Thermosyphon Circulation 4

[EmmanuelTop]

Dear members of Eng-Tips,

I have noticed cyclic behaviour of naphtha stabilizer during the last few days. The main cause of tower unstable operation is erratic heat input, delivered by bottoms reboiler. The column is operating at 80% of design capacity, with regular feed and regular temperature level of reboiler heating fluid (250C). The reboiler is horizontal H-type, recirculating without baffle (bottoms product has the same composition as reboiler feed).

The symptoms were:

1) Initial state: Tower bottoms temperature was 165C, while the reboiler outlet temperature was 170C. These parameters indicate normal reboiler operation.

2) After increasing heating fluid flowrate through the reboiler, reboiler outlet temperature increased to 175C and tower bottoms temperature dropped to 160C, indicating that process fluid flow through reboiler was reduced. Moreover, bottoms product flowrate started to oscillate between 0 and 60m3/h, followed by column overhead receiver pressure swinging (from 5.5 to 6.0 barg). Column behaviour has confirmed very erratic heat input, resulting in some sort of cycling. Cycle frequency is 5-10 minutes.
After decreasing heating fluid flowrate, everything came back and the tower operated again in its initial state. This phenomenon does not occur at lower feed rates (70% of design tower capacity and lower). Actually, this was the first time operating the tower at maximum bottoms temperature and feed rate higher than 70% of design.

Since reboiler is designed to achieve tower bottoms temperature of 175C at 100% tower capacity, what are possible explanations for diminished thermosyphon circulation?

- Undersized reboiler?
- Slug flow in reboiler outlet piping (cycling in tower operation)?
- Changes in boiling regime, or something else?
- Composition induced cycling?

According to A. Sloley, vaporization blanketing (binding) and critical flux limitation are the main causes of vaporization limits in reboilers. Since we are facing vapor phase superheating in reboiler (while net circulation rate drops in cycles, followed by light components in bottoms product), blanketing could be an explanation. Calculated (operating) heat flux is well below maximum design practice of 45kW/m2.

 

[katmar]

It could be an undersized reboiler, but my guess is that it is more likely a process piping problem and you are not getting sufficient liquid circulating through the reboiler. It sounds like this is a new installation, and it could be something as simple as construction debris left in the piping causing a restriction. My first steps would be to check the sizing of the process piping into and out of the reboiler, and to do an internal inspection of the physical piping.

Harvey

Katmar Software
Engineering & Risk Analysis Software

http://katmarsoftware.com

 

[reena1957]

Dear EmmanuelTop,
We have also faced similar problems in a Naphtha splitter of similar reboiler design. The basic problem was found to be the bottom level control oscillation. The bottom product was going through two streams with LC acting on FCVs in either stream with a selector switch. When we put the LC on the larger flow the problem vanished. The problem started when LC was the smaller flow and the bottom flow controls started influencing each other and started dancing to a tune of their own. The reboiler and top pressure controller also joined in and there was havoc. Get the controls decoupled OR see that the influence one on another is negligible. Another solution is to put a baffle on the reboiler side of bottom and see that the bottom tray downcomer goes in there.
Best wishes

 

[EmmanuelTop]

Reena & Katmar, thanks for your replies.

Katmar: reboiler inlet piping (U-tubes) could be fouled, at some extent. During the last turnaround we found deposits only on the bottom side of reboiler shell (process fluid inlet). Piping is 15" in diameter.

Reena1957: Bottoms level control is maintained via FCV on the product with smaller relative flowrate (stabilized naphtha to Petrochemical plant). Larger product flow is downstream naphtha splitter feed. Switching cascade between these two FCVs does not help, unfortunately.

Level fluctuations are very small (+/-1% deviation of LC set point), compared to bottoms product flowrate.

 

[reena1957]

Dear EmmanuelTop,
Pl see the following threads for some ideas:
124-66970,124-95533,124-100423
Best wishes

 

[katmar]

Emmanuel, Your initial post made it seem as if this was a new installation but now you speak of "during the last turnaround". If this column worked correctly as some time in the past then you can eliminate all design related causes. If the reboiler was not undersized before, it cannot be undersized now.

