Surface temperature measurement

[Graybeard]

Our plant has had several problems recently related to accurately measuring high temperature (~500C) surfaces using spring loaded flat tip stainless sheathed thermocouples. In several cases the reported surface temperature is up to 100C below what we think the actual temperature is. How? When you have an externally heated vessel or pipe and the material inside the pipe measures higher temperature than measured wall temp, we suspect error. We have moved away from thermocouples welded to wall due to maintenance difficulties. Is anyone else having problems like this? Is there fix/suggestion?

 

[m777182]

Try to check the surface temperature with an IR contactless thermometer.It samples and averages temperature over a small area of your surface.It excludes possible bad heath contact of a thermocouple. Then compare the measurement with the thermocouple readings not with what you expect to be or believe should be.
My experience is that "expected values" or "believe values" are very likely a consequence of some past missconcepts.
m777182

 

[phex]

m777182 was faster than me, but i fully second his suggestion to use contactless IR temperature measurement. Very reliable, very exact but of course also very expensive.

hth,
chris

 

[moltenmetal]

I presume it's pipe, and at 500 C I also presume it's insulated.

IR pyrometers aren't so expensive any more...but they don't take skin temperatures under insulation. You need to remove a patch first, which necessarily lowers the temperature of the surface being measured.

The issue of contact resistance and heat transfer along the sheath and conductors of a thermocouple is often neglected. The thermocouple sheath itself is a thermal conductor and the one end is at ambient temperature and the other is hot- clearly there's a heat flux along the sheath. If the contact between the junction and the object being measured is anything other than perfect, there will necessarily be a temperature difference across the contact resistance to drive that heat flux, which means the tip where the junction is will be at a lower temperature than the surface it's in contact with. Considering the heat transfer problem a bit more, a thin-sheathed thermocouple bent along the pipe for a few inches and strapped to and co-insulated with the pipe, rather than a spring-loaded flat-tip thermocouple, is going to give you a more accurate measurement.

 

[hacksaw]

You raised a couple of issues here:

1. What sort of thermocouple are you using?

2. How is the thermocouple terminated, where are the terminals, are the terminals polarized, what extension wire are you using, etc.

3. What sort of device is being used to read the temperature.

4. What form of cold junction compensation are you using.

5. Are all sensors showing the same defect?

6. What level of accurracy are you trying to achieve.

7. Are the sensor mounted on an external surface of the vessel or on the external of say a condenser tube?

Etc.,etc.

IR surface readings are pretty straight forward, but you will still need to have some means for independent checking at least initially.

 

[phex]

as a second thought, try to get some tft-based temp probes (i have to look up some manufacturers names) and try to find a metal-based glue which can take 500 °C. closing the air-gap between surface and thermocouple should solve your problem without changing much.

 

[Graybeard]

Some of the questions I can answer now, some I'll have to collect more info. This is not just one case, but the constant thread is that when we are trying to measure surface temperatures that are very low (-20C and lower) or very high (500 C and higher) as compaired to ambient, we see 10 to 20% closer to ambient than expected, and we have good reason to believe that the instruments are inaccurate. The latest case involves an electric heated solids dryer. Two external skin temp. TI's - indicate 350C. Stainless Steel sheated TC's, spring tip loaded in a SS tube welded to the shell of the insulated electric heated dryer. TC is supposed to be spring loaded tip against heater shell. TC's inside the dryer report solids temperatures of 450C. No exothermic reactions. No hot solids supplied, and not enough friction to account for temp difference. The equipment is near identical replacement for item that reported 350C shell and 200C bed solids. We have triple checked all wiring and externally verified that thermocouples are good. Two internal TE's report near identical temps, and Two external TE's are both reporting low. We didn't have this problem with other near identical equipment. Really old equipment had external TE's welded along wall. That provides best contact, but heck to replace.

 

[hacksaw]

have you properly insulated over the external sensors?

