The Best Test Is The Test You Run

By Graham Roeber

What plant operators really want from field measurement tools, and why so many good methods end up on the shelf.

I spent nearly four years as an engineering officer on crude oil tankers. Water quality on a ship was never just about compliance. We needed reliable results while there was still time to act on them. Contaminated water could damage machinery or spill oil over the side.

That experience continues to shape how I think about measurement. Since then I have helped laboratories on five continents put oil and grease testing into routine use, and the same challenge comes up again and again. A validated, technically sound method may still be abandoned in daily practice if it asks too much of the people who have to run it.

Labs don’t run tests as often when a test is too hard to do. They pool samples together or send the work to an outside lab. Results that come back in days are of little use for day-to-day process decisions. In those cases, the problem is rarely a lack of measurement capacity. It is about whether the method suits the way the facility actually works.

What Operators Want From A Field Tool

Any assessment of a water-monitoring method starts with accuracy, detection limits, and regulatory acceptance — and for good reason. But from the operator’s point of view, the practicalities of running the test are just as important. Usually three of them decide whether a method gets used.

1. Time to result. A test that takes most of a shift will usually wait until someone has a free afternoon. A test that slots in between other tasks gets run when the result is needed. The number of steps counts too. Each transfer, rinse, and weighing increases the opportunity for error and makes it harder to get consistent results from one shift to the next.
2. Hazards and precautions. Hazardous reagents bring training, protective equipment, ventilation, storage, and waste disposal. Each adds cost, and together they can push a test out of the plant and into a contract lab. Every new chemical that comes onto site also has to be ordered, tracked, and documented.
3. An actionable result. A result a week later only tells you what already happened. A result the same morning gives you the chance to find the source and correct it while the conditions that caused it are still present.

Practical issues such as these tend to receive little attention in method validation reports, but they often determine whether a method succeeds in routine use.

Why Oil And Grease Testing Gets Delayed

Oil and grease in water is a good example. It shows up in discharge permits for utilities, refineries, power plants, and manufacturers. For decades, the conventional approach — EPA Method 1664 — has been to extract the oil from the sample with hexane, evaporate the solvent, and weigh what remains. It works, but it is slow, requires a flammable solvent and a fume hood, and takes up a trained analyst for much of a shift.

That is why many facilities do not run oil and grease in house. Instead, they send samples to a contract laboratory. That is fine for periodic compliance reporting, but operators come up short when they need to trace the source of an oil problem. One result per week or per month doesn’t tell you much about what goes on between sampling events, and by the time it comes back the upset is usually over. To identify an intermittent source, such as a leaking heat exchanger or an upstream process upset, you need several results close together in time, ideally from different points in the system.

Several alternatives now aim to close that gap. Solvent extraction analyzers read the extract by infrared and reduce the analysis to minutes, but they still use a solvent that must be purchased, handled, and disposed of. ASTM D7575 is a solid-phase extraction method that captures the oil on a membrane and measures it by infrared without a solvent. Online monitors give a continuous signal but need frequent calibration and maintenance, and their readings are not always directly comparable with permit methods.

None of these automatically replaces the method named in a facility’s permit. Typically, the permit method is kept for compliance reporting and a faster method is added for process monitoring. The plant tests more often, catches problems earlier, and still reports the way its regulator expects.

Questions To Ask Before You Try A New Approach

These questions are worth answering before you commit to any new method:

• Who is going to run it, and on what shift? Conduct the trial with your own operators. They know what the results mean for the plant, and they will quickly tell you whether the method fits their day.
• What does it need that you don’t already have? New instruments, reagents, fume hoods, and waste streams all cost time and money. Add the ongoing costs too: consumables, training for new staff, and the hours spent keeping the method in control.
• How often will you really use it? If the honest answer is less often than your process calls for, it is probably not the right fit, no matter how accurate it is.
• Is it backed by a published standard method? When it matters, data produced under a published standard holds up better with regulators, auditors, and your own management.
• How does it compare with what you do today? Run split samples side by side on your own water before you switch. Results from your own matrix tell you more than any brochure.
• What happens if something goes wrong? Find out who answers the phone, how quickly, and whether they can troubleshoot your particular samples.

Measure What You Can Influence

My experience on tankers taught me that the most useful measurement is one that leads to a decision. The same applies to treatment plants and industrial facilities. Accuracy and technical capability are still important, but a monitoring program only succeeds if operators can run the test often enough to find a problem before it is too late to fix. When you consider your next method, give the person doing the work as much weight as the spec sheet. The best test is the test that gets run.

Graham Roeber is CEO of Orono Spectral Solutions, a spectroscopy company in Hermon, Maine. Earlier in his career he was an engineering officer on crude oil tankers. He holds a BS in Marine Engineering Operations from Maine Maritime Academy and an MBA from Fitchburg State University.