Something has to have changed. Have you changed the feedstock? the control strategy? the product spec? the flowrate? Clearly it is impossible for anyone to give you helpful advice based on the information available.

Katmar Software
Engineering & Risk Analysis Software

http://katmarsoftware.com

 

[25362]

I recall this subject was raised in a previous thread.

Have you considered the pendulum "U" tube effect that tends to appear on horizontal reboilers on higher vaporization rates ?

 

[EmmanuelTop]

Katmar, the column had worked correctly in the past because feed rate was always lower than 70% of design, with bottoms temperature of 175C (Reboiler outlet = 180C), or the bottoms temperature was lower (below 170C) at higher feed rates.
Operating the column with >80% of design feedrate and bottoms temperature of 175C is not possible - I do not know if this was ever possible to achieve - but it should be. I have looked in equipment datasheets, and the system is designed to achieve 175C bottom temperature for 100% (design) feed rate. What confuses me the most is the fact that cycling takes place when we increase heating fluid flowrate through reboiler, followed by vapor phase superheating and reduced process fluid flow through reboiler, with all this happening well below equipment design maximum and critical heat flux for naphtha stabilizer reboilers.

25362, your post is also very useful. It is strange, though, that cycling repeats after each 5-10 minutes, which is a much bigger period of time than the one calculated by Shinskey equation. What do you think about it?

 

[25362]

This is first time I've learned about the cycle time. There are many causes of oscillating "surges" in distillation columns with reboilers, beside the U-tube effect. Some of them are connected with the pinch point reached when increasing reboiler loads, and the effect this has on tower flooding and dumping sequences, as well as on bottom liquid level and its feedback action on the reboiler's operation and boiling r[é]gimes.

I'd recommend reading Henry Z. Kister's Distillation-Operation- (McGraw-Hill) to get a wider view of the possible reasons for the noted oscillations and ways to cope with them.

 

[katmar]

OK, so this is actually a "day 1" problem, even though the column has been operating for quite some time because the column has never achieved design throughput. The fact that you are getting superheating of the vapor points to the probability that you are getting insufficient liquid circulation - as you have said yourself.

Why would there be too little liquid? The only reasons I can think of are a low level in the column, or a restriction in the piping. You have said that the level control is working well, so the level is unlikely to be the problem. However I have come across problems of the type referred to by Reena. The downcomer from the lowest tray should discharge on the reboiler side of the baffle, and then you control the level on the other side. You may have leaking around/through the baffle, or from the downcomer, which could starve the reboiler at high rates.

This leaves the piping as the other prime suspect. As I said before, it could be a design problem or a physical restriction. What size is the vapor return piping? What circulation rate are you expecting? What is the heat duty?

The long cycle time indicates to me that it is a problem with the baffle in the column. Once you have lost level on the reboiler side it can take quite a while for it to build up again. If the problem were in the piping I would expect the cycle time to be shorter.

Katmar Software
Engineering & Risk Analysis Software

http://katmarsoftware.com

 

[reena1957]

Dear Katmar, There is no baffle(See the first post).

Dear Emmanueltop,
It might NOT be a reboiler problem at all. You are guessing it is so. But if the fluctations start anywhere in the system, similar oscillations will result. Before oscillations start every thing will be alright, but once it starts, you cannot say this is OK OR that is Wrong. It will be a self-feeding oscillation and you cannot segreagte cause from effect. How do you say it is the reboiler problem and not something else? Read 25362's post carefully before blaming the reboiler. Put all the controls on Manual and see if you can get rid of the oscillations by preventing mutual coupling effects. OR raise the capacity above 70% kepping the controls on manual. Comment please.

 

[EmmanuelTop]

Reboiler is fed from the column bottoms. There are no baffles (hot and cold side) in the column bottom - it's the most simple and most common reboiler configuration, called "recirculating without baffle". Heating fluid flow is controlled by three-way FCV. I apologize to Katmar and other participants if I presented this problem incorrectly at the first place.

