The original column normally has less than 7 mol percent i-C5 in the overhead and less than 3 mole percent n-C4 in the bottoms product when operating at a distillate rate of D/F = 0.489. Can these product purities be produced on the smaller column at D/F = 0.489?

Pressure drops in the column will be neglected, and the K values will be read at 827 kPa (120 psia) in both column sections from the DePriester nomograph in Fig. 13-14b. When constant molar overflow is assumed in each section, the rates in pound-moles per hour in the upper and lower sections are as follows:

Top section

Bottom section

D = (0.489)(100) = 48.9 V = (1.75)(100) = 175 L = 175 - 48.9 = 126.1 V/L = 1.388

B = 100 - 48.9 = 51.1 V = V = 175 L' = L + F = 226.1 V'/L' = 0.7739

note: To convert pound-moles per hour to kilogram-moles per second, multiply by 1.26 x 10-4.

Since the feed enters at the middle of the column, M = 5 and M + 1 = 6. Application of the FUG method is demonstrated on the splitter. Specifications necessary to model the existing column are:

1. N = 10, total number of equilibrium stages.

2. Optimum feed location (which may or may not reflect the actual location).

4. Split on one component given in the following paragraphs.

The solution starts with an assumed arbitrary split of all the components to give estimates of top and bottom compositions that can be used to get initial end temperatures. The a^s evaluated at these temperatures are averaged with an assumed feed-stage temperature (assumed to be the bubble point of the feed) by using Eq. (13-36). The initial assumption for the split on i-C5 will be DxD/BxB = 3.15/16.85. As mentioned earlier, Nm usually ranges from 0.4N to 0.6N, and the initial Nm value assumed here will be (0.6)(10) = 6.0. Equation (13-32) can be rewritten as a 6'°(0.1869)

The evaluation of this equation for each component is as follows:

Fxi F

The evaluation of this equation for each component is as follows:

Component |
ai |
a™ |
0.1869a 60 |
FXf |
DXd |
BXb |

Was this article helpful?

## Post a comment