The water content is the indicator that is measured and measured again most often of all the acceptance parameters, and also the one that most easily produces a disagreement. The reason is not complicated: there is more than one way of measuring water, they measure different things, they report different numbers and they carry different degrees of credibility. Putting the results of two methods side by side and comparing the numbers does not usually support the conclusion.
Karl Fischer: the strongest quantitative capability and the heaviest cost
The Karl Fischer method is based on the quantitative reaction between water and iodine, and it converts the consumption into the mass fraction of the water by titration or coulometry. Its value lies in its capability to quantify at the trace level: it measures a water content in the ppm range in a stable way. This is also the method route on which a test basis such as GB/T 7600 relies.
The cost shows in several respects: a dedicated instrument and matching reagents are needed, the reagents have a shelf life and have to be stored as specified; the operation needs trained staff, and the judgement of the titration endpoint or of the coulometric stop affects the result directly; the sample volume is usually very small, and every step of sampling, transfer and sealing can introduce an error; a single test takes no short time, which is especially noticeable in batch testing.
Indicative methods: they give a clue, not a number
On site there is also the practice of using an indicative reagent or a microbial activity criterion to assess the moisture condition of the oil. The principle is to determine whether a water condition that can be used by microorganisms is present in the oil, and the reagent gives the indication as soon as it changes colour. The real use of such a method is quick screening and watching the trend.
It has to be said clearly that this gives a qualitative conclusion, it cannot be used to read a ppm value and it cannot serve as acceptance data. Its value lies in answering the question "is the moisture condition of this tank abnormal" at very low cost, which gives the basis for deciding whether to send the sample for test.
On-site quick methods: fast, at the price of scatter
The on-site quick methods include visual inspection, the heating method with the observation of bubbles, the reading of the condensate water and conductivity testing. Their common advantages are that they are portable, quick and almost cost-free, which suits them for following the trend of the same batch of oil continuously on site.
The limitations are just as clear: the result depends on the subjective judgement of the observer, the repeatability is poor, the influence of the temperature and of the moment of sampling is large, and they do not reach down to the ppm level. These methods are therefore suited to judging a trend, not to making the final decision.
Why the results cannot be compared directly
Three reasons are the most common. First, the object of the measurement differs: a test method may give the total water, the water that can be vaporised or the water that can be used, and there is no fixed ratio between the three. Second, the sampling differs: the sampling point, the sample volume, the sealing of the container and the transfer all change the result, and a water content below the ppm level is in particular very sensitive to the adsorption and the release on the wall of the container. Third, the test conditions differ: the temperature, the atmosphere and the time the sample has been standing all bring a deviation.
So a sentence such as "we measured 5 ppm and the other side measured 2 ppm" means nothing in itself. It is necessary first to state which method each side used, how the samples were taken and under what conditions the measurements were made.
How the acceptance clause should be written
At least five things should be written into the document: which test method is used, with the reference given; the sampling point, the sample volume and the requirements on the container; whether the sample has to be taken on site and how long it may stand; the acceptance threshold and the way of retesting after an exceedance; the arbitration procedure in case of a dispute. A vague wording leaves all the risk to the day of the acceptance.
| Method | Principle | Output | Accuracy | Time and cost | Application |
|---|---|---|---|---|---|
| Karl Fischer | The quantitative reaction between iodine and water | A quantitative value | Quantification at the trace level | Takes longer, needs reagents and training | Laboratory testing and formal acceptance |
| Indicative method | An indicator or an activity criterion | A qualitative conclusion | Can only give a trend | Quick, low cost | On-site screening, deciding whether to send the sample for test |
| On-site quick method | Visual, heating, condensate, conductivity | A qualitative or rough reading | Wide scatter | Quick, almost no consumables | Continuous tracking of the treatment process |
- The acceptance is based on the quantitative method, the on-site methods are used to follow the process.
- Before comparing data from different sources, check the method, the sampling procedure and the standing time.
- The sampling container and its sealing decide directly how credible the measurement of the trace water is.
- For the same machine and the same oil, establish your own baseline data, it makes an abnormal trend easier to see.
What the water figure is worth depends on how it was measured, so align the methods before you compare the numbers.