The problem with a calendar interval is not that it is inaccurate but that it is wrong at both ends. Changing early discards oil that could still be used, wasting both the cost of the oil and the downtime. Changing late means that the deterioration of the indicators is only noticed when the filter medium, the seals and even the transformer windings may already have been affected together. The reason an oil change is so difficult to convert into a condition-based decision is that the site often cannot obtain the data on which to base a judgement.
Which end the calendar interval gets wrong
A calendar interval holds only on the premise of the same machine, the same oil and the same duty. On a real site the loading rate, the number of starts and stops and the condition of the seals all change year by year, and the batch of oil changes with the machine. A register interval of one change every three years is clearly too long for a machine that was later switched to frequent starts and stops, and clearly too short for another machine that runs lightly loaded at low temperature.
The question that really has to be answered is not how long the oil has been in service, but how far the present oil has degraded compared with new oil. That requires a set of indicators that can be measured repeatedly and compared across machines, not a date.
The quantities that condition monitoring watches
The oil indicators commonly used in industrial service each have their own part to play: moisture shows whether water has come in from outside, acid value shows the degree of oxidative deterioration, particle count shows the level of solid contamination, the dielectric characteristics show the ability of the oil as a whole to withstand an electric field, and the gas content shows the state of the dissolved gas. They are not independent of one another; a sustained rise in any one of them usually corresponds to a developing problem at some point inside the equipment.
| Monitored item | Common test basis | What a rise or fall in the reading indicates |
|---|---|---|
| Water content | GB/T 7600 | A rise usually points to a failed seal or a damp breather |
| Gas content | GB/T 423 | A rise usually points to partial discharge or a tendency to overheat |
| Breakdown voltage | GB/T 507 | A fall shows that the overall withstand capability of the oil is deteriorating |
| Cleanliness | NAS 1638 | A worse class shows that solid particles have entered the oil circuit |
| Acid value | Method selected according to the type of oil | A rise shows that oxidation products are accumulating |
From data to action: how the purification process connects
The monitoring data does not solve the problem by itself; its role is to decide whether the machine is to be touched, which one, and how far. The smoothest practice we have seen takes three steps. First, use the test data to establish the nature of the problem: is it water ingress, is it particles, or is the oil itself ageing. Then choose the means: water ingress calls for dehydration and degassing, particles call for staged filtration and flushing, and exhausted additives call for an oil change rather than purification. Only then are the parameters set, working back from the quantity of oil in hand and the target indicators to the processing time and the temperature.
- Where the water content is high and the breakdown voltage falls at the same time, schedule vacuum degassing first rather than changing the oil directly.
- Where the cleanliness class has got worse but the moisture is normal, concentrate on the source of the particles in the sealing and assembly stages and flush first.
- A continuously rising acid value means that the oil has degraded to a point where purification can no longer recover it, and an oil change should be considered instead of further treatment.
- A rising gas content is often not a problem of the oil but a problem inside the equipment, and treating the oil cannot replace the investigation of the fault.
What condition monitoring cannot replace
It should be said clearly that condition monitoring answers when to act, not whether the action was right. After a purification action on site, the same set of indicators still has to be retested with the same test method to confirm that the round of treatment was effective. If only the moisture is measured before and after the purification, problems other than moisture will never come to light.
In addition, the consistency of the sampling directly determines how credible the monitoring is. Only samples taken at the same measuring point, by the same sampling method and, as far as possible, at similar temperature and load conditions are comparable with one another. This point affects the result more than the choice of laboratory equipment.
Changing an oil change from the calendar to a condition-based decision requires a chain of indicators on site that can be measured repeatedly, compared and used to trigger an action, and not merely one more sensor fitted.