Vibration Induced Fatigue
Vibration induced fatigue, and fatigue failure, are dependent on a number of parameters that should all be considered at the start of an investigation into a fatigue issue. As vibration consultants, with expertise in structural dynamics, we are often asked to both asses fatigue risks in design, as well as analyze fatigue failure when it occurs. Stress level, number of cycles, geometry, and temperature should be considered along with the structural dynamics of the system. Efforts to lower stress levels without the big picture often fall short. For instance, due to the shape of the SN curve, for most metals, reducing the number of cycles will increase fatigue life. Conversely, if we reduce the stress levels we may be comfortable greatly increasing the number of cycles. A load spreading modification to the structure may appear to reduce point stress levels given a constant applied load. However, adding the load spreading plate can change the structural dynamics and result in greatly increased loads at another location.
Cracking of Generator Housing Due to Stress Induced Fatigue
Cracking of Generator Frame Due to Resonant Amplification And Resulting High Stress
For example: An Engine/Generation Set system was showing cracking of large frame member and also cracking of the generator housing. We performed Modal Analysis Testing to show that the system had a resonance at the operating speed and a mode shape that involved the generator and frame deforming in a way that would cause large strain at the failure locations. With this analysis we understood why these areas had fatigue cracks.
An FEA Model was built and tuned to match the existing dynamics of the engine/generator/frame system. A "fair" FEA model from which design modifications can be extrapolated requires significant expertise. Detailed work on boundary conditions, and connections, as well as the structural dynamics experience to quickly know what to pay attention. We experimented with modifications, and iterated upon them in the FEA model, to optimize the resonant response and mode shapes of the systems so that the fatigue issue would be resolved. The structural modifications were then made on the real system. The system was tested and found to match closely with the FEA predictions. No failures have since been reported and none are expected.
Comparison of our FEA model that was created using our modal analysis of the physical system to fine tune the model to match the important deformation shapes and resonances of the system.
The SN Curve, Useful in Analysis of Fatigue Issues
Recuperator Heating Tube Failure - A wood chip particle board manufacturer was having repeated failures of the long tubes that were used as a cross-flow heat exchanger to recapture heat that would be lost with the combustion waste. Vibration and thermal expansion were suspect.
Inspection of a Large Gas Duct During Investigation Into Heating Tube Failures At A Wood Products Plant
Cracked and Broken Heating Tubes During Investigation of Heating Tube Failures
We used Strain Gauge Testing to estimate the stress levels at the top of the tubes near the cracking locations while the system was cycled up and down in temperature. We also measured the Vibration Levels on the tubes at multiple locations. From the dynamic and static strain levels were able to show that the problem was due to near static stress levels. The vibration levels were high (and obviously concerning) but were not high enough to explain the cracks in the tubes. The high static stress, however, was not suppose to be present by design. We suggested that the tubes must somehow be constrained against free expansion. A close inspection showed that this was indeed possible as deformations of the tube guide and tubes could cause enough stiction to constrain the tubes and explain previous failures. A solution path was mapped out.
Boiler Refractory Cone failures - We were called to trouble shoot repeated refractory cone failures on a two boilers used in ethanol plants. For this project we used Operating Deflection Shape (ODS) testing, Dynamic Testing, and failure analysis. Our testing showed that the boiler burner support structure was vibrating at higher levels than necessary due the design of a couple of support legs that were not well grounded in the floor slab. The ODS testing showed that the current design of these support legs resulted in rocking motion at the structural connection of the refractory cone. Analysis of the failed refractory cones provided a hypothesis for a mode of failure. This hypothesis involved the rocking vibration of the burner due to the poor attachment of the supporting legs, which, in turn, caused the top braces to dig though the cone material, given the moment preload associated with the overhung cone design.
Refractory Cone in a Boiler, Three Supporting Brackets on Top Gripping the Front Edge of the Cone
Failed Cone with Brackets Worn Through the Cone
A prototype with More Support On the Cone Face To Reduce Stress Concentrations
We also showed that the support of the cone could be redesigned to significantly reduce stress concentrations and reduce cone failures at the mounting locations. We made design suggestions to fix the boiler support legs to further reduce the inertial loads on the refractory cone associated with the rocking motions excited during boiler operation. - Thus far we have not heard of any more failures.