Heat exchange area is 357m2
Heating fluid inlet T: 250C (stripped gas oil)
Heating fluid outlet T: 170C
Heating fluid flow rate: 60m3/h
Heating fluid (average) mass heat capacity: 2.89kJ/kgC
Heating fluid (average) mass density: 649.2kg/m3
Process fluid inlet T: 165C
Process fluid outlet T: 170C

Process fluid is stabilized (debutanized) naphtha, with C5-175C boiling point range.

As I have been told, this column operated with 110% of design feed rate in the past, but with bottoms temperature of 160C - and with lower operating pressure (5.3 instead of current 5.8barG - for the last 2 years, because of reduced condenser duty).

There are no problems to achieve bottoms temperature of 175C at column feed rates below 70% of design. I think
flooding/dumping process is not taking place in the column, because there are no overhead product yield changes, and pressure drop accross the column never falls below 0.2bar and also never crosses 0.4bar.
Column has 45 trays: 18 one-pass trays in rectifying section and 27 two-pass trays in stripping section.

Bottoms LC works satisfactorily, so it shouldn't be the cause for reduced circulation rates. System behaviour is always the same, regardless of the type of applied control (Manual or Automatic).

I found in Kern's "Process heat transfer" the following statement for oversized reboilers: "When a reboiler is oversized, it may operate by breathing. As liquid enters the reboiler, it may be completely vaporized very quickly because of the overdesign. New liquid replaces it and cools the surface down. The new liquid remains in the reboiler momentarily and is heated and completely vaporized also, so that intermittent bursts of vapor issue from the reboiler outlet instead of smooth continuous flow of vapor and liquid mixture. This can be overcome by placing an orifice on the shell outlet flange so as to cause an increased pressure drop on the vapor."

I would like to think of this problem as an example of composition-induced cycling or vaporization blanketing, but legendary N. Lieberman would probably say: "Your theory is nice, but without proof your theory is /\/\."

 

[reena1957]

Dear Emmanueltop,
How do you control the heating? Based on reboiler outlet temperature OR 4th OR 7th Tray (from bottom) temperature? Normally, the reboiler ouilet temperature control can lead to risky wrong control when the T vs vol percent vaporized curve flattens out and is similar to flat pump disch pr vs Flow curve at low flow rates of the pump. 4th or 7th tray twmperature gives better control depending on its slope with reboiled vapor. Do a bit of literaure/E-tips survey on this so that the hot fluid is controlled by something other than the reboiler outlet temperature! In the mean time try making the hot oil flow proportinal to the feed rate when you raise the feed rate above 70%. Check also the bottom Level indication against the Level gauge at two OR three levels to be absolutely sure that is not a "dead level" indiaction. What is the smaller bottom flow LC/FICV opening at 70% feed and can it take extra feed when you increase the feed OR do you have to proporionately increase the other bottom flow to the stripper manually?
Best wishes

 

[EmmanuelTop]

Hello Reena,

Tower top temperature is on cascade control with reflux flowrate. Reboiler is always on manual control mode (i.e., three-way FCV which controls heating fluid flowrate is always on manual mode of operation). So, there are always slight variations of reboiler outlet temperature and tower bottoms temperature, but I consider this insignificant (deviations are 0.3C) for tower operation.

Smaller bottom flow (naphtha to Petrochemical plant) FCV opening is held between 20-30% (cascade control), while bigger bottom flow (feed to downstream naphtha splitter) is on automatic mode. In case of increasing naphtha stabilizer feed, additional flow of bottoms product is routed to naphtha splitter, while smaller bottom product flow remains unchanged.

When stabilizer (and reboiler) are working fine, fluctuations of smaller bottom product flow are quite normal (18-25 m3/h, for valve % opening between 20 and 30). In such circumstances, there are no tower pressure fluctuations, and pressure drop accross the column is 0.25-0.30bar. In case of reboiler/tower upsets, smaller flow oscillates between 0-60m3/h while naphtha splitter feed remains unchanged, as I described before.

