The Importance of Your Fuel Sampling Procedure

Testing is only as good as the sample. It’s impossible to get reliable or accurate fuel testing information without a good fuel sample. You might think that’s self-evident, but you would be surprised at the number of people who should know better but do not.

"Sampling" is the process of taking part of a whole, examining it, and then drawing conclusions about that sample that are intended to apply equally to the whole from which the sample came.

This concept has been important in "statistical sampling" for a long time. For many years, experts in statistics have explained that the size of the group we study and who is in that group can change what we learn from the data. Some people spend a lot of time learning how to choose groups in a way that gives them the most trustworthy information. When we say "trustworthy," we mean that the characteristics of the small group we study should closely resemble the characteristics of the larger group it comes from.

All of this applies just as much to fuel sampling as it does to, say, statistical sampling of people in the United States about what political candidate they like. You want the size and characteristics of your sample to be sufficiently robust enough that you can have confidence that its characteristics are close to those of your overall fuel from whence it came.

Fuel Sampling For Microbial Testing – Why It Matters

Because of the way fuels have changed, it’s virtually impossible to effectively manage stored fuel without doing some kind of fuel testing along the way. That also means that it’s virtually impossible to effectively manage that same fuel without good fuel sampling.

Nowhere is fuel sampling (and handling) more important than with controlling microbial growth in stored fuel. It’s recommended that stored fuel be tested for microbial content on a regular basis. How often can depend on the system’s ‘control interval’ – how long the system has been shown to take before the microbial levels pass problematic levels. In the absence of this kind of information, once a year is recommended at a minimum. Twice a year is better, quarterly is great. More than quarterly is probably too much.

Why does good fuel sampling matter so much? Why is “good” fuel sampling so critical to accurate microbial content results? And what defines “good” in this case, anyway?

If we’re going to be drawing conclusions about a larger system from what we observe (through testing) in a sample, then a good sample is going to be one that accurately represents what the larger system is like.

Yet, when we’re sampling to do microbial fuel testing, we have to realize that there’s a lot less uniformity with microbial presence in a tank than there is with physical fuel properties. That makes drawing a good sample for microbial testing even more important.

What do we mean when we say this? Let’s talk about three key things.

Key thing #1 - sampling where the microbes are

Physical fuel properties don’t really vary that much in different parts of a fuel storage tank, whereas the level of microbial content can vary tremendous from one part to another.

Let’s say you need to verify the cetane rating of some stored diesel fuel. Cetane rating isn’t something that’s going to be different at the top of the tank vs. the bottom of the tank. Viscosity wouldn’t change. Water and sediment content can differ, and so we account from that by making sure we take a sample from above the fuel pickup (since we know that water and sediment tend to settle on the bottom). Generally speaking, if you’re doing physical properties testing (like the tests specified in ASTM D-975), the fuel in different parts of the tank is not going to vary that much, especially on a horizontal plane.

This isn’t the case with microbial levels. It’s known that microbial levels will be higher in certain areas. Near the walls. At the interface with any free water that’s present. And even in places that we can’t really know about. Microbes grow where they grow, and they’re never distributed evenly within a tank.

What do we take away from this?

The most important thing is that if you’re sampling to monitor or detect microbial levels, you want to sample from areas in the tank that are likely to have the highest microbe levels. That means the bottom or the bottom-middle, near the interface. You wouldn’t take a sample from the top of the tank and expect it to be “representative” or “of diagnostic value”.

Representative means the sample accurately represents how the overall system looks.

“Of diagnostic value” means the sample is going to tell us something useful if we’re trying to assess or answer a question. If the question is “What are the microbial contamination levels in my tank?”, a top fuel sample isn’t going to be able to help you accurately answer that question, so it doesn’t have “diagnostic value”.

Back to the three key things.

Key thing #2 - don't screw your sample up by handling it wrong

It's important is to handle your sample correctly if you’re intending to use your sample diagnostically to assess microbial contamination in your tank. Once you draw a sample out of the tank and put it in a sample container, the microbial population is going to starts changing. Think about it – you’ve removed microbes from the environment they were used to and put them in a completely new environment. One that has different light levels, different amounts of foods and other substances they were using to grow, different temperatures, and different water levels.

Some of the microbes may decide they like their new home better, but some of them will not. The change may even kill some of them. Something will change in that population, and your job is to handle the sample so as to minimize how much of that happens.

What are the best practices for handling this? The big ones are to test the sample as soon as possible – preferably within 4 hours of drawing it. Put the samples on ice as soon as possible, but don’t store them in the freeze (put them in the refrigerator instead if you need to store them) – many microbes don’t really freezing cold and you’ll probably kill some of them.

Key thing #3 - you can't avoid natural variability

If you’re lucky enough to be using a quantitative microbial test like ATP testing, you have to keep in mind that there may be differences in microbial levels even within the sample that you take. Let’s dig into that a little more.

Say you take a 500 mL sample out of your storage tank and put it in a clean storage jay, taking it back for ATP testing. With any microbial test, you’re not testing the entire sample, you’re testing a portion of it. If you dip a culture slide into your container, what you’re actually testing is the fuel and liquid that came in contact with the slide and stayed on the growth medium. That’s a small fraction of your sample, which itself is a small fraction of what’s in your tank. For an ATP test, if you’re using a first-generation ATP “pen”, you’re only going to be drawing less than half an mL into the pen – a tiny fraction of your sample. And yet, that’s what the test is going to assess – it’s going to tell you how many microbes are in just that tiny fraction of the sample. Second-generation ATP tests (ATP-L) use a much larger sample – at least 20 mL, but you can test higher amounts if you want.

The point to remember here is that there’s a possibility that microbial levels can vary within even the very sample container you’re looking at. That means that if you’re going to run more than one test (especially if you find that you have to repeat a test from the same sample), you should really try to be consistent on where in the container you draw the sample from.

Also, you should expect some natural variability in the results from test to test. That doesn’t make them wrong – every test of any kind has a degree of variability. How much? If you don’t know what to expect, it might freak you out. You could have one sample that comes back with 1500 microbes per mL. You repeat the test, even drawing from the same spot (at least as much as you can remember that it was), and the result comes back at 3500 microbes per mL. What’s going on? In reality, that’s a pretty small difference that’s not enough to really worry about. Compare a difference of 2000 microbes against a threshold where you need to have, for fuel, 100,000 microbes per mL in order to take action. Now you realize that a difference of 2,000 wasn’t that much.

This kind of variability is a fact of life. When you know what to expect, and you know how to handle samples correctly, you can better account for it when deciding what to do.

You may be interested in this related post:

• Good fuel maintenance starts with fuel sampling

• For best fuel results, use an improved ATP test