When attempting to schedule maintenance actions based upon oil analysis data, simple statistics can be a powerful tool to simplify data, identify relationships between oil analysis parameters and increase confidence in conclusions. Statistical techniques like correlation analysis can really help ensure that we are making the right decision. They can also help focus our efforts to uncover the root cause of the abnormal condition.
Review of the oil analysis data from nine identical hydraulic machines performing the same function in the same environment reveals substantial variation in zinc levels as a function of the time the oil is in service. Further investigation leads us to conclude that total acid numbers (TAN) also decline as a function of time. Upon calculating the correlation, we see that zinc and TAN values are highly correlated (see Figure 1).
We know that the zinc dialkyldithiophosphate (ZDDP) used in most anti-wear oils reacts with the potassium hydroxide (KOH) reagent used to measure TAN, elevating the numbers when the oil is new. The TAN number declines as the additive is depleted. Once the ZDDP anti-wear/antioxidant additive is completely depleted, it leaves the base stock with reduced protection from oxidation, and TAN numbers will begin to rise from their minimum point as the base stock degrades. Also, once the ZDDP is depleted, the machine is subject to increased wear due to lost anti-wear protection from the fluid. Zinc and TAN tend to correlate well in most oils equipped with a ZDDP anti-wear additive. It is important to quantify this correlation with test data specific to an application.
We notice from the analysis that one of the machines is running with very low zinc levels and low TAN numbers. Because both zinc and the TAN numbers have depleted, and knowing that the correlation between these two parameters is strong in this application, we have high confidence that our ZDDP additive is depleted, perhaps to the point of exhaustion. This is a situation that warrants maintenance action. It may well be that the oil has simply reached the end of its life. Alternatively, abnormal stress might have expedited degradation. Additional oil analysis and inspection of the machine should identify if the degradation is normal or abnormal. If abnormal, the process should reveal the specific root cause of the problem. Once the root cause is identified, a maintenance action can be scheduled to correct the situation. If the rate of degradation is deemed normal, we simply change or reconstitute the oil without further investigation.
By understanding how various oil parameters correlate, we can investigate abnormal symptoms and make decisions with a strong sense of confidence that we are addressing real maintenance problems, not just chasing false alarms.