RESEARCH TOPICS

Vibration and Power Flow in Complex Vehicle Structures

"Complex" structures are assemblies of coupled component structures. Analyzing vibration transmission between connected structures in the mid-frequency range is often a daunting prospect. Since the smaller wavelengths lead to greater model discretization, the size and computational cost of a full structure model (e.g., a Finite Element model) can become prohibitive. Also, as the wavelengths approach the scale of the structural variations, uncertainties (tolerances, defects, etc.) can significantly affect the dynamics of the structure. Starting at what may be called the mid-frequency range, deterministic models fail to predict the response of a representative structure with uncertainties. Therefore, in the mid-frequency range, a statistical analysis of vibration transmission may be more appropriate. However, simplifying (high-frequency) assumptions used in current statistical energy methods tend to significantly reduce their accuracy in the mid-frequency range. Therefore, new methods are needed to predict and simulate the vibratory response of ground vehicles (in particular, the transmission of vibration energy through vehicle components) and to assess the attendant durability. The objective of this research is to develop computationally efficient models of low- to mid-frequency vibration and power flow in complex structures that can be applied systematically and account for parameter uncertainties.

Vibration of Mistuned Bladed Disks

Bladed disk assemblies found in jet engines feature random variations in the blade geometries. This variation, called mistuning, is caused by manufacturing and material tolerances, as well as operational wear. Mistuning can lead to forced response amplitudes which are much larger than would be predicted by the analysis of a tuned assembly with identical blades. There is therefore a need to predict the effects of mistuning in the design process. Typically, a finite element model is made for a bladed disk assembly of interest. However, because mistuning is a random phenomenon, thousands of assemblies must be simulated in order to adequately capture the response statistics of mistuned systems. The attendant numerical costs make it infeasible to use traditional finite element analysis to predict the statistics of the mistuned forced response of bladed disks. Currently, a variety of reduced-order modeling methods are being developed for bladed disks. The resulting reduced-order models have very low number of degrees of freedom compared to the parent finite element models, yet mistuning effects are well captured in the order reduction process. In addition, advanced vibration analysis and design tools are being developed for bladed disks.