Experimental investigation of mistuned bladed disks system vibration
Résumé
Bladed disks are critical structural components in jet engines and other turbomachinery. The nominal design for a bladed disk is typically assumed to have identical blades. However, there are always small, random variations in the blade properties due to manufacturing tolerances, material defects, and operational wear. These blade-to-blade discrepancies, called mistuning, can have a dramatic effect on bladed disk vibration. In particular, mistuning can cause localization of the response in a small region of the bladed disk, leading to higher blade stress and high-cycle fatigue concerns. While comprehensive analytical and computational studies of mistuning have been performed, relatively few experimental investigations have been conducted. The primary objective of this research is to experimentally investigate the fundamental structural dynamics of mistuned bladed disks, and to achieve a physical understanding of mistuning effects by accounting for the influence of important phenomena that have been largely neglected in previous mistuning models and system identification algorithms. First, a systematic experimental approach is presented to validate a new mistuning identification and model updating algorithm for single-piece bladed disks, or blisks. It is shown that only a few system response measurements taken at resonant frequencies are required to identify the blade stiffness mistuning parameters and the model updating parameters referred to as cyclic modeling error. By incorporating a model updating procedure, the accuracy of the mistuning identification results are significantly improved. Second, an alternative approach for vibration testing of many mistuning patterns is proposed and validated. In particular, varying the external forcing function provided to the blades is used to mimic the influence of structural blade property mistuning on the vibration response. Since it is much easier and more efficient to vary the external excitation than to physically alter the blades, this work opens the possibility of running an experimental analogue of a Monte Carlo simulation. Finally, the mistuning identification method is extended to also identify the forcing amplitude and phase applied to each blade. This approach shows promise as a powerful tool for accelerating calibration procedures, as well as for improving the accuracy and capability of experimental methods for bladed disks.