Transformer oil and hydraulic oil normally arrive with a test report attached. The report states figures such as the water content, the breakdown voltage and the cleanliness, and it looks complete. But once the oil reaches the site, the numbers on the report often do not match what is measured there. The reason is usually not that the oil has deteriorated, but that the report itself does not state three things. Filling in those three gaps removes a good deal of the argument at the point of acceptance.
The first gap: the condition the test was carried out in
The same oil sample can give different water content and gas content readings depending on whether it has just been discharged into a tank, has been in transit, has been stored in a drum for two months, or has already run through a filtration unit. If the report gives the values but not the sampling time and the state of the oil at the moment of sampling, there is no way to tell which batch and which moment the data belong to. Water content is measured to GB/T 7600 and gas content to GB/T 423, and both figures are sensitive to the sampling conditions: whether headspace was left in the container and whether the sample was shaken will both change the result.
The second gap: which set of standards was applied
Whether the breakdown voltage is written as 65 kV or 70 kV, and the water content as 3 ppm or 5 ppm, means nothing without the standard, because those limits are obtained under a specific test method and specific electrode dimensions. GB/T 507 specifies the method for determining the breakdown voltage, GB/T 7600 the water content, GB/T 423 the gas content, and cleanliness is graded to NAS 1638. If the report carries only the numbers and not the standard numbers, both sides read them by their own habit. Writing NAS 1638 class 5 is not the same as writing class 5 alone, because cleanliness also has another coding system in ISO 4406, and the grading of the two systems cannot be copied across directly.
The third gap: whether the tested oil and the delivered oil are the same batch
This one is the easiest to overlook. The report may be a factory test carried out on one drum, while several other drums arrive; in between there has been transfer, repacking or standing. If the number of drums, the batch number and the filling date are not on the report, there is nothing to check them against on arrival. The reliable procedure is to draw samples from the drums by batch number after arrival and repeat the test, rather than letting one factory report stand for the whole delivery.
| What the report says | What is missing | How to close the gap on arrival |
|---|---|---|
| Only the water content and the breakdown voltage values are given | The standard numbers and the test conditions behind them | Ask for a statement based on GB/T 7600 and GB/T 507 |
| Test values are given, but no sampling time and no state at sampling | The moment and the conditions the data refer to | Add a sampling record stating the sampling point and the state of the machine at sampling |
| Cleanliness is given as a class number with no standard system named | The grading system | Confirm whether it is NAS 1638 or ISO 4406, the two cannot be mixed |
| No batch number and no number of drums on the report | The link between the tested oil and the delivered oil | Draw samples by batch number on arrival, retest them and keep a retained sample |
| All the figures on the report pass, but there is no description of the sampling method | The sampling method and the container requirements | Take new samples as required, with the container and the sealing method following the standard |
How to read the report
- Read the standard number first and the numbers second: a value without a standard can only be taken as a reference, not as a basis for acceptance.
- Check the batch numbers and the number of drums on the report against what actually arrived, and sample the batches that do not match separately.
- Fix the sampling point when retesting on site: the top of the tank, the middle of the tank and the oil outlet can give different results, and without a fixed position there is nothing to compare.
- The retained sample matters more than the result: when two reports do not agree, the sample is the only thing that can state the facts.
- Read the delivery report and the report on the treated oil side by side, because the change between the two sets of data is what the machine actually did in that run.
What a test report can prove depends on the part it does not state. The fewer the gaps, the less there is to argue about at acceptance.
