17 replies to this topic

We are considering the installation of a new side reboiler in one of our distillation columns. We have a process stream that is now being cooled down in an air cooler and we could use this stream as a heat source for this side reboiler, with high energy savings.

In the column the traffic is high in both the stripping and rectifying sections. We have a very pure component as a distillate and a mixed hydrocarbon as a residue.

Feed to the column is in the 21st tray. It is a flashing feed (25% wt. vapor). Temperature profile is quite flat in the rectifying section with a sudden temperature increase just below the feed tray. So we consider the liquid in 21st tray as the optimal feed to the side reboiler.

Simulation of the column shows that 22th tray is the optimal return tray for the liquid/vapour mix exiting the side reboiler (the higher reduction in bottom reboiler duty with a lighter increase in reflux)

In your opinion, is that a feasible configuration? I believe the return to the column from a side reboiler is usually above the side-draw location. In this case, returning one tray below seems to be the optimal solution (the reduction in vapour traffic in the stripping section is almost the same as the amount of vapour generated in the side reboiler)

Thanks in advance for your answers.

[Zauberberg]

Hello MTR,

This is very interesting question. Your observation about shifting reboiler duties is correct: using low-temperature heat source in side-reboiler will unload the bottom reboiler. It's very similar to higher feed preheating. Being in your position, I would consider the following:

1) Tray loadings profile. Since the ammount of tower heat input remains basically the same, you need to consider new tray loadings on both stripping and rectifying setions. By installing side reboiler at the feed tray, you will unload the stripping section, while your rectifying part of the tower will face higher internal traffics. This will be benefitial if your stripping trays are close to flooding point. But, if you operate your rectifying trays very close to the incipient flood point, side reboiler will give you a headache, forcing you to cut feed or reflux (=product purity) rates in order you could achieve design product specs. Since you mentioned that you have process simulation tools, it would be very wise to compare tray loadings for both cases (with and without side reboiler).

2) According to Henry Kister, reboilers and feed arrangements are primary causes of tower flooding. Generally, you will have very complicated velocity profiles at the feed zone(s), so special care needs to be taken. You can read one interesting article at: http://www.fluor.com.....entry ups.pdf

3) Tray Efficiency. By shifting reboiler duties, you will change overall tray efficiency profiles. By unloading bottom reboiler and stripping trays, it is very likely for you to see your overhead components in tower bottoms product, if any part of new configuration is designed improperly.

When you put reboiler return below the reboiler feed tray, you will have higher reboiler inlet temperature (compared to base case), which lowers available LMTD and reboiler heat duty. Did you account for new temperature profile when calculating available low-temperature heat source for the side reboiler?

Please let us know how your project is developing. Distillation tower revamps are always challenging.

Hi

Traffic loadings will really improve. Now, the stripping trays are close to flooding, so the side reboiler will help to unload this section. We are far from flooding on the rectifying section.

We expect 5°C higher reboiler inlet temperature (compared to the base case). It has already been accounted in the simulation.

How could I estimate the new tray efficiency profile? I guess with the new configuration (feed to reboiler from 21st tray and return from reboiler to 22nd tray) and with the installation of the distributor, acumulator, etc...we will lose one tray. We are revamping a 40 tray column so we don´t expect a high lost in separation efficiency.

Thanks for the answers and for the article. I´ll take a look.

[Zauberberg]

You have analyzed your situation very well.

So now, in my opinion (it is based on what you have said about stripping trays load), efficiency will not be an issue, since you will reduce stripping trays internal traffic - and they now operate very close to the flood point. Reducing tray loadings in such circumstances will generally cause slight reduction in tray efficiency. Start the new simulation cycle by retaining the efficiency values from the original model, and then, on each new cycle, lower the estimated efficiency value by 0.05 and watch your bottom reboiler heat input increase at fixed bottom product specs. I can't say more at the moment because I do not know the nature of your system - tray efficiency can be as high as 0.97 and as low as 0.1, which depends on specific service.
By the way, did you estimate the percent of bottom reboiler heat duty reduction (roughly)? In some cases, this value can be a guide to efficiency reduction. Anyway, I think you have plenty of space to optimize between energy savings and product specs/additional heat input - as far as we are talking about tray loadings and efficiency. Of course, all this applies only if your process model is set and calibrated accurately.

