This will be our final chapter dealing solely with Aspen Plus, and as such, we have decided to give an example that tries to pull together most of the work and information gleaned from the previous chapters into one. We will try to show you how to model an entire process, which you could hopefully translate to your own work.
Remember that your system might be a little more rigorous than ours, or not have the exact same pieces of equipment, but it isn’t impossible to model it if you take what you have learned here and try to adapt it to your own specifications.
Problem Statement:
In the process pictured below, we have a stream of natural gas at 400 psig and 110oF. This stream contains numerous saleable propane, butane, and C5+ (pentanes and above), as well as a large percentage of methane, all of which could be sold at a higher price in its purified liquid state.
To extract these, we first cool our feed to -20oF by heat exchanging with our cold residue gas, and then with propane refrigerant, thereby causing condensation of the heavier hydrocarbons. The feed is then introduced to the bottom of an absorber, where chilled lean oil (nonane) is
introduced at the top. The lean oil will absorb the heavier hydrocarbons from the gas. The methane-rich Residue Gas can then be resold at a higher price than purchased. The lean oil and hydrocarbons leave the bottom of the absorber, and are introduced to the top of a stripping column, where some propane and lighter hydrocarbons are removed by a vapor stream created by a reboiler. The stripper tops are recycled back and mixed with the feed, whereas the stripper bottoms are then fed into a depropanizer, where saleable propane is removed in the tops. The bottoms are then sent to a debutanizer, where saleable butane is removed. We could, later on in our process, remove any excess nonane from our process with another distillation column and recycle it back to the absorber, but let’s just stick with what we have now. Our goal in this example is to recover more than 85% of the propane in the feed. Its product stream can contain at most 1 mol% ethane.
Solution Methodology:
You may realize that you probably don’t have enough information to completely design this system. The absorption/stripping column can be pretty challenging, because, aside from not knowing the column dimensions, you also do not know the flowrates of the absorbing oil or the distillate rate in the stripping column. One of the best suggestions we can give is to use the GPSA Handbook, and find initial estimates for your specific column. Listed below are some good estimates for the first two columns.
Absorption and Stripping Column Specifications: (Partially suggested by the GPSA Handbook Fig. 19-20) Absorber Height: 23 ft. Stripper Height: 15 ft.
Diameter: 36 in. Diameter: 18 in.
Packing: 2 in. Plastic Pall Rings Packing: 1 ½ in Plastic Pall Rings
• Setting Up Your Absorber:
First, let’s setup our flowsheet. Our recommendation is to input one piece of equipment at a time, instead of putting in every piece all at once.
It is easier to try to figure out why your column dries up with one column than with 4. Without going into details because this should be somewhat obvious, column optimization is also much easier separately. So, we are only initially going to deal with our first main column, the absorber, and not worry about the recycle or the stripper. However, since there is a recycle, it wouldn’t hurt to add a stream mixer before the feed, as shown below in our flowsheet. Trust me, it will save you time in the long run, when you have actually added the recycle stream.
Next, you need to setup the specifications for your column in the Data Browser (If you are confused on how to do this, or are unsure, see Chapter 7 for details on setting up your absorber column).
o Enter all of the components into the component setup page, realizing that, if you are entering iso-Hexane you would need to Find using
“Methyl -Pentane.” Make sure to also add n-Nonane.
o Enter the Thermodynamic Property Method you will be using.
o Note: As well as the recommended Thermodynamic Property Tree, an index of most common abbreviations for the models is given in the online help. However, the list is not complete. A complete list of abbreviations is given in the online manuals, accessible by going to Start -> Programs -> AspenTech AES 11.1 -> Aspen AES 11.1 Documentation. Look for the pdf file Physical Property Methods and Models. This manual will have details of the parameter names and the models.
o Enter your stream specifications. Remember that both nonane and feed streams will be at 400 psig and -20oF. We don’t know what our flowrate of nonane will be, but we can take a fairly educated initial guess by using the suggested L/G ratio in the GPSA Handbook Fig.
