After an accident, a flood or a fire, the contaminated oil usually faces a decision: send it to the laboratory, or purify it. The order looks like a matter of procedure, but it decides the cost and the risk directly. Read the situation the wrong way and the waste is time and oil; get the order wrong and oil that does not qualify may find its way back into the system.
The first step is always safety and sampling
The first thing after an accident is not to go into the machine but to make sure it is safe: has the system been depressurised, have the live parts been isolated, is there still residual pressure or oil vapour in the tank. Until that is confirmed, neither sampling nor any other operation should begin.
After the safety check comes sampling. This step alone can ruin a test result: a container that is not sealed, a long storage time, or air moisture entering the sample while it is taken, and the figure the laboratory reports no longer describes the oil in the tank. The sampling point matters as well: the tank bottom, the middle of the tank and the outlet can all give different states.
The source of the contamination should be sorted out first. Rain water and ground water bring water into the oil, while the residues left by fire fighting, cleaning agents and metal oxides bring solid particles and chemical substances. The two kinds of contamination follow different paths: water is dealt with mainly by vacuum degassing and dehydration, solid contamination mainly by filtration, and a mixture whose composition is unknown has to go through both operations. Working out what has been added to the oil before deciding on the order of treatment is more reliable than getting the equipment running first and finding the answers afterwards.

| Test item | Basis of the judgement | Direction of the treatment when it is out of limit |
|---|---|---|
| Water content | ≤3 ppm (GB/T 7600) | Drying and degassing are needed, then a retest |
| Breakdown voltage | ≥65 kV (GB/T 507) | Retest to rule out bubbles and residue on the electrode, then decide on treatment |
| Gas content | ≤0.01% (GB/T 423) | Vacuum degassing is needed, then a retest |
| Cleanliness | NAS class 5 (NAS 1638) | Filtration is needed, then the particle count is checked again |
| Filtration accuracy | ≤1 μm | An indicator of the equipment side, read apart from the tests of the oil |
Let the result decide the next step
- The water content is clearly out of limit while the particle count is acceptable: vacuum degassing and dehydration first, then the water content and the breakdown voltage are tested again.
- The particle count is out of limit while the water content is normal: filter first, then check the cleanliness again.
- The breakdown voltage does not pass while the water content is acceptable: test again first and rule out the effect of bubbles and of residue on the electrode before drawing a conclusion.
- Several items fail at once: deal with the water content first, because the breakdown voltage is the more sensitive of the two.
Two situations in which the oil should not be released
The first is when the sampling itself was faulty: the container was not sealed, the sample was stored too long, or air was mixed into it while it was taken. No test result is enough to support a decision to release the oil in that case, and the right answer is to take a new sample.
The second is when the state inside the machine is unknown: if residual water, foam or sediment is left in the tank after the accident, going straight into the oil treatment carries the problem into the filtration stage and into the vacuum stage, and the cost and the risk of that are larger than emptying the tank and looking first.
There is no standard answer to how the oil after an accident is treated, but there is an order: safety first, then sampling, then the treatment decided by the result, and a retest to confirm. If any one of these four steps is left out, the conclusion behind it does not hold.
The retest is the step most often left out, and it is the only way to judge whether the treatment worked. Taking a second sample after the treatment and setting the new figures beside the ones from before shows whether that round of work actually got there. If the retest is still out of limit, the contamination has not been cleared, and the source has to be looked at again instead of running the same operation a second time.
Whether the oil after an accident is tested first or treated first is decided by the test result itself; with the order wrong, the whole batch has to be done again.