The tricky point that remains is reboiler and tower hydraulics at transition spots. Please reffer to the article I posted in my previous message. Also, feel free to contact me if you need more articles concerning distillation towers/process modeling, which I cannot upload here without written permission of original website owner. The articles were free for download, but not for uploading to other web locations.

[Zauberberg]

I am posting a few articles, which are free for distribution. Although very general in nature, they may help you in some way:

[attachment=291:attachment]
[attachment=292:attachment]
[attachment=294:attachment]

With the future configuration, the bottom reboiler duty will be reduced from 60 MMBTU/h to 48 MMBTU/h. The side reboiler duty will be 13 MMBTU/h (the air cooler will take the extra million because of the increase in reflux rate) Unfortunately, because of the sharp temperature profile just 2 or 3 trays below the feed tray we can´t take the side-draw from a lower tray. So the traffic around the feed tray and the tray below will be "complicated" and as you say, an special care will have to be taken during the new distributor and acumulation design

Anyway, with the current NG prices the savings will be high enough to justify the project.

Thanks again for the articles and the answers

[Zauberberg]

From your previous statement, I presume you have a gas-fired, forced circulating reboiler.
Do you preheat the tower feed by bottoms product?

How about additional feed preheat instead of new side-reboiler? Since you already have mixed feed (L+V), you would not need to make any significant tower internals modifications. Also, according to existing feed zone temperature profile, as you wrote, the low-temperature heat source in such case will be quite appropriate, maybe even better solution. Did you consider it as an option?

Feed is already flashing. If we preheat the feed even more, the reflux will be much higher, and some energy will be lost through the air cooler. This produces the same effect as taking the side draw from the feed tray and returning from the side reboiler also to the feed tray.

However, if we return one tray below, our reflux will not increase so much and energy savings will be much higher. But I don’t understand well why. Maybe it is better because if we return to the column one tray below the feed, the produced vapors will “boil” the liquid from the feed tray and so, much less vapors will have to be produced in the bottom reboiler and a high reflux will not be necessary. What do you think?

Thank you.

[Zauberberg]

Hello again, friend. This is how I see things:

Heat In = Heat Out. In other words, heat that must be removed by column reflux is equal to the feed enthalpy + heat generated by reboiler(s). A well-defined simulation approach will give you comparison of both cases: in step one, build a column model with side-reboiler, generating the Q amount of additional heat. In step two, add to the feed enthalpy exactly the same ammount of heat Q, and watch for changes in reflux rates, for fixed overhead and bottoms products specs. What are the temperatures of feed tray with side-reboiler, and with additional feed preheat (by the models)?
By preheating feed, you will raise the temperature of feed tray, which you will also do by returning reboiler outlet stream one tray below the feed inlet. Even more, I think you will put more heat into the column by preheating feed, since you have slightly higher LMTD (compared to the side-reboiling option). See the column temperature and vapor/liquid loading profiles in both cases. I am very interested to find out the results.

Hi

What you propose is exactly what we did when we started doing simulations in order to find the best scenario for the column revamp

Adding heat (Q) to the feed increases the feed tray temperature from 125 to 133°C. And the reflux changes from 20 to 29 m3/h

If we take the side draw from 21st tray, add the same amount of heat (Q) in the side reboiler, and return to 21st tray (feed tray) we have the same temperature profile as preheating the feed and also the same change in reflux.

However, if return location from the side reboiler (after addition of the same heat Q) is 22th tray, the 21st tray temperature will be 127°C and my reflux will be 22m3/h. That seems the best option, but I am really not sure why.

So, even if I can put more heat in the feed, I would lose much of it through the air cooler.

