by hkooijman » 01 March 2016, 09:30
There are two ways to simulate a quench column: only for the heat transferred, or for both the mass and heat transformed. I recommend you to try to simulate the quench column the first way but then with one or two flashes is series. This will already get you a very good approximation for the heat being transferred, and the water saturation of the reaction mixture. It also allows you to check whether you are applying sufficient water to the quench, as we suspect is not the case in this flowsheet. The consequence of using too little water is that the liquid "dries up" at some position in the column and creates a numerical problem for our column model, as it cannot deal with stages w/o any liquid or vapor flow. Note that you added more than two stages in the quench. The number of stages you can get is very much dependent on the type of internals in the quench and the typical internals installed are sprays or packings. A layer of sprays will only get you 1 equilibrium stage if it is sufficiently high i.e. at least 1-2m. Also packing will only get you maybe 1 stage per 2m of installed bed (as typically large 3-4" packings are applied). If the cooling liquid is pumped around - as it often is - you might find the use of flash unit operations even more appropriate than using a column.
Furthermore, it is always advised to not add in recycle streams straight away but to act as an engineer and guess what you intend the recycle to be. This will not only enable you to only close the recycles at the very last moment but also to fully understand what you are accomplishing. Once you have the flash or two flashes in series working, you can try to replace it with a column operation again.
Note that your EoS model only accounts for the physical solution of HCl in water, which is a poor approximation. However, any more detailed modelling requires you to also use a thermodynamics model that can account for the solubility of HCl in water as well as for the heat of absorption the acid gas has. This probably requires you to use an electrolyte model which currently neither TEA nor ChemSep offers. This will not change in the immediate future.
Harry / ChemSep