Thank you all very much for your assistance.

 

[katmar]

EmmanuelTop, As Reena pointed out, you made it very clear that there was no baffle. Sorry, I should have re-read the thread from the top before I commented again.

With all the additional info you have provided it is most unlikely that the reboiler and piping is undersized. If anything, it may be oversized and as you found in Kern, this can lead to "breathing". But I would expect the effect of the oversizing to have been worse when the column was previously operated at lower pressures and that was not the case. Also, my experience of "breathing" is that the cycle times are less than a minute.

You pointed out that the bottoms product flow varies from 0 to 60 m[sup]3[/sup]/h during the cycle. Doesn't this imply that the column boot level is varying by more than 1%? How does the phasing of the bottoms flow and level compare with the phasing of the bottoms temperature? From Reena's latest questions it seems they are also suspecting the level control. If you raise the set point on the level control do you get the same cycling phenomenon?

Katmar Software
Engineering & Risk Analysis Software

http://katmarsoftware.com

 

[EmmanuelTop]

Thanks, Katmar. I apologize to forum members because of misunderstanding during our discussion. This is the consequence of my inadequate knowledge and operational experience concerning process heat transfer and reboiler thermohydraulics.

I will write again all facts (observations) related to the naphtha stabilizer problem:

1) At some operating point of the system, increasing heating fluid flowrate through reboiler results in higher reboiler outlet temperature, but also in lower column bottoms temperature. (Up to that point, increasing heating fluid flowrate normally results in both higher reboiler outlet temperature and higher bottoms temperature.)
In my opinion, this reduction in tower bottoms temperature at observated operating point, is the symptom of reduced thermosyphon circulation.

2) At the same operating point, tower pressure and bottoms product flowrate cycling begins to take place.
Level in the tower bottoms remains the same in all operating regimes, while smaller product flow (in cascade with LC) experiences quite large fluctuations within range of 0-60m3/h. Other bottoms product flow (to downstream naphtha splitter) is held on Automatic mode (50m3/h set point), without changes or fluctuations.

3) This phenomenon occurs only at tower feed rates above 70% of design capacity. With feed rates <70% of design it is easy to achieve maximum bottoms temperature of 175C (reboiler outlet 180-181C), without any cycling or surge in tower operation.

Answer to Katmar's questions: phasing in bottoms product flowrate results in +/-2C change of tower bottoms temperature. Increasing LC set-point is of no help. Unfortunately, I do not know if this is the problem from "Day No.1" as you say, because this process unit is older than me

Thank you again.

 

[reena1957]

Dear Emmanueltop,
The last suggestion: Check the the level indications match with the actual and THEN raise the level as close to HLL as possible. If this does not solve the problem, take a shutdown and check the reboiler liquid and vapor lines for obstruction. One more thing! Is it possible to insert the tube bundle in a wrongly rotated orientation -- damaged rail ,etc.? Now I rest and when your problem is solved tell us the full and complete story! You wouldnt hear from me till then, at least on this thread!!
Best wishes

 

[25362]

The fact that a shift in boilup rate over a relatively narrow temperature range induces large swings in the V/L traffic with the recorded cycle times, may be the result of the combined hydraulic lags of all the trays, quick on vapor flow, slower on liquid downflow.

I'm inclined to think of a process lead-lag effect beyond a certain point (70% load) for an unstabilized feed whose composition may have changed over the years to include, beside CDU overheads, also light-ends-rich naphtha from other units such as the FCC or the hydrocracker; in particular since -being a total overhead condensing column- its pressure changes with constant top temperature maintained by the reflux, and a changing overhead composition.

 

[EmmanuelTop]

Bottoms level control was the culprit of heat input fluctuations. To confirm what I have just said, I need a little more time to monitor naphtha stabilizer operation.

Reena and Harvey, thank you for your genious assistance!

 

[25362]

If only real systems weren't so cantankerous !