The laboratory only measures what it receives
A perfect analysis of a poorly taken sample gives an exact result for something that is not the oil in the equipment. That is why what looks like a deviation in the transformer is, on many occasions, actually a sampling artifact.
How each error distorts each measurement
| Error | Frequent effect |
|---|---|
| Exposure to air or a poorly closed container | Measured water rises and dissolved gases are lost |
| Not purging the sampling valve or line | Stagnant oil, with dirt or moisture from the sampling point, instead of the oil in the equipment |
| Bubbles in the syringe or cylinder | Loss or gain of gases; unreliable dissolved gas results |
| Dirty container or hose | Particles, fibers, and contaminants the oil did not have |
| Sample taken with cold oil or without recording the temperature | Misinterpretation of water, which depends on temperature |
| Prolonged exposure to light | Changes in some oil properties |
| Incomplete identification | Confusion between units and loss of the historical trend |
What an engineer sees in practice
Results that do not agree with the previous trend and have no operational cause: a jump in water with no change in load, gases that fall without intervention, or two samples from the same day with very different results. This is the signal to suspect the sampling before the equipment.
Good practices in principle
- Use clean, dry containers suited to each analysis, and keep them closed until sampling.
- Purge the sampling point and rinse with the oil itself when the method calls for it.
- Fill so that no air enters and no bubbles form, using a closed circuit for gas analysis.
- Record the date, oil temperature, equipment, sampling point, and load conditions.
- Ship the sample protected from light and in the shortest possible time.
Sampling on energized equipment must be carried out by qualified personnel, following the applicable regulations.