Remember your elementary chemistry? Lubricants are nothing but a mixture of long-chain hydrocarbons that somebody dug out of the ground and some specialty chemicals. Everyone knows that you can’t destroy hydrocarbon molecules by grinding them between gears and mashing them between steel plates. Hydrocarbon molecules are immune to physical attack. So, what can go wrong with the durable molecules in your oil? What can justify spending money on oil analysis?
Nevertheless, someone convinced you that sending an oil sample to a lab for analysis is a smart thing to do. After some discussion, you dispatched your best technician to collect the sample. The technician used the best possible sampling techniques and properly labeled the bottle.
The shipping department very carefully packed the bottle to minimize the transit problems then dispatched the carton to your favorite lab. A week later, the lab sent you a brand new, clean sample bottle, another blank label, and a sheet of paper filled with numbers they claim represents your previous sample.
Now, I ask you, what are you going to do with that sheet of numbers? What does it all mean? How are you going to convert this presumably important information into some action that justifies both the lab cost and in-house cost charged against your department?
The big picture
Like any other fluid, oil has certain physical and chemical properties. It’s the specific combination of these properties that prompts people to use oil for lubrication. Unlike the physical gnashing of gears, oxidation is a chemical process that does break those durable molecules into smaller pieces to alter the lubricant’s properties. Therefore, oil analysis should address the intrinsic properties of the lubricant itself.
The sump and system from which your technician drew that sample constitutes a container. There are other fluids besides oil that can reside in that same container. Because these fluids are not so good at minimizing friction, they are considered to be contaminants. Oil analysis should address the fluid contaminants.
The only reason a machine requires a lubricant is that somewhere in its internal workings are some pieces of metal and other materials rubbing or banging against each. This mechanical action induces stress that exceeds the metal’s elastic limit and microscopic metal pieces break off. The oil carries these pieces away so more stress can break off more pieces. Therefore, oil analysis should address the bits and pieces of solid contaminants.
Oil analysis report forms present a lot of information covering the properties of the oil, the fluid contaminants mixed in the oil and the solids suspended in the oil.
The specific gravity is nothing more than the mass per unit volume referenced to a standard temperature. Oil contaminated by solvents, for example, has a specific gravity different from that of clean oil. It’s a nice number to know, it’s easy to measure, but it doesn’t really tell the whole story.
Viscosity is a measure of the oil’s relative resistance to flow. Its value depends on the temperature at which viscosity is measured, the viscometer used for the measurement and the technique of the lab tech. for best results, always measure the viscosity at the same temperature on the same viscometer operated by the same technician. Because there are so many variables that affect the trend data, the smartest thing to do is to buy your own viscometer and train someone to use it in-house.
Measuring viscosity at some constant room temperature yields the greatest resolution. At elevated temperatures, the difference between the viscosity of degraded oil and good oil is too close to differentiate. The correct viscosity is application dependent and has both a lower and upper limit. The specs on the lubricated machine should indicate the proper viscosity to be used. Find an oil that matches it.
Don’t count on your viscometer to give you readings shown on the label on the oil drum or on the oil spec sheet. Measure the viscosity to get a baseline value for the current lot of clean oil. Get another baseline reading each time you start drawing from a different load of oil.
Total acid number
This measurement, in milligrams of potassium hydroxide per gram of oil, is related to the level of oxidation present in the oil – more oxidation, higher acid number. Get a baseline reading on a sample of clean oil. One caution though – clean oil has a measurable acid number because the lab test responds to some of the antioxidants and antiwear additives as if they were acid contaminants. This is called the interference effect.
As the additives deplete over time, the reported acid number drops, goes through a minimum, then rises. It drops because the additives are being depleted. The curve bottoms out when they are gone. It rises again because there’s now genuine acidic material in your oil.
Take action just before hitting the bottom of the curve. Replace the oil completely or refortify it either with concentrated additives or with a partial replacement.
If you miss the bottom, don’t risk getting into trouble. Even if you drain the sump and replace the oil, acid in the residual oil rapidly depletes the additives in the new oil. The best action to take if you miss the bottom of the curve is to drain the sump and lines, then clean them well before adding new oil.
Total base number
Acid buildup in oil is a bad thing. Chemistry being what it is, bases react with acids to neutralize them. The total base number is an aggregate measure of the concentration of components in the oil that counteract acid buildup. It’s measured in equivalent milligrams of potassium hydroxide per gram of oil.
This test is generally performed only for oils taken from internal combustion engine crankcases. It’s the fuel’s sulfur content that forms sulfuric acid during the combustion process. In the past few years,the sulfur content of diesel fuel has been very low. This means, for example, that some over-the-road engines operate for 20,000 miles and reduce the total base number from an initial value of 8 down to a final 6. in any case, if you monitor this variable, then you should be trending using a baseline measurement on the clean oil.
There seems to be some controversy about changing oil on the basis of total base number. Some proponents advocate changing oil when the total base number depletes to 50% of the total base number of the virgin stock. Others advocate waiting until the total base number drops to somewhere around 2 mg. KOH/g. My advice – be conservative. Oil is cheap, machinery isn’t.
