Thermocouple performance on Visbreaker Furnace outlet

[petrotrinsmerv]

We have a Visbreaker furnace with four coils. On each outlet the temperature is monitored (COT) and also on the main transfer line. A year or two after startup, we noticed spurious COT readings, i.e. 30 to 60 degF lower than the transfer line. The instrument department checked using a test thermocouple, but readings were about the same. We observed that the thermocouple tips were originally installed near the pipewall, which is different from where we typically place these tips, i.e. near the centre of the stream being monitored. So we changed the positioning accordingly. The result was: the spurious temperatures were seen even sooner than before. I have no explanation for this problem, I can surmise that it has something to do with the impact of coke laydown (we decoke every 12 to 14 months). Note that when the heater is down, all the temperatures equalize to ambient. I would like to know the following:
Does anyone have a similar experience with temperature monitoring in this type of service (thermal cracking)?
Are there any guidelines or Rules of Thumb for temperature monitoring in a coking/fouling service like this?
I must point out that we have a small continuous barrier steam flow into the main transfer line upstream of this thermocouple. This is to prevent backflow of hydrocarbon in the event of a tube rupture.

 

[EmmanuelTop]

Locating thermocouple near the pipe wall in two-phase systems (such is the outlet/transfer line in Crude/Vacuum distillation units etc.) almost always results in erroneous process temperature information.
In my opinion, the answer lies in heater hydraulic profile - due to highly non-ideal flow regime, it is not possible to achieve complete liquid-vapor equilibrium inside the heater coil and transfer line. In other words, there is always some amount of vapor superheating, compared with liquid phase equilibrium temperature at process conditions. Because of such phenomenon, for example, predicting wash-oil rates in Crude/Vacuum/VB fractionator becomes a very hard task - if vapor superheating is not accounted for in design stage.
If your thermocouple is submerged in liquid phase (annular flow?), I would expect significantly different temperatures than those measured at the pipe centerline.

 

[petrotrinsmerv]

I hear you. However, we did place the thermowell in the middle of the coil outlet, and it did work, for a while. As a matter of fact, the thermowell tip near the wall also worked for a period of time. The deterioration of the accuracy is where the question arises. The only thing that happens to the visbreaker coil outlet is that, over time, coke builds up.
Also keep in mind that, since this is a visbreaker, the degree of vaporization is much less than that of a crude, or even vacuum furnace. We have no problems measuring COT in these furnaces. I would estimate that the vapor volume for the visbreaker furnace outlet coil lies between 5 and 15% depending on unit severity. In fact this is a soaker type, where the majority of the thermal craking takes place in the downstream soaker vessel.
For these reasons I have doubts as to whether the flow regime is the issue, but I may be wrong. I am not in fact advocating to put the thermocouple near the wall, which is not our standard in any case. I just wonder whether the original designers may have known something I don't.

 

[25362]

At the temperature levels of a visbreaker, a small (usually negligible) amount of heat is radiated to the atmosphere from the external cap.

To verify if there is coke fouling of the surface of the measuring element, Lieberman and Lieberman, in their A working guide to process equipment, suggest to cover the outer thermocouple assembly with a piece of loose insulation to reduce the heat loss.

If, as a result, the temperature rises by 5 to 10^oF, they say, fouling is proved.