The objective is to design the optimum conventional distillation column to separate
2x10^8 pounds per year of a mixture of 70% mole propylene and 30% mole propane into
a distillate that is 99.7% propylene and recovers 99.8% of the fed propylene as distillate.
Notice that the question doesn’t concern whether to build the distillation column; the
question is what the optimum design is. The constraints are that the utilities available in
CapCost are the ones available at the refinery where this will go.
When you did the propylene/propane alpha plots before the start of this project, you
should have seen that increased temperature (caused by increased pressure) gave a
decreased alpha which, in turn, gives an increase in minimum number of theoretical
stages for the separation or an increased reflux. However, if you can operate the overhead
at 45° C you can use 35° C cooling water ($0.354/GJ) instead of the least expensive
refrigeration, 20° C cooling at $3.44/GJ. You also have refrigeration available at 5° C,
-20° C, and -50° C. These suggest that you should look at towers that have distillate
temperatures at 15° C, -10° C, and -40° C. You can change the distillate temperature by
adjusting the distillate pressure.
What we have done so far:
Used aspen to determine the column pressures needed to produce overhead condenser temperatures of approximately 10 degrees above the available cooling utilities in Capcost. We have produced 3 aspen models for each of these column pressures: 130 stages, 140 and 150 (theoretical). We have costed these in capcost and got NPV values for each. So far, the best NPV value has been at around 65 psi which is on the low end of our selected column pressures.
We have also determined the relative volatility as it relates to increased temperature and in effect pressure values. We have found that decreasing pressures return higher volatility which should give an easier separation.
My question is this:
So far, the column pressure of around 65 psi has returned the most favorable NPV values. So should we now settle on that pressure for our column and proceed to find the best amount of theoretical stages? What is keeping us from going even lower for the pressure (which gives increased volatility)?
Any insight, recommendations, tips or general comments would be greatly appreciated.
Our calculation spreadsheet is attached as well.