Attached is an image looking into the fill plug of upper bearing oil reservoir of a large vertical motor.

The camera was pointed to see the boundary between the oil and the wall.

Based on the image assuming 0.95 emissivity, we suspect 146F oil temperature and 131F temperature of the adjacent wall.

By the way, the oil is very turbulent at it's surface.

Quiz: What do you think is the real oil temperature that would be measured if we inserted a thermometer into the top inch of the oil at this location.

I'm going to try it tomorrow so I'll know then... but interested to hear comments on what you guys think. I am very suspicious of emissivity difference because of the steep change in temperature at the boundary of the wall and oil. Or maybe this is explainable by the fact that the wall is relatively thin and cooled by ambient on the other side of the tank wall?

Pete,

Not knowing the material on the inside of the tank, it is good to question emissivity. Also consider that an infrared camera is designed to return a reasonably accurate temperature when you have compensated for emissivity, reflected ambient, AND camera is directed to a flat surface normal to the viewpath. In this case, you are on a very sharp angle, which makes accurate temperature measurement difficult.

Having said that, I still believe that it is quite possible for such a gradient to exist at the lube level. Oil has a much higher thermal capacitance than the steel of the tank. This means that the heat will not migrate quickly from the oil to its surroundings. The steel of the tank has a high thermal conductivity, which means that it can quickly transfer the heat from the oil it is in contact with through conduction. With a cooler headspace, and even cooler surface on the outside of the tank, it is not surprising to see a very sharp drop in temperature between the area in contact with the oil, and the area of the tank above the oil. Since the outside of the tank appears to be painted, you can probably be quite successful at finding the level in the tank by aiming the camera at the outside wall, and looking for this sharp thermal gradient. This is the very technique, by the way, that is used in IR thermography to find liquid levels in tanks.

As far as emissivity is concerned, remember that emissivity of a surface less than 1.0 returns a temperature which is a combination of the emissive and reflective portion of incident radiation. In the angle that you have taken the picture, the surface of the oil is indeed the reflected portion anyway, so even if the inside of the tank were of very low emissivity, the temperature you would measure on the wall would be close to the emissivity-corrected temperature of the oil surface. In this case, I believe that the inside of the tank, which is actually steel with a thin oil coating, is probably pretty good, perhaps >0.80, and as such, is presenting a fairly accurate temperature of the inside of the tank wall.

When you check with the contact and immersion thermocouples, check both the inside, where you've pointed the camera in this picture, and on the outside of the tank with contact probe, both slightly below and slightly above the oil level. I think that will confirm what you are seeing with the camera.

Good luck,

Rich Wurzbach
Certified Lubrication Specialist (STLE)
ASNT PdM Level III certified IR
Maintenance Reliability Group, LLC
rwurzbach@mrgcorp.com

Thanks Rich. That's a great answer and gives me some things to think about.

I did forget to mention that the inside surface of the reservoir is painted also.