Crack Detection in an Embedded Spindle using Broadband Modal Excitation

This paper demonstrates a methodology for detecting cracks in a metal spindle, which is housed within a military vehicle wheel end assembly. A finite element model is used to estimate the undamped natural frequencies of the raw spindle. Additional mass and stiffness is then added to a simplified model of the spindle to simulate sources of variability within the assembly. Spindles are tested outside the wheel end assembly to compare the responses of undamaged and cracked spindles. A hypothesis testing approach is developed to detect the presence of a crack based on damage indices involving the summation and multiplication of spectral energies. Experiments on wheel assemblies mounted on a vehicle with the vehicle lifted off the ground are performed to demonstrate that the wheel assemblies can be nondestructively evaluated to identify cracks of varying depths despite sources of variability due to the presence of the brake drum, lack of repeatability in the sensor attachment, and other issues.