Measurement point selection and modal damping identification for bladed disks

Abstract

A novel measurement point selection (MPS) technique for bladed disks (blisks) is presented and applied to a new modal damping identification method. When gathering data to be used for applications such as mistuning identification in blisks, it is important to measure points which provide sufficient and accurate information for the analysis. However, to reduce the experimental time and cost, the measurement points should be chosen optimally so that the minimum number of measurements have to be collected. This paper discusses a modified form of the effective independence distribution vector (EIDV) method presented by Penny et al. and adapted by Holland et al. The key novel aspect of the proposed method is that it uses only single sector-level calculations instead of the whole system. A residual weighting optimizes the MPS technique for noisy measurements. The method presented is equivalent to the full system EIDV method, but it decreases the computational cost, increases the robustness of the identification, and minimizes the measurement time. Also, a novel method to identify damping parameters for each mode in a frequency range of interest is presented. This method utilizes the proposed MPS technique to increase the accuracy of the identification. Measurement locations and modal damping results for a 30 degree of freedom system and a blisk with a complex geometry are presented. Using the proposed methodologies it is possible to obtain an accurate modal damping identification with a decreased computational and measurement cost.

Hybrid modal damping identification for bladed disks and blisks

“…This section summarizes the modeling approach used for modeling blisks with small mistuning and viscous modal damping. This description closely follows that of Holland et al (2012b). The starting point in modeling blisks is to exploit the fact that the mistuned modes for systems with small mistuning can be represented as a linear combination of tuned system modes (Yang and Griffin, 1997).…”

“…The types of excitation systems used for vibration testing of blisks typically do not provide the magnitude of the forcing, and hence c has to be determined by other means. The following section uses a modified version of the damping identification method presented by Holland et al (2012b) to solve for the unknown forcing coefficient.…”

“…As two methods are required, the proposed approach herein is referred to as the hybrid damping identification method. Holland et al (2012b) derived the following result…”

“…Appendix A provides details on obtaining Equation (10). As also discussed by Holland et al (2012b), one way to extract k is to use the frequency transfer function for the speakers or magnets used for excitation, and assume that the force applied to the structure in a given frequency band is proportional to the signal applied to the speakers or magnets in that frequency band. Solving Equation ( 6) for f, one obtains…”

“…The new method identifies first modal damping values relative to each other, which makes these values known (only) up to a multiplicative constant. Then, to determine that constant the novel approach makes use of the method of Holland et al (2012b). Distinct from other identification methods based on frequency response functions, the proposed approach introduces a reduced frequency response matrix which can be constructed without requiring that all degrees of freedom (DOFs) of the system model be both forced and measured.…”