I feel very confident with Hysys (I have used it with many distillation columns, and results fitted very well with process conditions) My only big concern now is clearly understand why the scenario with the return location below the feed tray is the best.


Thank you, Zauberberg

[Nirav]

Hello,

I have been reading this post and find it very useful and interesting. I also read your question on "eng-tips" and anser from Mr.Mathur on "eng-tips". I am looking at your concern about lower reflux in case if you add side reboiler return on 22nd tray. I would like to explain as follows.

In any of cases, Heat in (Feed + Reboiler) = Heat out (Bottoms + Distillate + Air fin cooler)

Case [1] : There is no reboiler or over heating of feed::
Consider following for reference.
Q1 = Amount of heat in column top vapor outlet (overhead)
T1 = Column overhead temperature
R1 = Reflux

Case [2] : Side reboiler return on feed tray :
Consider you are providing "Qr" amount of heat by means of side reboiler. Which means this Qr is available for a section of 21 no. of trays (top section above feed tray).

Now, to maintain same Column overhead temperature of T1, with higher heat content of Q2 in overhead (due to Qr), there will be higher mass flow of vapors.

Q2 = Q1+Qr = Amount of heat in column overhead.
You will need higher mass flow rate of column overhead vapors (based on m.cp.dT)

This higher mass flow is coming back to column in terms of "reflux" (R2 = 29 m3/h)because your distillate product rate is also constant.

Case [3] : Side reboiler return on tray below feed tray :

Here, Qr is available for of 22 no. of trays (top section above feed tray + one tray below feed tray).
Therefore, slightly less amount (copared to case [2] above) of heat is availalbe for top section above feed tray.
This is the reason why increse in feed tray temperature is less in comparision with case [2].
Q3 = Q1+Qr - heat loss in 22nd tray = Q2 - heat loss in 22nd tray.

Therefore, less amount of "additional" mass flow rate of overhead vapor (compared to case [2]) is required to maintain T1 (top temp.). Therefore, less amount of "reflux" (R3 = 22 m3/hr) compared to case [2] is coming back to column.

SO, ULTIMATELY,
Q2 > Q3 > Q1

therefore, R2>R3>R1.

It makes sense to me. However, I'm looking ahead for other member's analysis.

Regards,

There is an interesting article : "Use a side reboiler to increase tower capacity" from Nye, Herzog & Cheaney. Hidrocarbon Processing, September 1999, p51 to56. There you can see how column traffic changes with the side reboiler location.

The lowest the location of the heat addition, the lowest the increase in tower reflux. That´s what we also see in our simulation, and that´s why it is better to return to 22th tray.

However, in typical side reboilers, the return to the column is to the same side-draw tray, and not to the tray below. Why? It is, maybe, because if you return to the tray below, your side draw tray could run dry? (that would not be a concern if the return is to the side draw tray)


Regarding traffic and temperature profiles, it is the same:

-taking the side draw from 21st tray and returning to 22th tray
-taking the side draw from 22th tray and returning to 22th tray.

However, from the point of view of temperature difference with the heat source, it is better to take the side-draw
from 21st tray.

So, why it is not always done in this way in industrial columns?

Thanks.

[Zauberberg]

Required reflux rate is lower with side-reboiling (reboiler outlet below the reboiler feed tray), because heat generated by reboiler is lower, as compared to the case when you preheat the feed, or returning reboiler outlet above the feed tray: for fixed heating fluid flow and temperature (which is exactly our case), available LMTD is lower because of higher reboiler feed inlet temperature. You will never have the same heat Q entering the column (by preheat or side-reboiling-pumpdown) in these two cases.

