Grease Sampling Methods Matter

After developing a lab in the power plant and focusing primarily on analysis of oils from the plant’s critical components like turbines, pumps and large motors, I was approached one day by one of our valve engineers.

The question: Could I also analyze greases? I had been training in analytical ferrography and had the equipment at my disposal. I was also aware of a “grease solvent” and a special process to dissolve the grease using glass beads. Armed with that equipment, I instructed the engineer to start bringing me samples.

After a month, I had received nearly 400 grease samples and had dissolved them all. I spent countless hours at the microscope identifying the particles I saw. I wrote reports for each sample and created a summary, showing my “dirty dozen” of the worst valves.

I felt satisfied that I had taken on this new challenge and come up with valuable information for my plant. However, after reviewing my report and investigating the valves I had targeted, the valve engineer had bad news for me. Some of the valves that I had identified as the worst actors were found to be in good shape. In other cases, there were known bad actors that I had identified as being fine. In other words, I was told that they had no further need for my services on valves.

How could I have gotten things so wrong? I looked back over my notes and took a second look at my ferrography slides. Had I misidentified the samples? Had I missed some clear signs that the separated particulate was telling me? After a review, I couldn’t see where I had fouled up. I returned to the engineer and asked a question that should have been the very first one I asked at the start of the project: How were the samples taken? The answer gave me insight into the problem.

“We shoved cable ties into the housing and wiped them off on the mouth of your oil sample bottles.”

Over the next hour, we both came to the realization that the sampling method was at the heart of the problem. It was likely that the grease that was adhering to the cable ties was “opportunity grease,” the material closest to the access plug. As we began to understand how the grease moved in the housing, it became clear that the samples I analyzed were in many cases grease that was not currently active in the lubrication of the valve gears.

A few of the “dirty dozen” turned out to be valves that had failed in the past but had since been rebuilt and were operating very well. What we had discovered was old particulate from the previous damage that was not completely cleared from the housing and ended up on the cable tie.

We learned a lot through this failed project about the valve grease and how it functioned in the gearbox, but we learned very little about the condition of the valves. I found that my failure to first address the sampling challenge had resulted in the previous month of my career producing very little actionable data for my company.

Several years later, I was afforded the opportunity to participate in a research project for the Electric Power Research Institute (EPRI). EPRI member utilities had identified a need to better understand greases and how they worked in power plant equipment.

From this, the Nuclear Maintenance Application Center (NMAC) of EPRI assembled the funding for the “Effective Grease Practices” research project and guideline. Among a number of areas of investigation was a focus on the challenges and best practices for motor-operated valve (MOV) grease sampling and analysis.

This gave me the opportunity to determine where my prior efforts to sample MOV grease had gone wrong and share those findings with the industry. Nick Camilli of EPRI provided the insight and leadership that resulted in the construction of an MOV test stand (Figure 1). This test stand allowed us to try various methods of sampling and characterize the movement of the grease in the valve while in operation. Sampling tools and methods were developed that were incorporated into a new ASTM standard for in-service grease sampling (ASTM D7718).

Additionally, the test stand was used to create a known higher wear condition to see if this could be detected by the grease sampling and analysis process. After the motor to the gearbox was deliberately misaligned, the result was a clear change in the measured wear content of the grease obtained by the new recommended sampling method, as seen in Figure 2.

A few years later, while attending an MOV conference, I overheard some engineers discussing the lubrication of MOV valve stems. Their concern was with short-stroking valves and the challenge in ensuring that an effective coating of grease was being delivered to the valve stem and stem nut when it was impossible to move the stem to expose most of the threads.

Why was the application of grease so critical in this application? Because the MOVs are tested by measuring the force of the motor through power signal analysis. By looking at the power signature of the motor, the closing force of the valve was inferred.

This is only valid when the coefficient of friction for the stem and stem nut interface is properly accounted for. If a stem is not fully lubricated, and the friction is higher than the calculations assume, then the force (and thus the power signature) normally sufficient to fully seat the valve is suddenly inadequate.

Knowing the challenge at hand, I returned to the lab to try out some solutions for relubricating the stem. Our first efforts failed, as we were unable to design a fixture that could be connected to the valve stem and deliver the high pressure of new grease required to displace the old and provide a new, reliable layer of lubricant. It was at a picnic where I got the chance to discuss this problem with my uncle.

Roy Leitz had a long career as a welder and pipefitter. He was also used to solving difficult problems. One such challenge he had was in sealing the high-vacuum valves of NASA’s space simulation vacuum chamber in Sandusky, Ohio. As he told me the story of how his team had overcome the challenges of sealing out the atmosphere to produce the world’s largest vacuum chamber for simulating the conditions of space, we came upon a design solution for fixing the leaking fixture to relubricate the valve stem.

The result was a new way to lubricate valve stems in critical applications like the MOV by taking a completely different experience and applying what had been learned. Not only had we developed a way to relubricate the valve stem, but in the process had created a new sample of the grease that had been in service, offering the potential for new condition data from the grease.

It took a long and painful experience in analyzing hundreds of meaningless samples to help me realize the importance of lubricant sampling. The best analysts at the top laboratories in the world cannot produce meaningful results from samples that have been taken incorrectly. It is essential that we focus not only on the task of generating analysis data but also in ensuring that the data will have value. This can only be achieved with accurate and representative samples from motor-operated valves and all other critical grease-lubricated machines.