Inspection of the oil pan found a crack along the bottom plate of the oil pan, adjacent to the weld. The pan was sectioned and the crack was opened up to examine the fracture in detail. Towards the middle of the crack, fatigue from multiple origins propagated primarily from the bottom (painted) side of the bottom plate. The ends of the crack showed fatigue propagation along both sides of the bottom plate, towards the center of the plate thickness. Microscope inspection of the weld near the crack and away from the crack showed good fusion, acceptable fillet leg and throat size, and no defects. Fatigue propagated from multiple origins along both sides of the bottom plate, towards the center of the plate. Fatigue progressed primarily from the bottom of the lower plate. This could indicate higher stress along the bottom side of the bottom plate.
Having reviewed the parts, and the failure mode identified, CAT contracted 6DOF to analyze the vibration signature of the failed component on site, and propose a plan to make the oil pan more resistant to the stresses induced from the vibration signatures identified.
Unfortunately, access limitations did not allow for placing any instrumentation in the oil pan failure region. The gap between the bottom of the oil pan and the base structure did not provide sufficient space to install any instrumentation.
Standard vibration measurements were made surveying the entire engine, compressor, base system of two CAT 3612’s and one CAT 3616 at the location of the first oil pan crack. This allowed for a comparison of units to see if this was a system or unit issue.
In general, the measurements obtained on three different units indicated that the vibration was not particularly high. The primary vibration was at the 1st and 4th orders (multiples of engine-compressor speed). Vibration measurements will typically either be referenced to another instrument or un-referenced. In this situation both types of measurements were taken but the difference between each are important when troubleshooting a vibration related issue.
Phase represents half of the vibration data available to you and can provide insight into the cause or possible solutions for reducing it. In this case, two accelerometers are used to get phase, one as a reference and the other as the measurement point. The phase angle is the position of the peak relative to the reference point. The measuring accelerometer is “roved” along discrete points to create an operating deflection shape (ODS) model. The speed of the unit is swept, from 900 to 1000 RPM to create an illustration of the motion pattern at a particular operating speed and engine-compressor order. The numbered points identify the measurement locations on the package. This allows you to create an animated model to see the direction (phase) of the vibration (motion).
The ODS were computed from vibration data acquired during speed sweeps. The finite element model was used to determine if this motion could cause a stress pattern consistent with the cracking pattern. Although, some system operating deformation shapes looked like they might be related to the pan cracking, no conclusive “smoking gun” was identified.
The key result of the first round of testing was that there was not a system wide issue that could cause the oil pan to crack. Rather a more detailed look was required at the local region. To do this, the finite element model (FEM) was reviewed to determine the mode shapes that could cause high dynamic stresses in the orders that were measured during the first round of testing. The FEM would be reviewed to determine strain and measurement locations directly on the oil pan.