Fourier Transform Infrared Spectroscopy is a versatile way to determine contamination levels and additive depletion in oil. Infrared radiation interacts with materials on a molecular level, not the atomic level. Molecules in the sample absorb IR at specific wavelengths in the IR spectrum – more molecules of a substance, more absorption. It’s possible, therefore, to get a correlation with concentration in the oil sample.
The spectrometer measures the infrared light passing through the sample. The spectrograph is nothing more than a curve of the intensity of the transmitted IR as a function of wavelength. The Fast Fourier in the name refers to the algorithm that processes the signal for output as a curve.
This analytical method monitors the depleting concentration of antioxidants and antiwear components. It also measures increasing levels of contamination from moisture, oxidation products and dissolved metals. If the sample is from an internal combustion engine, it also measures ethylene glycol and soot levels.
Of course, you already have a base spectrograph for the virgin stock of oil. This brings up the issue of differential spectroscopy. This refinement mathematically deducts the spectrum of the base stock from that of the oil under test. Given suitable reference spectra for each sample, the resulting difference better highlights the additive depletion and the contaminant buildup in your oil.
This test reveals the type of wear debris in the oil and its particle size distribution. Keep an eye on how the distribution curve shifts over time. The current calibration standard for the particle counting test is ISO 4406. as is the case with any contamination, the test results have a natural lower limit of zero but there’s no upper limit. This is a measurement that ought to be monitored regularly and trended.
Wear particle monitoring depends on the magnetic properties of the particles. One test passes oil through a magnetic field to collect the ferromagnetic material. Measuring the flux of the collected material gives a reading that relates to the level of solids in the oil. Another type of test consists of passing some oil through a glass tube mounted in a magnetic field. The ferromagnetic solids remain behind. Shine a light through the glob of iron dust, measure how much gets absorbed and convert that to a measure of the concentration of these solids. Solids in oil is bad. Ideally, you want this reading to be no larger than zero.
These test rely on statistical process control to establish upper and lower limits on the acceptable level of solid contamination. However, the tests don’t give readings that are directly applicable to machine or lubricant adjustments.
If the bad news is a high level of ferromagnetic solids, then use analytic ferrography for exception verification. Analytical ferrography is another form of wear particle analysis. This approach identifies the composition of the solid chunks in your oil sample. The lab tech prepares a glass slide using a dab of your oil sample and examines it under a microscope. Shine a little light of one color on the top of the slide and light of another color on the bottom. The objective is to identify the solid pieces seen in the microscope eyepiece. Getting meaningful results as the identity of wear particles, their sizes, counts, size distribution and shape is a matter of experience and training.
Because this test is a visual inspection, experienced technicians can identify nonmetallic and other nonmagnetic wear particles. Because analytical ferrography is a direct look at the particle, it provides more objective information about the condition of the machine from which the oil was taken. The test is neither inexpensive nor quick because there’s a great deal of handling and expertise in testing this way. That’s why it’s best reserved for verification testing.
Producers also test their oil
Refineries also know about oil analysis. They must if they are to make a consistent product. Refineries blend oils and thus maintain consistent from one batch to another. Every grade of oil must conform to SAE and other industry standards. The standards provide a range for each variable.
The refineries attempt to hold their variation to a small fraction of the permissible range and try to keep the mean value centered in the range. It doesn’t matter whether they’re filling quart cans or tank trucks, they test at the beginning of the filling operation, in the middle and at the end. The load of oil doesn’t leave the refinery until the lab confirms that the batch that was just packaged meets the standards. It also assures users that the oil is consistent from lot to lot.
So, what now?
There are more test available – for instance, the entire series of ASTM standards for measuring oil. Space doesn’t permit a full exploration of the options available. The key to making sense of oil analysis is that no single test fives you the total answer. You must play the role of a physician. Take into account the interactions among the diagnostic results. Use 10 to 12 reports to establish a trend for each of the variables shown on the report sheet. Periodically submit a virgin sample of lubricant for the same round of testing.
Alarming test results
Remember, there’s an awful lot of random chaos sloshing through the universe. Remember, too, that no matter how careful and consistent you are in taking and packaging samples, no matter how rigorous and repeatable are the lab procedures, someday some of that chaos is going to rub off on the lab results.
The reported data will come back way out of line. When the reported data makes you gasp, it’s time to be a rational human being. Don’t panic and revert to the reactive mode of maintenance management. Getting away from that kind of nonsense is why you instituted high tech maintenance practices in the first place.
No, the machine isn’t self-destructing before your eyes. Don’t panic quite yet. Instead, take three giant steps backwards and look at it again from a new perspective. Talk to the laboratory. They might be able to retest any retained sample. Most likely the other diagnostic data and the predictive maintenance results are still within control limits. Watch the trends. Send out another sample to validate and confirm the bad news. Then act -– with the wisdom of an enlightened asset care professional. Then, thank those people that convinced you to get involved with oil analysis.
By the way, be sure to read the article in the September 1996 issue of Plant Services that delves into data trending in more detail.