19-19, and scaling it up for this process. A good starting guess for the mass flowrate of the nonane would be 1,600 lb/hr (this is a number we arbitrarily chose), but, as this number will be overridden later on, try not to marry yourself to any initial guess made.
o Note: If you are unsure of the volumetric flowrate of your gas, you can find a pretty reliable online weight, flowrate and volume converter for crude oil, its products, and various hydrocarbons at http://www.processassociates.com/process/basics/oil_vw.htm, or one of many other online unit conversion websites. You could alternatively use Perry’s Chemical Handbook, but using a conversion
calculator, such as the one on the website, will save you precious time.
o Now you need to enter the number of segments and the conditions for your Reboiler and Condenser (you have neither, so choose None for both). More segments usually translate into more precision, but make your preliminary guess. We are using 10 as our number.
o Important Note: From the problem statement, we are missing two important pieces of information: the column pressure, and the flowrate of our absorbing oil, Nonane. Since we are trying to optimize our system, while remaining within the specifications set by the problem, we will first need to choose one of these two variables to keep constant, while we optimize the column via the other variable. In our example, we chose to set the pressure to be our constant, and, later on, tried to find the “best” flowrate of nonane to obtain a desired percentage of our component(s) retained.
o So, first you will need to enter the pressure at the top segment, which can often be tricky. If you use a pressurehigher than that of your feed streams, your segments could dry up because there isn’t enough driving force into the column, which will give you severe errors, or stop calculations entirely.
o Note: Since this wasn’t really stated in the problem, we are going to set some goals for the amount of some components in our packed columns. That said, we will arbitrarily try to achieve a minimum of 95% recovery of propane in the absorber, and between 85 – 90% in the stripper, so that when it gets to the DePropanizer, it has more than enough room to achieve the purity specification, and still maintain 85% recovery.
o Shown below is a diagram of how some important components will be retained in our absorbing oil when we vary the pressure of the column (We created this using a Sensitivity Analysis after completing our first run, so don’t worry about where this came from right now).
o Note: I have only shown the effects on Methane, Ethane and Propane, since we are attempting to retain most of our propane, while getting rid of most of our lighter hydrocarbons. Our heavier hydrocarbons, you will find later on, should have an even better retention rate than propane (near 99%), so, for now, we will mainly be concerned with these three components.
% Component Retained vs. Column Pressure
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0 100 200 300 400
Column Pressure (psig)
% Component Retained
Methane Ethane Propane
For now, let’s assume that the added benefits to our separation gained from using a lower pressure won’t change too dramatically upon changing the flowrate of our absorbing oil. You may notice that, with a column pressure greater than 80 psig, there is no appreciable loss in propane, however a significant amount of our lower weight hydrocarbons are removed. For our current flowrate of nonane (1,600 lb/hr), at 80 psig, we have removed 95.3% of the methane and 48.5% of the ethane, while still retaining 99.4% of our propane. This is a pretty good number to use.
o Next, you need to enter your Packing S pecifications, Col umn Diameter, and Feed Locations (remember to use On Segment convention).
o Now enter any specifications for the mixer. This includes the stream pressure and temperature conditions. After you are done, Run your simulation. Your results should look something like those shown below.
We have successfully removed a fair amount of the methane and nitrogen, but some of the other components that we wanted to remove (ethane and carbon dioxide) are still present in relatively large amounts.
We will minimize this by now running a sensitivity analysis on the flow rate of our absorbing oil, but you could take a possibly more difficult path by creating a design spec and solving for the flow rate of a specific component. The latter is potentially more complicated because you have the risk of going crazy from running your design spec and finding it doesn’t work, or it doesn’t get the separation you want, etc. You may take either path, however, we specifically chose to use a sensitivity analysis on the flowrate of nonane because we can see easily see how many components change with varying the nonane flowrate. (For more information on performing a Sensitivity Analysis, see Chapter 2) o Firstly, and always remember to do this before you run your simulation, Reinitialize. You will find out that if you don’t reinitialize you
will get fatal errors, and Aspen will terminate its calculations. You probably won’t realize this at first, but if change your input and then run your fairly complicated simulation, Aspen will freak out! You will, at the very least, see warnings everywhere.
o Next, go to your Model Analysis Tools folder and then Sensitivity Analysis. Under the Define tab, create a flowsheet variable and input the information for the first component you want to look at; here, you will want to find the “Mole Flow” of Methane in TOSTRIP. When you are done, repeat this for the other two components we were looking at earlier (Ethane and Propane).
o Under the Vary tab, you want to input the information for the Mass flowrate of Nonane. I would recommend varying the flowrate from 500-4,000, incrementing by 100. Finally under the Tabulate tab, enter the name of each variable and the order you want their spreadsheet column to appear as. Run the simulation. Our results are shown in the plot to below.
