressor blade to be more prone to high cycle fatigue, which can derive in blade failure. Moreover, the high sensitivity of the transonic flow subjected to a narrow operating range increases the difficulty of their design. Thus, the prediction of forced response behaviour with the inherent high reduced frequency is of great importance in the transonic compressor design. The forced response performance of the blade is calculated by solving the structural dynamic equations with the aerodynamic forcing and damping as inputs. One of the critical steps is to calculate the aerodynamic damping accurately. The most accessible approach is the controlled flutter method. This approach controls one or several blades vibrating as the unsteady pressure is obtained on the blade surface. Typically, this can be done in the traveling wave mode or the influence coefficient method. The aerodynamic work can be directly calculated when the traveling wave mode method is used. All blades oscillate with a specific phase shift while the unsteady pressure and the blade motion of a blade are calculated. The advantage of this method is that the blade can vibrate at different nodal diameters. In contrast, in the influence coefficient method, only one blade vibrates in a rigid body mode or with a prescribed mode shape at a specific natural frequency, and the unsteady pressure is calculated on all the blades. The blade’s unsteady pressure fluctuation is influenced by the vibration of the moving blade. With the assembly of the influence coefficient matrix, the aeroelastic performance can be finally analysed. Panovsky et al studied that the coupling behaviour is mainly on the vibrating blade and the nearby two blades [2]. Theoretically, a three-blade cascade could be satisfied as long as the cascade flow is near periodic. The linear cascade is the simplified model of the annular cascade. The annular cascade is normally recommended because the inlet and outlet flow varies along the span according to the twist angle. In the flow through a linear cascade, there is no variation in the primary flow angle at the exit of the cascade [3]. As a consequence, the secondary flow in the annular cascade is closer to the real compressor flow. This advantage of the annular cascade is not present in the current study. The test rig aims to simulate the rotor flow with a stator cascade. Therefore, the inlet flow angle needs to be determined by combining the stagger and incidence angles. Compared with the annular cascade, the linear cascade represents a more robust design. Besides, the blade profile changes significantly along the span in the modern compressor rotor. Typically, the mid-span profile is designed for high subsonic flow, while only a small tip part for transonic flow. Since the research only focuses on the transonic region, the annular cascade needs