HYSYS Number of Stages

[C26M13]

Dear Engineers,

I have got one basic question on defining the number of stages in HYSYS simulation for a distillation column.

As an example,
- on the design datasheet of the column it says that the the number of trays is 22 and the feed enter at tray 10.

My question is why when we sepcify the number of stages in column in HYSYS, sometimes we reduce the number of trays (i.e. from design datasheet: 22 trays, but inHYSYS it specified as 11 trays). Is that because of the correlation of the tray efficiency (we give some assumption of the tray efficiency)?

And what is the difference between actual trays and theoretical trays? Which is actual and theoretical stages if I refer to the design data datasheet trays and in HYSYS trays?

I am not really sure with this concept. Hope someone can feed me with your explanation on my questions? Appreciate for your help.

Regards,

 

[don1980]

Simulations model stages, which means complete ("perfect") thermodynamic equilibrium between the up-flowing vapor and the down-flowing liquid. Trays are the hardware that's installed in an actual column. Tray designs vary because the internal vapor and liquid traffic varies. It varies within a given column, and from one distillation application to the next. The tray design attempts to maximize efficiency (maximize attainment of perfect vap-liq equilibrium) while at the same time handling the specific vapor and liquid loads. No tray achieves complete vap-liq equilibrium (no tray is 100% efficient). Therefore the number of trays must be greater than the number of simulation stages. Assuming the simulation has a sufficiently accurate physical property package, the simulation will accurately tell you the number of equilibrium stages required to do a given separation. Then you evaluate the tray efficiency to determine how many trays are needed to actually achieve that number of stages. The tray efficiency will depend on the particular tray design, the vapor-liquid loading, and the properties of the fluid.

 

[C26M13]

Thank you very much don19080 for your response on my question.

Further question on that, if i have 22 trays and then if i want to specify the number of stages in HYSYS (i.e. I specified as 11 stages foe a column in HYSYS), why 11 stages is used instead of 10,11,12 or any numbers? is there any reason to select the number as i mentioned?

 

[don1980]

If you are trying to model an existing column, then you enter the number of stages that matches the actual performance of the column. You know the chemical analysis of the overheads and bottoms streams, and you know the reflux ratio. So, vary the Hysis stages in the rectifying and stripping sections until you match the actual column performance. Then you have the correct number of stages in the simulation model.

 

[zahra-yari]

hi dears. i have the same problem. i want to define the tray number of an absorption culmn, for an amine sweetening, but i dont know how can i enter that in hysys? please answer me!

 

[chemlite]

I agree with don1980,
many authors as Kister (see for example the chapter Tray Efficiency Fundamentals in Distillation Design etc.) suggest to model stages with 100% efficiency, there are several correlations as Murphree and corrections for non-ideal equilibrium as amines etc. but in practice it is difficult to define the efficiency in different sections as that depends from many local factors, you can solve column and then convert to some real design (procedure is not too different from that adopted for packed columns)

for existing columns in order to estimate configuration you may regress existing data, see for example this discussion

https://www.cheresources.com/invision/topic/26796-aspen-plus-sensitivity-analysis/

I adopt a similar procedure with Prode Properties and Mathcad, some simulators include optimizers as discussed in thread.

 

[georgeverghese]

Note the condensor and reboiler, if included in the total no of theoretical stages, is 100% in real operations. So if the simulation says you need 11 trays total, inclusive of reboiler and condensor, then apply the overall tray efficiency to 9trays only.

 

[C26M13]

Thanks all for sharing your knowledge.

So, the number of stages in HYSYS model will be determined after we do trial simulation with the data and match with the column performance. Once it matched with column performance, then that's the number of stages to be used in HYSYS. Correct me if i got it wrong. (That means that i should have the number of stages in HYSYS will be different with what I have in my design datasheet)

 

[chemlite]

with a simulator you can model an existing column or design a new column,

when modeling an existing column you wish to match the actual performance of column,
for that you can define ideal or real equilibrium stages (see my previous post) and change the number of trays or other parameters to match the measured values,

in design stage you have some specifications (as, for example, purity in overhead liquid product etc.) but the procedure is about the same, you guess the number of stages and the software solves the mass and energy balances,

as said you can automatize the whole process working with internal optimizers or external tools (I use Prode with Mathcad or Matlab), the optimization process allows to solve problems such as minimum costs etc.

the mentioned Distillation Design (Henry Kister) or similar books can provide additional information.

 

[georgeverghese]

It is not likely that the simulation for a normal operating case would be the design case for the max no of column trays. The design case would be when you possibly have all of the following occuring at the same time in this column
a) Min concentration of heavy component in the overhead product
b) Min concentration of light component in the bottoms product
c) Feed composition should be close to the design case feed
d) Reflux ratio corresponding to highest normal utility coolant temp at the condensor
e) Lowest normal operating column pressure or highest normal operating pressure.

There may be other constraints that have a bearing on the no of trays also, which you'd probably know better for this column.

 

[zahra-yari]

i think we should at first design a shortcut column to find the number of trays(theoritical) then we can simulate our process...am i right dear Chemlite?

 

[chemlite]

yes, that is a possibility,
or you can solve n-times the column with different number of trays / parameters in order to identify the best configuration,
also, you may be interested to specify a range of operations for your column as suggested by georgeverghese.