Experimental Study on Impeller Blade Vibration During Resonance—Part I: Blade Vibration Due to Inlet Flow Distortion

Abstract

Forming the first part of a two-part paper, the experimental approach to acquire resonant vibration data is presented here. Part II deals with the estimation of damping. During the design process of turbomachinery components, mechanical integrity has to be guaranteed with respect to high cycle fatigue of blades subject to forced response or flutter. This requires the determination of stress levels within the blade, which in turn depend on the forcing function and damping. The vast majority of experimental research in this field has been performed on axial configurations for both compressors and turbines. This experimental study aims to gain insight into forced response vibration at resonance for a radial compressor. For this purpose, a research impeller was instrumented with dynamic strain gauges and operated under resonant conditions. Modal properties were analyzed using finite element method and verified using an optical method termed electronic-speckle-pattern-correlation-interferometry. During the experiment, unsteady forces acting on the blades were generated by grid installations upstream of the impeller, which created a distorted inlet flow pattern. The associated flow properties were measured using an aerodynamic probe. The resultant pressure fluctuations on the blade surface and the corresponding frequency content were assessed using unsteady computational fluid dynamics. The response of the blades was measured for three resonant crossings, which could be distinguished by the excitation order and the natural frequency of the blades. Measurements were undertaken for a number of inlet pressure settings starting at near vacuum and then increasing. The overall results showed that the installed distortion screens generated harmonics in addition to the fundamental frequency. The resonant response of the first and the second blade mode showed that the underlying dynamics support a single-degree-of-freedom model.