% Component Retained vs Nonane Flowrate
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
500 1000 1500 2000 2500 3000
Flowrate of Nonane (lb/hr)
% Component Retained
Methane Ethane Propane
o By using a flowrate of Nonane of 1,000 lb/hr, we will retain 95.2% of our propane, 36.6% of the ethane, and 3.4% of the methane. Using this new flowrate, run the simulation once more as a check.
We have managed to optimize this column just by varying the pressure of the column and the flowrate of the absorber oil. If you want to try varying another aspect of this column (number of segments, packing type, temperature or pressure of the feed, etc) go for it. You might very well achieve a better separation. Using the new chosen flowrate, our results are as follows:
With your work on the absorber ostensibly completed, you will need to add your second piece of equipment, the stripping column. However, since you need the feed going to both columns at 400 psi, you will require the addition of a pump as well. So, go back to your flowsheet, and attach a Pump (found under Pressure Changers item directory) to the absorber bottoms, then add another RateFrac column to the stream coming from the pump. The diagram on the next page is a visual description.
o Note: In the final design, it is probably a good idea to add a pressure changer to the Residual Gas Stream, which exits at 80 psig. Re-pressurizing your residual gas (or adding a multi-stage mixed-propellant cascade refrigeration unit) will cause it to condense, thus making your methane saleable as well as yielding a 600-fold decrease in volume and making it easy to transport.
o Reopen the browser and add the setup information to the Pump block. This will really only include the discharge pressure of 400 psig.
o Now you need to set up the stripping column. This should be very similar to the manner in which you set up the absorber, except that you now have to include a Reboiler when you get to the Setup page. This also means that, if you have 10 segments, the Ending Segment is now 9.
o Note: Instead of using 10 segments, I recommend you try using 5 – 7. I speak from experience when I say that you will probably receive errors otherwise. I am not trying to dissuade you from using 10, merely saying that it happened when I ran my column.
o You have to assign it an Operating Specification, such as distillate flow rate, boilup ratio, reboiler duty, etc, so that it can minimize the number of variables. I personally think that using the Distillate Flow Rate makes it easier to find out later on how large your recycle stream has to be in order to achieve your specifications. A good initial estimate for your distillate flow rate should be around the size of the molar flow rate of the amount of impurities you are trying to remove (Methane, Ethane, CO2, N2), so try using a molar distillate rate of 1.75 lbmol/hr.
o Again, here we don’t know the column pressure. Shown to the right is a sensitivity analysis plot for % of component retained in the bottoms vs. column pressure. According to the chart, the optimum pressure to operate at would be approximately 30 psi, since the minimum ethane composition occurs there.
o Reinitialize, and run your simulation. Hopefully your results match ours below.
% Component Retained vs Column
While we do manage to remove a lot of the ethane bottoms stream, we also remove more propane than our spec allows. For right now, this is fine, since we will be recycling the stripper vapor back to the feed. So, now connect the STRIPVAP stream to the mixer, reinitialize and run the simulation. The results you obtain are ok, but still do not manage to meet the specifications that we were hoping for. To find the best separation possible, try running a Sensitivity Analysis comparing component molar flow rates to distillate rates. If you do, you should find that the best operating range for your distillate flow rate is between a flow rate of 1.99 to 2.4 lbmol/hr. Our diagram is shown to the right.
If you re-Run your simulation, you should obtain a bottoms stream fairly free of ethane that yields approximately 85% recovery. One thing to take note of is that the higher the distillate rate you use, the less room the depropanizer has to meet the recovery specification.
