Aerodynamics issue (at least, I hope )

[Micko]

Hello experts,
I hope you can help me with the following problem.
There is one huge pipe (diameter about 3 meters). In that pipe there is a flow of mixture ash and air. That mixture is very hot (about 300 °C) and abrasive. I need to measure temperature of that mixture. I placed thermocouple with its protective thermowell. I was surprised that thermowell was damaged soon. It seems that velocity in pipe is so large that in combination with abrasive ash, it practically cut off the thermowell in few weeks.
Thermo well is placed normal to the pipe (90°). There was some guy who though he'd solve the problem by putting something in front of the thermowell to protect it, so he placed heavy thick steel pipe whose diameter is about 3-4 cm bigger then diamter of the thermowell. He was reasoning that mixture's flow will first hit that protective pipe, disperse to the right and left and in such way protect thermocouple. This was better solution and now it can take about 3-4 months.
I wonder if this can be better.
Information about temperature is used in control logic so, real information must be fast and accurate. Heavy armor about thermowell is expensive and cause dead time.

Can you suggest some kind of smart protection using physics knowledge?
So far, it is made like this:
,,,,,,---->
flow ---->O o - thermowell
,,,,,,,,,,,,,,|
protective pipe

What would you suggest?
Basically I need something that will disperse ash particles and slow them down not to damage thermowell, or at lest increase periods of maintenance.

Thank you very much!

 

[FredGarvin]

The first thing that comes to mind is that, in stead of using a fully shielded thermocouple, go to a semi-shielded or aspirated TC. Keep the upstream side shielded the way you have it now but expose the TC element to the downstream side flow either by taking out a large section of the sheath or by putting holes in the sheath. This will help with the response time issue. It may mean a new design of the thermowell.

The other thing you could try is to take wall temperatures of the pipe with a welded TC wire if you feel the temperature gradient across the pipe is not too large.

Another thing to look at would be infrared, non contact measurement of the pipe inner wall.

 

[LURCH]

Yeah, infrared was the first thought to come to my mind as well. Alternatively, you could try putting the TC in a recessed position, so it isn't sticking out into the flow of ash. If I understand you correctly, the damage is being done by the abrassive action of the ash pushed by the airflow (so I'm assuming this is a fairly high-speed airflow, yes?). If you "T" in a short length of pipe normal to the main pipe, the temp inside should be the same as that in the main pipe, but any ash that hits the TC will be striking with greatly reduced force.

Just make sure you T-in the short pipe sticking up from the existing pipe, or the pocket of nearly-stationary air inside will cause the new pipe to fill up with ash. Can something like that be added on?

If the temp reading has to be taken at or near the center of the main pipe, then perhaps the smaller-diameter pipe could be replaced with a new one that is bent 90^o, with the open end pointing "downwind", and the TC placed inside the small pipe.

 

[russ_watters]

I like the surface temperature idea. If you insulate the pipe near the thermocouple, there won't be much of a gradient. And if it is that big of a deal, you can always calibrate it out.

 

[Danger]

The surface temperature and IR both sound good to me. I'm wondering, though, if replacing the steel shield with a rounded bit of some ceramic material might work.

 

[Cyrus]

Could someone retitle this thread fluid mechanics issues?

This is not aerodynamics related.

 

[Cyrus]

Try placing a ring of thermocouples along the outside of the pipe. Find out the thermal resistance value of the pipe itself, k, and back out the temperature of the inside wall indirectly. You can find this by taking the temperature at the wall with your thermocouple right now, and putting a new thermocouple on the outside wall at the same spot. Then find the k value. Afterwards, try and compare the value you get for the flow from the outside of the wall using equations indirectly to the ones you get with the actual one in the airflow.

If its within 5% for a week or so of operation, I think u can assume you did it correctly. Then you will never have to change thermocouples. Does this thermocouple measure the centerline aifllow tempearture, or the wall temperature?

 

[LURCH]

And, please keep us in the loop on this, let us know how it's going.