Sudden malperformance of kettle type reboiler 1

[bloem]

All,

We are currently starting up a chemical facility after a revamp. For debottlenecking reasons, we've installed a kettle type reboiler to evaporate excess toluene from a reactor effluent. Process pressure is 3 barg (boilingpoint 170°C)
The kettle does not have a weir so we can freely choose the liquid level and thus the level of submersion of the tube bundle.

We observe the following:
Kettle operates perfectly at 10 tph 35 barg steam at say 50% of the bundle in the boiling pool.
When further increasing the steamflow to 11 tph, the flow falls to nil. Boiling stops and level increases sharply. after some time (minutes) we can start again with lower steam flowrates. The unit operates very calmly at the lower rates.

At a later stage we tried a fully submerged bundle and reach 13 tph steamflow before the problem recurred.

From an earlier day we learned that a too high liquid level leads to heavy entrainment... so the operating window becomes really small.

I don't believe we have a technical problem: Steamtrap is OK (checked today), steam control valve is OK (checked yesterday), condensate return shows no signs of mishaps (normal presures, other column reboilers operate perfectly)

We think we are "vapor blocking" the tube bundle, causing heat transfer to stop.

Has anybody ever observed this and can he give me some key data (heatfluxes or whatever)or observations to validate this phenomenon?

Regards
Bloem

 

[25362]

Before attempting to analyse whether there is a film-boiling condition or if and how the uncovered tubes act as desuperheaters for the throttled entering steam, allow me to say that logically an increase in steam flow rate, if totally condensed in the process, means more heat transfer, not less.

How do you know or measure and control your steam flow rate ?

 

[katmar]

Hi Bloem,

If you are using a standard arrangement for your steam flow control then you will have a control valve in the steam line to the reboiler. Although you are using this valve as a flow controller, it is actually a pressure controller. As you set the controller for a higher steam flow rate the valve opens and allows more steam into the reboiler, increasing the pressure and temperature in the reboiler. The idea is that as the pressure and temperature increase the heat transferred to the boiling liquid increases, thus keeping in balance the amount of heat coming in with the steam and leaving in the boiled vapor.

As you set the flow rate up to (say) 11 tph the steam pressure and temperature in the reboiler increase and the temperature difference across the tube wall may increase to the point where you get "vapor binding" and the rate of heat transfer will decrease. This means the steam pressure in the reboiler will increase further because the steam is not condensing. This will cause the steam flow to drop off and the controller will respond by opening the steam valve further. This further increases the pressure and temperature in the reboiler making the problem even worse.

Your maximum temperature difference looks to be about 70 deg C, which is very high.

This was discussed recently in thread124-105064 Have a look at the (probably conservative) maximum heat flux numbers recommended from Kern. NB maximum heat flux = minimum area.

 

[skoutso]

Hi bloem.

It sounds like a “vapor binding” of “film boiling” problem. As the steam flow increases, the duty increases as does the heat flux up to a point (11 tph steam) that a film of vapor is formed on the outer surface of the tubes. This film reduces heat transfer and the steam flow into the tubes. That explains why the steam flow is next dropping. With a totally submerged buddle you get more steam flow because of the more available area. Assuming that the 13 tph is your design duty I estimate that the available transfer area is around 85% of the required one. Calculate the heat flux (design duty / available area). The max heat flux must be limited to 12000. Ideally it should be around 9000 for kettle type reboilers. A vapor disengage space height of 12" (as minimum) is usually provided above the bundle. Consider the use of two vapor outlets to reduce the liquid entrainment.

 

[unclesyd]

Look at the tube spacing on the tube bundle as this condition is aggravated by close tube spacing, especially on triangular pitched bundles.

 

[25362]

Whether we are speaking of a boiler or a reboiler, the arguments for film-boiling are compelling enough, in particular when considering a revamped, clean, still not-fouled unit.

If I'm not wrong Bloem is speaking of a kettle designed to recover more than 50 ton/h toluene, what % is that on the reactor's effluent, and how much toluene remains dissolved -in equilibrium- in the kettle's liquid effluent (!?)

I'd feel much more satisfied if Bloem would have given us a complete picture including factors such as:

1. How is the process-side 3 barg pressure maintained constant?
2. How is the kettle being fed and at what temperature ?
3. How is its liquid level controlled ?
4. Is toluene the only boiling chemical ?
5. Why is it that Bloem suspects of "vapor-blocking" of the bundle at half-level when the unit apparently works OK also at full-level and ~30% more heat load ?
6. Some dimensions: what is the diameter of the bundle compared with that of the shell; what are the vapor line pipe dimensions; what's the length of the shell, does it have a dome with a demister ?
7. Is the 35 barg steam superheated or dry ?
8. What's the pressure in the condensate collecting header ?
9. Why was a (re)boiling step adopted instead of stripping?
10. If Bloem continues operating the kettle smootly and quietly with 10 ton/h steam, does it mean some of the toluene is lost ? Does the process include recycle of the (re)boiled stream ?

From personal experience, when detailing all the pertinent data, sometimes the solution to the problem pops up by itself. I hope Bloem succeeds.

 

[srfish]

Bloem,

It appears that the critical heat flux has been exceeded when the bundle is 50% submerged. What is the amount of heat exchanger surface ? Knowing the amount of surface will allow service heat flux calculations to be made. But before a comparison with the critical heat flux is made, more mechanical information is needed like tube spacing and bundle diameter.

By reading between the lines of your post, I believe there is not enough space above the top row of tubes.