An Experimental Study of the Effects of Vane Count and Non-Uniform Vane Spacing on Rotor Resonant Response

Papa Aye Nyansafo Aye-Addo, Purdue University

Abstract

The objective of this research was to experimentally investigate forced response conditions of an embedded stage rotor (R2) in the Purdue 3-Stage Axial Compressor Facility. Different symmetric stator vane counts excitations of the R2 1T vibratory mode, different symmetric vane count excitations of the R2 1CWB and 2CWB vibratory modes, and non-uniform vane spacing (NUVS) excitations of the R2 1T mode were compared to improve the current understanding of multistage compressor forced response and to provide an invaluable data set for the validation of computational prediction tools. The effect of a reduced vane count upstream stator on the resonant response of R2 was first evaluated with an Agilis NSMS Blade Tip-Timing system. With a higher mechanical resonant crossing speed and lower blade solidity, the reduced upstream 38 engine order (EO) and 76EO excitations, yielded larger rotor 1T and 1CWB blade tip deflections, respectively, and larger S1 wake forcing functions, compared to the baseline downstream 44EO excitation. The larger blade tip deflections from the 38EO-1T and 76EO-1CWB resonant crossings compared with the 44EO-1T and 88EO-1CWB resonant crossings, suggest that the vortical forcing function has a stronger effect on the excitation of the 1T and 1CWB modes than the potential forcing function. The effect of non-uniform vane spacing on rotor resonant response was evaluated with six different NUVS configurations (18-20, 20-22, 18-19, 19-20, 18-22 and 19-22). Each NUVS configuration consisted of two 180° sectors with different vane counts on each half, with uniform vane spacing on a particular vane half. With a vane count difference of two between the two vane halves, the symmetric harmonic response from the symmetric excitation was completely reduced, and its vibrational energy was spread into ±1/rev, ±2/rev and ±3/rev neighboring frequencies at reduced amplitudes. This was the case for both the 18-20 and 20-22 NUVS configurations. In the case of only one vane count difference between the two vane halves (19-20 and 18-19 NUVS configurations), the rotor resonant response was not beneficial for blade vibration reduction. The rotor vibration due to the symmetric harmonic response was not reduced. Additionally, the ±1/rev excitations were increased to the level of the symmetric excitation vibration response. The case with a difference of four between the two vane halves showed the best reduction in R2 forced response. Only two engine orders corresponding to vane counts in the compressor were shown in the blade tip deflection results. The 18-22 case performed better than the 19-22 case because the 36EO had a larger reduction in blade tip amplitude compared with the 38EO excitation of the R2 1T mode. The blade tip-timing measurements with the introduction of NUVS configurations also showed that the installation of a particular vane half under the optical probes was necessary to measure the maximum blade vibration of the corresponding engine order.

Degree

M.S.A.A.

Advisors

Key, Purdue University.

Subject Area

Aerospace engineering|Mechanical engineering

COinS