While the value we used, D = 2.38, was on the higher end of this operating range, this was OK since our depropanizer does a very good job of separating out our propane from our butanes. The final results for our absorber and stripping column with recycle are shown below.
To produce saleable propane and butanes, we’ll now need to introduce and design in Aspen the two remaining columns: the DePropanizer and the DeButanizer.
• DePropanizer
o The feed into this column will be the bottoms from the stripper:
Component lb/hr lbmol/hr Mole %
Ethane 0.43 0.0144 0.14
Propane 49.16 1.1158 10.84
iso-Butane 17.46 .3004 2.92
n-Butane 35.00 .6022 5.85
iso-Pentane 4.35 .0603 0.59
n-Pentane 14.50 .2009 1.95
iso-Hexane 1.73 .0201 0.20
n-Hextane 10.39 0.1206 1.17
n-Heptane 6.04 0.0603 0.59
CO2 0.017 .0004 0.004
n-Nonane 999.9 7.7960 75.75
Total 10.2914 100
o You should be decently familiar with how to set up a distillation column in Aspen, so we won’t walk you through the basics. Refer back to Chapter 5 to navigate yourself through the simulation.
o After inputting the feed rates, we’ll use the depropanizer in Chapter 5 as a start to simulating this column. We’ll start with 35 trays, feeding on tray 11, with a reflux ratio of 4.2 and distillate rate of 1.3 lbmol/hr.
o While running this column, stages below the feed started to dry up, so we lowered the condenser pressure slowly until our column converged at 160 psia. Our first converging run produces the following results:
o All of the propane is removed out the tops, but the purity is not at 99% yet.
o We tried lowering the distillate rate, and lowering the condenser pressure if the column dried up. While doing this, we maintained total recovery of propane, while increasing the purity of the tops to over 98.7%. We ended up with a condenser pressure of 140 psia (below).
We cannot really improve this separation, since ethane will come out with propane, and and the molar flowrate of ethane is 1.29% of that of propane, we won’t ever reach a 99% purity level in the tops.
Lastly, decreasing the size of the column to 30 trays and continuing to feed at tray 16 results in the same level of recovery and purity.
• Debutanizer
Next, we desire to recover as much of the butanes as possible from the bottoms of the depropanizer, with at least 97% purity.
o We’ll start the column with 30 trays, feeding at tray 15, and setting our reflux ratio to 1.5. and distillate rate to 0.91 lbmol/hr. We set the condenser pressure to 100 psia, which is the vapor pressure of iso-Butane at 120oF.
o The results from our initial run are shown on the next page.
o Here we see that we’ve recovered 100% of the butanes, and the purity of the stream is over 99%. Excellent results for the first shot.
o We’ll try to make the column a little smaller, by running a Sensitivity Analysis on reflux ratio, and feed tray for several column sizes.
o We were able to decrease the reflux ratio to 0.9, and use 25 total trays, while feeding on tray 9:
Our final results for the debutanizer are almost 100% recovery of iso-Butane, 99.8% recovery of n-butane, and a stream purity of 99%.
When you are modeling your process and you find that nothing seems to work, just try making slight adjustments, perhaps in error tolerance, or increment size. Mastery of ASPEN comes in time, but also understand that small things like that have, in reality, major consequences on how ASPEN operates. Fiddling with, or, as I like to call it, “fine-tuning,” can mean the difference in getting severe ASPEN errors on the first iteration and seeing your 7 column system with 3 recycles converge in 4 iterations. If you are still stuck, possibly consider modifying your process slightly. This could include adding a small purge stream, to aid in convergence, or specifying the data associated with a stream.
Wow, wasn’t that fun? Sorry for the blatant sarcasm (we thought it would introduce an element of levity), but I am sure you will have as much fun with this example as we did. Recycle streams are easily the worst possible curveballs you will have thrown at you, and to conquer that is
Wow, wasn’t that fun? Sorry for the blatant sarcasm (we thought it would introduce an element of levity), but I am sure you will have as much fun with this example as we did. Recycle streams are easily the worst possible curveballs you will have thrown at you, and to conquer that is