 

[6748]

Hi Graybeard,

01. The first thing to understand is “A thermocouple measures its own temperature”. It is least concerned with what is around it. It is our job to ensure that the thermocouple tip attains the temperature of the object whose temperature is to be measured.
02. A thermocouple picks up the heat from a zone/region/object that is at a higher temperature than its temperature. It release heat to a zone/object that is at a temperature lower than it. In this process of attaining a thermal equilibrium its tip temp reaches a steady value. Then it gives its millivolts or milliamperes which we convert to Temperature values.
03. For example: If we make the thermocouple tip to face an opening from a hot furnace at a distance of 4 or 5 metres, it will pick up the radiation heat and its tip temperature will go up. Since the ambient air that is flowing across the tip is at a lower temperature, it will loose its heat. Ultimately it will attain a temperature of say 40 Deg C, which is neither the hotness we physically feel (when we go that close to a furnace) nor the 1000 and odd Deg C that is inside the furnace, nor the ambient air temperature of say 30 Deg C. Keeping the location distance etc constant, if the ambient air flow increases the thermocouple will show a lower temperature, because the ambient air is picking up more heat now, whereas the heat pick up by radiation from the furnace is same.
04. Usage of thermocouple calls for locating the tip such that a) The tip will attain the temperature of the object/zone whose temperature is to be measured. b) that heat pick up from any other source,( like direct radiation heat from the electrical heating element) is avoided c) there is no way that the tip can loose its heat to an external zone that is at a lower temperature.
05. In this case your statement that – Externally verified that the thermocouples are good – implies that they are accurate. So try to make sure that the tips are placed/fixed/located/confined in a manner that they will attain the temperature of the thing that you want to measure--- without any heat pickup or heat sink from outside that will change the temperature of the thermocouple tip.
06. All the suggestions given by the members are concerned with the practical problem is ensuring the para 05 is complied to the best. Since the equipment is known to you, you are the best judge/designer to ensure compliance of para 05.
07.By the way what are you concerned? The inside solid temperature OR the outside wall temperature?

 

[Graybeard]

Thanks 6748. I don't have time to respond fully to your post and others because I'm off trying to prove to folks than this latest crisis can be overcome. Part of my frustration is that we have a bunch of instrument engineers that have their head up an their own orifice. We have complext systems that are difficult to maintain, and the guys are off doing cook books and cant seem to do much more than check continuity. Thank God you don't know where I work.

It seems lately that every time there is a process problem it's because we can't accurately measure high and low temperatures, and it all get's back to the stuff in your response. Either the instrument engineer hoses it up, or the maintenance or construction folks hose it up by not doing to best job they can - almost good enough is good enough when it's time to go on break, lunch, home, whatever.

As far as the IR type items are concerned, we have portable units used to measure surface temps, but they cant help us in this case because it's insulated and we can't get to the temps we need to measure without removing the insulation. We have (to my knowledge ever used a permanent mounted IR device.)

Our concern in this particular case is that overtemperature results in poor product quality in a product that is very expensive to produce. In addition, although this device is effectively a calciner, our fire protection and safety people have classified it as a Class B overn per NFPA 86. Ovens have to have thermal protection to keep them from burning themselves up, or catching fire, or melting into a puddle of stainless steel, or at the very least burning out the calrods. So if the surface temperature measurement are "lying" (under reporting actual surface temp.) then how do we prove that we are operating the equipment safely.

I'm off to oversee some testing. I'll report back what I find.

Thanks again for your help.

 

[moltenmetal]

Thermowells are the bane of temperature measurement. In an effort to make replacement of a thermocouple element easier, they can often throw out the accuracy of the temperature measurement itself and render it useless.

Here's a practical solution to reduce the contact resistance of your thermowell to thermcouple junction: coat the thermcouple with nickel antisieze paste before you insert it into the thermowell. At lower temperatures there's better stuff to use, but at 500 C nickel antisieze is about the only option you've got. At 500 C the oils in this paste will burn off, but you'll still end up with a film of nickel powder- far more thermally conductive than air. That may help, but won't eliminate the offset you're observing because of heat transfer across this contact resistance.

 

[hacksaw]

Graybeard, sounds like a tpyical operating plant so whats the problem?

Agree in part with molten metal that thermowells are a pain, but you can have the same problems only worse without them.

 

[StoneCold]

Greybeard
Here is a thought for you. I have never tried this exact application but the concept sounds good to me. Get some Zeston Heat transfer compound. It is mostly graphite and bentonite so it comes out of the can as kind of a wet clay/paste. Make a cone of it around your stainless tube to help bring the tube up to the actual process temperature. Then insulate over the top of it. I have used this stuff several times to drastically improve the heat transfer between steam tracing piping and the line being heated. I think it could help you here. I also like the nickel based antiseize. We try to put a very small amount of mineral oil in our wells to help with the heat transfer, but we run closer to 150C not 500C.
Good luck.

 

[StoneCold]

Greybeard
Another thought on old equipment. Is that many of the old thermocouples are J type and get replaced with K type. This throws the values off if you don't chance the settings in the monitoring loop to account for the change. I was burned on this once before myself.