Despite everyone’s best efforts component failures can and do happen. While the initial concern is typically with restoring production it is important to identify the root cause of failure and prevent re-occurrence.
A failure investigation may involve initial data gathering, visual assessment of the failed unit, sampling of lubricants, analysis of wear debris and testing of the failed component. Some failure modes can be detected by a range of condition monitoring techniques, others may arrive unannounced. Failures related to wear or lubricant contamination may have indications in the oil sampling history (hopefully the potential for failure would have been identified and highlighted). Some failure modes, however, can be sudden and catastrophic without any advanced warning being visible (e.g. due to overloading). In either case there may be evidence of the failure event itself both in the lubricant and in the oil filters.
When conducting a failure investigation it is important to record all relevant information about the event, including its chronology and operational parameters. When collecting lubricant samples do not rely on the usual sample points alone – consider sampling from additional locations which may help identify origin of potential contamination, the extent of wear and other useful information. Retain oil filters – the entrapped debris can be retrieved and analysed to shed further light on the failure mode.
When removing failed components for further testing mark them to indicate orientation and fit, take photos in situ and during the disassembly process. Make sure that components are not damaged further during disassembly or, where this is unavoidable, record and identify disassembly damage to differentiate it from the original failure. If engaging assistance of experts do not clean the components until they had a chance to view them or at the very least take a sufficient quantity of high quality photographs. It is often difficult to draw conclusions when a lot of the evidence has been washed away.