Since you have fixed ammounts of distillate and bottoms product, the only heat removal capacity that remains is reflux rate (remember: Heat In = Heat Out). We may call it "an intent deception" - by returning reboiler outlet below the reboiler feed tray, you will put less heat into the column, than you actually could do. It would be the same as if you were cooling overhead vapor before introducing it to the fin-fans, and then say: hey, my fin-fans required duty has dropped! To prove this, you may perform the following simulation: vary the reboiler outlet above and below feed tray and watch for change in bottom reboiler duty, at fixed product specs - you will see that bottom reboiler required duty is lower in the case when you return reboiler outlet above the reboiler feed tray (or especially when you directly preheat the feed), simply because you put more heat into the column. In my opinion, that is why your reflux rate is higher. I couldn't think of anything more spectacular than that.

That is exactly what Nirav has also said: the aditional heat entering the column is lower in the case when you put reboiler outlet below the reboiler feed tray - it is simply thermodynamics, and the reflux rate must also be lower. You have much higher reflux rate only if you fix the bottom reboiler duty, and then add a new heat-source, which is unnecessary to achieve products specs.
What is interesting the most: if you have fixed products specifications, there should not be significant deviations in total heat input and total heat removal demand, compared to the base case. Using side-reboilers is efficient when you have low-temperature heat sources available, so you can cut the high-temperature heat source flow. Remember that your aim was to unload the bottom reboiler, and if so:

Side-reboiler Q + Bottom reboiler Q2 = Bottom reboiler Q1 (base case)
Reflux heat-out (side-reboiling) = Reflux heat-out (base case) at fixed feed & products flows and qualities

Bottom reboiler unload = Q1-Q2
Reflux 1 = Reflux 2 (roughly)

These are for 100% ideal systems, but you should not have great deviations from the above equations.
So, I can answer your question: why it is not always done in this way in industrial columns? (returning reboiler outlet below the reboiler feed tray) - it's because your heat input is lower than the one you could achieve in practice. And what would be the purpose of such complicated installation, when yu can simply reduce the heat source flowrate.
I would like to hear further comments from you. This topic is very interesting.

[Zauberberg]

I also apologize if my english is not so easy to be understand. Currently we are in the middle of process unit turnaround/revamp, which means 12+hrs working day and reduced writing/thinking capabilities:)

Best wishes to all participating members.

I think some of the statements are not clear. I´ll try to clarify them.

-- Products spec are fixed, and they are fixed specifications for the column simulation. Hysys will give us the bottom reboiler and air cooler duties (and so reflux, Tprofile,etc,...)

-- We fix also the heat added to the column with the side reboiler. It is always the same heat (Q). In our industrial process it can be achieved by varying the heat source flow. This heat source experiments some cross-exchanges with other process flows before arriving to the side-reboiler and flow can be ajusted to better fit with the needs of the side reboiler.

-- Feed to the column is in 21st tray.

--(case a) If we take the side draw from 21st tray and return from SideReb. to 21st tray with a heat addition Q, the effect is the same as preheating the feed to the column with a heat addition Q

-- (case If we take the side draw from 22st tray and return from SideReb. to 22th tray with a heat addition Q, my reflux will not increase so much as in case a. So bottom reboiler duty will be lower

--(case c) Simulation shows that if the side draw is from the 21st tray and return to 22th tray, with a heat addition Q, the profiles and traffic in the column will be the same as case b

--Temperature in 22st tray is 10°C higher than 21st tray so in the case c we have a higher temperature difference with the heat source, and an "easier" exchange with this external heat source.

So, for me, the case c is the optimal.

So my questions are still there. What I understand from your last e-mail is that in case c the heat input will be lower. I think it will not be lower but higher (if needed)(higher LMTD with the external heat source)

Have a nice turndown (which is not always easy)

[Nirav]

hello,

I agree with MTR in his following remarks.

I think it will not be lower but higher (if needed)(higher LMTD with the external heat source)

When i compared Q, i did not mean heat input by reboiler. I meant "heat content of overhead vapors". It is little difficult to explain by writting sometimes. I would again try to explain my understanding.

For better understanding, I assume following.
[1] side draw rate & heat duty of side reboiler is constant. This will make outlet temperature of side draw variable. (which is entering back in column).

