Date of Award

Fall 2013

Degree Type


Degree Name

Doctor of Philosophy (PhD)


Mechanical Engineering

First Advisor

Douglas E. Adams

Second Advisor

Jeffrey F. Rhoads

Committee Chair

Douglas E. Adams

Committee Member 1

Jeffrey F. Rhoads

Committee Member 2

R. Byron Pipes

Committee Member 3

Peter H. Meckl


There is not a complete understanding of how damage mechanisms in composite materials react to nondestructive testing inputs. A deeper understanding of composite damage mechanisms and their responses to external vibratory excitation is sought using nonlinear modeling of damping and stiffness characteristics related to the damage. Different types of damping and stiffness single degree-of-freedom models are presented with analysis of their behavior at specific harmonics of the primary resonance of the system. After an understanding of the approximate behaviors of the models is obtained, experimental tests on a damaged specimen are conducted. An aluminum honeycomb sandwich panel is damaged by applying a heat source to the top face sheet. The expansion of the heated area on the face sheet creates a disbond of the face sheet and the honeycomb core. This damaged panel is tested by exciting the damaged area with a shaker at known frequencies and amplitudes. The resulting responses of the panel are measured and compared with the nonlinear predictive models through a direct comparison of response amplitudes to the analytical solutions and by examination of restoring force curves. The disbonded aluminum honeycomb sandwich panel exhibited behavior similar to a pure quadratic stiffness as well as a smaller influence from a cubic stiffness. The quadratic stiffness is the result of the facesheet experiencing two distinct stiffness regimes, the first as the facesheet moves away from the core, and second as the facesheet presses into the core. The smaller cubic nonlinearity is thought to come from the additional stiffness imparted into the single degree-of-freedom system by the epoxy fillets that hold the facesheet to the core. As the face sheet vibrates, the epoxy fillets contribute a small additional stiffening as the displacement grows larger. It is also confirmed that the displacement of the damaged area is able to be modeled using single degree-of-freedom models. This enabled the use of single degree-of-freedom equations of motion, which simplified the nonlinear analysis. Two further observations are made with regard to potential damage detection applications. First, that lower frequency excitation of panels may be able to excite the damaged areas more easily for nonlinear measurements than higher frequency excitation. This is supported by the nonlinear analysis showing that additional response peaks are more readily obtained through superharmonic excitation than subharmonic excitation. Second, that smaller damage sizes have higher quadratic and cubic stiffness coefficients, which produced relatively larger responses at the primary resonance when excited at the superharmonic frequencies than the larger damage sizes. These relatively larger responses may be a key in decreasing the size of damage detectable using vibratory excitation.