The Power of Root Cause Pre-failure Analys
Abnormal wear is not like a bad rash, which tends to go away on its own in time. Instead, it’s more like early-stage cancer, which requires intervention and treatment. Oil analysis has exceptional abilities to detect abnormal conditions, both root cause (like dirty oil) and predictive (active failure in progress).
Root cause failure analysis is post-mortem. It starts with failure and works backward in search of one or more root causes. The knowledge gained reveals a plan of needed change that will prevent or delay the recurrence of similar failures. Failure is indeed a strategic teacher of better ways to design, manufacture and maintain machines.
The whole purpose of machine condition monitoring, like oil analysis, is to enable organizations to foretell the future. It produces data that points to the existing problems and the seriousness of these problems. Action is required to confirm a problem’s existence, determine and verify the root cause, and finally to remedy the problem. Sadly, this is where most oil analysis programs are delinquent. The fault lies equally with the laboratory and the end user.
The Lost Art of TroubleshootingOil analysis can’t fix a failed machine. That’s what mechanics do. It can provide pre-failure alerts, both cautionary and critical. If a root cause is detected, such as the wrong oil, dirty oil or wet oil, these conditions can be remediated quickly. This is proactive maintenance at its best.
Approximately 10 to 20 percent of samples analyzed by oil analysis laboratories have one or more non-conforming conditions. The vast majority of these are predictive in nature, like abnormal wear. However, the origin of the impending failure (e.g., a particular bearing) and the root cause(s) typically remain uncertain to both the laboratory analyst and the end user. Routine oil analysis for predictive maintenance is an effective screen for abnormal conditions but ineffective at problem troubleshooting alone.
Presently, oil analysis laboratories are barely more than data generators. The evaluators used by these labs have limited time to spend troubleshooting and diagnosing individual samples with reportable conditions. Most evaluate several hundred data sets each day. They also rarely have a background in machinery lubrication, tribology, failure modes or machine design, and are generally unfamiliar with machine operating conditions and the exposures these machines and their lubricants face.
Be alert to the false promise or expectation that the laboratory can be your troubleshooter. Troubleshooting requires additional steps and persistent, timely action by the end user or designated troubleshooter. Unlike a bad rash, these problems don’t go away on their own. The offending condition must be identified, contained and surgically removed. The oil analysis report with the red alert only initiates the process. Be a pre-failure investigator and finish the process before the machine’s service life finishes first.
Critical machines that are in critical alarm need quick and effective troubleshooting by a qualified investigator. The following are examples of common conditions often found on oil analysis reports that require unrelenting investigative analysis:
- High/low viscosity
- High wear metals
- High particle count
- Coolant leak
- Fuel dilution
- Rapid oil aging
- Sudden high acid number (AN)
- Demulsibility issues
- Air-handling issues
- Dark oil issues
- Hot running conditions
Build the Case FileTroubleshooting requires discovery of anything and everything that might reveal the origin, root cause and severity of the problem. In the medical world, this investigative process is performed by pathologists. Forensic investigations go through very similar steps. Start by building the case file and leave no stones unturned. In addition to routine lab data and historical trends, seek the following:
- Companion condition monitoring data (vibration, infrared thermography, temperature history, pressure history, etc.)
- External inspections
- Internal inspections including a borescope
- Abnormal operating data (loads, speeds, etc.)
- Lubricant product data
- Service history (recent oil/filter changes, repairs, teardowns, etc.)
- Operator interviews/observations
- Preventive maintenance (PM) inspection history
- Past reliability history
More testing of the in-service lubricant could be required to dig deeper into the core of the problem. You may want help in selecting the tests that best answer your troubleshooting questions. Also, you might need a specialized laboratory that is equipped for complex analytical and tribological studies. For instance, one or more of the following tests may be necessary:
- Sediment analysis
- Used filter analysis
- Scanning electron microscopy (SEM)/energy-dispersive X-ray spectroscopy (EDS) of debris fields
- Inductively coupled plasma (ICP) acid digestion testing of large solids
- Composition particle counting
- Particle characterization
- Grease analysis
- Flender air-handling test
- Gas chromatography
- Metallography (machine surface morphology)
- Fourier transform infrared (FTIR) spectrum analysis
- Film strength analysis
- Organic solids analysis
Most importantly, don’t fail to take action. The problem won’t go away on its own. Take full control of the investigation and understand the urgency. Get the help you need. Use the knowledge gained from pre-failure as your teacher.
|88%||of oil analysis programs employ particle counting as an analytical method, according to a recent survey at MachineryLubrication.com|
About the Author
|Get 60 Questions to Ask Your Lubricant Supplier absolutely FREE when you sign up for our free Lube-Tips™ newsletter.|