Now, major heat content is in "vapors" from side reboiler. All the heat from this vapor is available to heat the content of column on and above the tray where it is introduced in column.

e.g.
-- if you enter vapor at 21st tray, vapors from side reboiler will transfer heat to material on and above 21st tray.
-- if you enter vapor at 22nd tray, vapors from side reboiler will transfer heat on and above22nd tray.
-- if you enter vapor at 23rd tray, vapors from side reboiler will transfer heat on and above 23rd tray.

So, as you go down in column, additional heat coming from side reboiler is being utilized by "larger" amount of material inside column. Which means addtional amount of overhead vapor produced in column top will keep on "reducing".
This will make reflux quantity to decrease. We can check this by simulation. I just had a small check and i found it. This will happen when you match overhead temperature and product flow rate.
That is the reason why "reflux" keeps on decreasing as you introduced side reboiler outlet below feed tray.

Now coming on to MTR's question. If reboiler return on lower trays is benificial, why it is not done frequently?

This could be because, as we go down, we need to heat the side draw "higher" to match column temp profile. However, it could not be feasible to achieve that higher temperature because we are utilizing heat from the waste sources (stream going to Air fin or water cooler).
So, there could be limitations on "heat source" to increase the temperature of side draw from column.

I'm still looking for more analysis, since i find it very useful and really interesting.
I'll read the article suggested by MTR.

Regards,

[Zauberberg]

Hello Nirav & MTR,

I have just checked column internal profiles, and the results confirmed what you said in your previous posts. I must say it is very interesting phenomenon.

From internal reflux side of view, this is what is happening:

- connecting side-reboiler anywhere above the bottom reboiler increases vapor flow upward the column (because it is easier to boil lighter components at the middle of the tower - with the same added heat Q - than in the bottom tray), compared to the base case. Higher mass flow of vapors causes higher degree of internal reflux vaporization. The higher the reboiler location is (for the same ammount of added heat Q), the bigger is the column vapor flow above reboiler return and higher internal reflux vaporization.

- In order to maintain sufficient internal reflux (i.e. = required level of separation), the external reflux from overhead receiver must increase. The net result is: higher vapor and liquid loadings in the column section between the side-reboiler return and column top, which has been stated at the beginning of this thread.

So now, the only question that remains is: why it is better to move side-reboiler return below the reboiler feed tray? The only thing that crosses my mind is - because of lower degree of internal reflux vaporization. This also means lower external reflux requirements. Your case is somewhat special, because you cannot place your side-reboiler below 22nd tray due to sudden temperature rise on tray 23rd and below. In such case, you would loose a great percentage of available heat in your side-reboiler.

Returning reboiler outlet below the reboiler feed tray should not pose any special concern: it is a common design practice for once-through reboilers.

What should be addressed further in your simulations, is:

- Separation efficiency if you loose 1, 2 or 3 trays because of side-reboiler installation. You didn't tell us how did you get tray efficiencies (calculated or assumed) in the base case, and if you have made any changes in tray efficiency when adding a side-reboiler? This is very interesting point and it should not be neglected, because you stated that most of the separation takes the place accross stripping section (sharp temperature profile).
- What parameters of distillate and bottoms product quality have you used in your simulation?

Hi

I think I found a possible explanation

The side-draw tray, must provide a confortable liquid feed to the side reboiler.That means we need some kind of acumulator in the tray. If I want to asure a liquid feed to the reboiler I can not use a tray with sieves or valves, because the liquid will be foamy and would jeopardize the reboiler´s performance.

If I can not use valves or sieves,I will have to use chimneys. And that means vapor by-passing the tray. So all the vapors returning one tray below will by-pass the tray, and the effect of returning below the side draw tray will be lost

Maybe that´s the reason why it is not done in industrial columns. Unless someone knows any kind of tray that could work for that scenario.

About the tray efficiency, I agree it is very important to have a good estimation. Specially because some trays would have to be removed with the consequent change in overall column performance

I specified weight percents of light and heavy keys in distillate and residue for the simulation. I think it is the most logical way to simulate.