Oil analysis is an overwhelmingly popular nondestructive maintenance tool, and for good reason. It identifies numerous problems that are fixable in their early stages while occasionally stemming a catastrophic failure in heroic fashion. In fact, oil analysis (OA) is perceived to be so beneficial that, often, inadequate records are kept to verify the benefits in financial terms. Savings are just assumed, or obvious. This highly prevalent nonchalance has, unfortunately, led to a number of unsupported conclusions and decisions, for example:
-
Programs perceived as marginal are unjustifiably dropped because the benefits are not sufficiently obvious.
-
Weak commentary on the report is tolerated, exacerbating the issue raised in point No. 1.
-
Price rather than performance has nearly always been the principal determinant in selection of a laboratory, contributing to the issue in point No. 2.
When the cost of the test is more important than the quality and accuracy of the conclusions to be drawn from the test data, something is wrong. Why does one test oil samples and gather data? The data are neither the goal nor the end result, but rather a path toward understanding what's going on within the system: the component and its lube.
Seeking ExpertiseA qualified expert is needed to understand what's going on, based on the OA data under scrutiny, yet few OA users investigate this area when they hire a commercial lab to test their samples. Make no mistake; the quality of evaluations varies widely between labs and between lab locations, where multiple branches exist. It's no secret, it's common sense. All evaluators are not equal, and never will be, because each evaluator comes from a different background and set of experiences, with or without various mentors along the way. Once this is understood, an automated intelligent system makes a great deal of sense.
Prescient is an auto-evaluation software application, developed to handle the most important, yet least addressed aspect of oil analysis: the evaluation of data; specifically the comment, the maintenance advisory, the work order; addressing the reason oil analysis is performed in the first place.
There are compelling reasons why an automated evaluation program is an ideal approach in the oil analysis process:
-
Hierarchical complexity is readily accommodated. One can introduce intricate wear, lube degradation and contamination scenarios and patterns with complete confidence they will be logically and appropriately applied. This ability clearly allows one to focus and limit diagnostic aspects to a narrow platform; that is, one can be specific for specific situations where knowledge is especially deep.
-
Consistency is virtually guaranteed - all the knowledge resides under one roof, applied precisely.
In rating data - Solid statistical analysis, tempered by specific experience, results in data rating and flagging that is both appropriate and consistent, allowing conclusions to be focused.
In issuing commentary - Like patterns and data sets will result in like comments, at maximum depth. It is true that the commentary could occasionally be less than ideal where knowledge was not (yet) adequately corroborated, but this is readily corrected, and then it doesn't waver. -
Knowledge can grow exponentially. Any number of experts can inform the knowledge base so that it can grow in numerous directions simultaneously, achieving progressively more powerful capability quickly.
-
Continuous Improvement
Prescient has roots in artificial intelligence, and can begin to rate itself once enough feedback is received (see Feedback section below). As the software application gains knowledge, it refines it. -
Auditing
Maintaining a proper audit trail for maintenance activity is a vital, and often required task. This requirement is driven by safety and financial considerations, as well as deliverables. The software application provides such a trail by generating specific reports, showing conditions found and making available date- and time-stamped access records for log-on or feedback entries.
Feedback - the receipt, collation and correlation of machinery symptoms and repairs, based on reports - has been a bane to oil analysis. Many users simply don't want to bother, or it may be difficult to coordinate the actual findings when several people are involved in the maintenance activity. Some private laboratories with a high degree of control and accountability, do an excellent job of seeking, gathering and using feedback, but they are in the minority. Commercial labs don't fare any better. Most users fail to inform the evaluator(s) in a complete and consistent manner. The consequence, of course, is slow learning, and difficulty with program validation, as mentioned earlier. This is a critical aspect of OA that the software application addresses.
In recognition of the difficulty in acquiring proper feedback, a mechanism is built into the DataView screen mode, making it easy for the maintenance group receiving and using the report to inform the software application as to work performed or not performed. This is one of the learning tools within the software.
Buttons can be easily checked to indicate pertinent maintenance performed. Additional comments can be added to a text box to further clarify mechanical work performed, or as-found conditions.
Pump ReportThis report for a hydraulic vane pump shows a detailed, logically presented comment stream, organized by action, reasoning, then lube maintenance items. Notational comments offer further clarification as appropriate. Mechanical action that might be considered is printed in red text so that it stands out, but follows diagnostics comments.
Mechanical action is only advised pending results of diagnostic actions, thus a mechanic should first take noninvasive action to ascertain if a mechanical inspection is justifiably indicated. Further, this is the opportunity for the mechanic or operator to insert his own valuable observations on-site into the diagnostic and decision-making equation.
Oil analysis will continue to be a highly-valued maintenance tool, but it could use some help in reaching 21st-century standards when it comes to data evaluation and report commentary rendering. In this area, technology has heretofore been sparingly applied; however, a new approach is now available.
Prescient is a significant step in that direction.
