Effects of thermowell on the flow profile and heat transfer

[reactorshell]

I have a catalytic fixed bed reactor with approximately 2400 tubes. There are several thermowells in the tubes of the reactor which measures the temperature along the bed. The question is, How i can ensure that the temperature measured by these thermocouples representative of the rest of the tubes within the reactor? Will the presence of the thermowell in a tube affect the flow profile of the process vapour flowing in the tube and thus the amount of heat transfer and therefore the temperature measurement. How can i validate that the thermocouple is reading a satisfactory reading?

Any comments, advice and suggestions are greatly appreciated. Thank U!

 

[hacksaw]

there are no pat answers to your question.

you will generally not be able to use conventional thermowell methods to get reliable measurements.

we had to fabricate custom thermowells with sensors as part of the equipment design.
they cannot be added later and they cannot be removed from service once installed, so all of your measurment needs have to be addressed up front.convential out sourced designs simply wouldn't work without creating the problems you've touched upon.


Even at that you will always have to deal with thermal environment your sensors are "seeing" and whether or not it is truely representative of the process.

good luck,

 

[reactorshell]

Hacksaw,

Does that mean there is no way I can validate the readings obtained from the thermocouples, except to perhaps reinstall the thermocouples and perform calibrations etc or even install new or better thermomeasuring devices?

How about doing a simple model to estimate the skin temperature at the thermowell and compare with the averaged bed temperature? Would that be useful?

 

[hacksaw]

reactorshell,

there are several issues here based on what you've said:

1. you need as accurate a temperature of the exit stream as you can get. two of the big players here are mixing and if you are dealing with a gas stream, thermal radiation from the bed and adjacent equipment.

the thermowell error for that measurement can be easily estimated. if the error is high there are thermowell design and sensor location options that can be used to reduce it. shielded /finned sensors can be used for that.

2. the temperature variations within the reactor are difficult problem. the measurement error (assuming you have calibrated and matched sensors) is a function of the thermal environment at each measurement location.

flow distribution and the temperature gradients within the in the reactor are critical. in one reactor we were seeing 50 Deg C changes in a 10 cm zone. the several sensors were only registering the slightly different placement within the reaction zone.

of course the reaction zone shifted with composition and catalyst activity, etc. and did little to contribute to measurement certainty or confidence.

without measurment errors, who could the process designers blame?

simply put, a traceable temperature measurement can only be obtained where the process temperature gradients are small. that is a tough requirement within a reactor.

3. you can measure tube skin temperatures but they have to be installed (welded in place) and they have to be the right sort for your application.

a thermowell temperature is okay, but you have to be able to define or identify what is being measured at the sensor location.

your concern about the thermowell blockage cause a change in flow distribution would have to be sorted out.

in the absence of radiation, most thermowell measurements report within a few degrees of the fluid temperature at the point of measurement. that is where the big disagreements begin: what is the sensor seeing in the reactor.

good luck,

 

[reactorshell]

Hacksaw,

I'm still pretty confused about the understanding of the thermocouples in a packed tubes. What should the thermocouples within the tubes read? The bulk averaged temperature? The center-bed temperature? Then the next question comes: How do we read when we are reading is true or the actual conditions within the tubes?

One more consideration: how would the flow into the tubes containing the thermowells be affected by the packing density within such tubes. My guess is that since packing density on such tubes will probably be lower than the rest of the tubes, a higher superficial velocity could be seen in the tubes and therefore, would the temperature readings be meaningful now? How can we make it more meaningful? Inert packings into such tubes to compensate for the lower pressure drop in these tubes?

THanks alot!!!

 

[hacksaw]

Those are all issues that you need to sort out be elimination.

thermowell block will reduce the flow in the tube it is placed(guess about your physical arrangement).

to the extent that you have now altered the flow rate in that tube you will bias the measurement. you wont know by how much until you try it.

blocking the other tubes is a bit over the top. you don't want your measurement to modify the process.

here is what you have: when you intall a temperature sensor you read a temperature, the question is whether that reading is representative of the process condition you are actually trying to monitor.

it is a tough meas. give it your best shot, make certain that it won't fail mechanically, and be ready for some challenges.

good luck,