Dissolved Gas Analysis-Gas Chromatography - Part 2
Now that we have extracted the sample of gas from the oil it must be injected into the chromatograph for analysis and identification of the gases. Gas chromatography machines use tubes packed with special material to separate the gases. These tubes are called columns. Columns come in varying lengths packed with varying materials depending on exactly what you’re looking for. You can’t just inject your gases into any old chromatograph and expect it to tell you what’s in the sample. The column is in the “oven “section of the machine and it’s heated to a specified temperature. This allows the gases to expand and stratify within the tube.
After a period of time long enough for the gases to separate and stratify they are “pushed” through the tube with an inert gas (usually argon) and then burned at the end of the line.
The chromatograph prints out a line graph corresponding to temperature and time. It looks sort of like a pencil line drawing of a series of mountains. Looking at time on the bottom horizontal axis tells you what gas you’re looking at (remember they stratified or layered and came out in a certain order). When you calculate the area under the peak you can evaluate the percentage of that particular gas. Now all that remains is to translate all of that data into parts per million of dissolved gas in the sample using the total amount of oil used along with the amount of gas extracted. That part is simple math.
All of the above is chemistry and can be performed by any qualified technician. The real art is interpreting the results. This requires usage of several different methods of evaluation along with a good field knowledge of the equipment and other reasoning skills.
I like to use Rogers Ratio, Sacramento State, Dornenberg Ratios, Key Gas and any other method available to come up with then verify my conclusions. No one method works best or all of the time. Conclusions need to be determined by gathering as much onsite physical data and equipment history as possible. In many cases when a problem is discovered it is only a snapshot. Another snapshot (test) should be performed at a later date to verify the problem and the extent of it. That could be as soon as a week or as long as six months depending on the problem.
That’s all for today.
Please write or comment on any subjects that may interest you for further postings.
Thank you,
Bob
Today’s quote is from Billy Connoly
