Abstract

This project sought to optimize the separation of 1-heptene from a mixed stream of hydrocarbons. The original design utilized three distillation columns to purify four product streams: a propane stream for fuel gas, a C4 stream for LPG, a hexene stream, and the 1-heptene stream. The primary change made to optimize this process was to combine the propane and C4 stream from the first distillation column into a single stream, and send all of it to LPG. Additionally, each column was simulated in Aspen^TM and the reflux ratios, number of trays, and operating pressures were adjusted to yield the lowest possible cost over a 20 year plant life. The number of trays in each column was decreased, and the improved design overall saves about $177,000, or 6.5%, during the life of the plant.

Table of Contents

1. Introduction

2. Removal of C3/C4 Split

3. Column Arrangement

4. PFD Summary and Product Specifications

5. Property Package

6. Economic Assumptions

7. Distillation Column Design

7a: Column Sizing

7b: Reboiler and Condenser Sizing

8. Final Economic Analysis

8a: Cost Savings

9. References

Introduction

The objective of this project is to economically optimize the separation of a hydrocarbon mixture to ultimately produce pure 1-heptene. Heptene is valuable in industry as an octane booster for gasoline and as a plasticizer. In the base case presented in the textbook, three distillation columns in series are used to separate a reactor product stream consisting of hydrocarbons from C3 to C11. The first column separates C3 and C4 in the overhead, the second column separates C6 in the overhead, and the third column separates the heptene product in the overhead, with a heavy waste stream from the bottom. Rather than increasing the purity of the products, this project seeks to improve the cost efficiency of this process.

Removal of C3/C4 Split

In the original PFD, the overhead product stream from the first column, consisting of propane and C4, was split in a partial condenser, and the propane stream was sent to fuel gas and the C4 stream was sent to LPG. This precise split between the propane and C4 is not possible, due to the very similar volatilities of propane and C4, and to achieve such a separation would require a fourth column. However, the need for a fourth column was ruled out by a simple economic calculation. Using the price for fuel gas of $2.00 / mmBtu and the price for LPG of $13 / barrel, both obtained from a personal interview with Exxon employee Jonathan Coombs, the values of propane as LPG and as fuel gas were calculated and compared. Using these values, propane is worth approximately $3.87 / kmol as fuel gas and approximately $7.10 / kmol as LPG. Since propane is worth nearly twice as much when sold as LPG than when sold as fuel gas, there is no economic incentive to add a fourth column to separate propane and C4. The best and simplest alternative is to replace the partial condenser with a total condenser, and send all overhead product from the first column to LPG: this was the most significant change made to the initial process flow diagram.

Column Arrangement

The arrangement of the three columns was left as in the original PFD, based on the heuristics given in Turton et al [1]. For optimum column configuration involving multi-component separations, heuristics dictate that the easiest separation (that is, the one requiring the least number of trays and smallest reflux ratio) should be performed first, with the remaining products drawn off individually as overhead products. Since the original PFD has utilized this configuration, performing the relatively easy C4 / C6 split first, it was not altered. Additionally, each column was optimized independently to minimize the total cost of the process, as discussed in more detail in the column design section.

PFD Summary and Product Specifications

The feed to the separation towers is a mixture of propylene, butanes, butenes, 1-hexene, 1-heptene, octenes, and undecenes. The primary product is heptene, though hexene and LPG are also valuable and are to be purified. The mole fractions of the various components of the feed stream are as follows:

The C3 and C4 components are split off as distillates in the first tower, T-501, and become the LPG product stream. This separation is the easiest of the three, as there are no C5 components to interfere with the C4 / C6 split. Only 0.8% of the incoming hexene, 0.17 kmol/hr, is lost in the LPG stream.

The bottoms stream from T-501 is passed into T-502, which splits off the hexene as a distillate. This separation is far more difficult due to the similar volatilities of hexene and heptene, and therefore requires a larger number of trays. The produced hexene distillate stream is 94.4% pure, an improvement over the 93% purity presented in Turton et al [1]. Hexene recovery is 99% in this column, leaving little hexene to interfere with the difficult heptene separation in the final tower.

This final column, T-503, separates off the primary heptene product to a distillate stream that is 99% pure, meeting the required specifications. 98% of the heptene fed into this tower is recovered as distillate, and the overall heptene recovery is 97% (the remaining 1% is lost with the hexene distillate of T-502).

A heat and material balance of all major streams is included in the appendix.

Property Package

The RK-Soave property package was utilized to model the vapor-liquid interactions in this system. Both the RK-Soave and Peng-Robinson packages are known as very good models for hydrocarbon systems, and the RK-Soave model was chosen because Aspen is known to have concentrated on its development, according to Aspen Tech representative Brian Hart [2]. Stream purities using the RKS package were compared to those obtained through the use of the PR and ideal equations of state, and the key results are shown in Table 1 below. This table makes it clear that the choice of property package has little effect on the simulation results, as the system is nearly ideal. This ideality is the result of low-pressure operation, and the presence of only nonpolar hydrocarbons with only minor differences in volatility: there are no long chain hydrocarbons in the system.

Adequate experimental VLE data could not be found for this system, as was expected given the large number of components. The limited data that was found from Laugier et al [3] was only valid for very high pressures and only a limited number of components. Given the ideality of the system, the effects of experimental data on the property package parameters would most likely be minor; therefore, the data’s absence is not considered to be a major problem.

Economic Assumptions

The only costs considered in this project were those associated with the separation towers and their associated condensers, reboilers, and pumps. The flowrate and composition of the feed stream were unaltered, and the compositions of the product streams were held very near those specified in the text. Therefore, because no changes were made to the feed and product streams, an incremental economic analysis, considering only changes in the equipment and utilities costs, was sufficient to determine the economic effects of our design changes.

Operations costs were determined using the correlations presented in Chapter 3 of the text, while capital equipment costs were calculated by the computer program CAPCOST. The plant was assumed to have a 20 year life, and all operations costs were assumed to be constant over this time. All capital costs were assumed to be paid up front, while operations costs were discounted at 7% each year to yield a present value of all cash flows. This present value, combined with the initial capital outlay, represented the dollar amount that would have to be invested today to construct and operate the plant over 20 years, and was the key number used to evaluate different plant designs.

Distillation Column Design

The problem statement specifies that the heptene product stream must be 99% by mole pure, the hexene product stream must be at least 93% by mole pure, and that no more than 0.2 kmol/h of hexene should be lost in the overhead stream from column T-501.

Additionally, the reboiler temperatures were constrained to be no more than 155 C, so that low-pressure steam at 160 C could be used for heating. Similarly, the condenser temperature were constrained to be no less than 45 C, so that cooling water at 40 C could be used for cooling.

For purposes of optimization, each column was assumed to operate independently so that minimizing the total cost of each column individually would result in the lowest cost overall. This assumption is justified because the conditions of the product streams from each column do not change significantly when varying the design of the columns; the conditions and compositions of the outlet streams are implicitly specified by the project guidelines. In order to optimize the columns, the number of trays and reflux ratios were varied and the costs of each combination (including capital costs and utility costs) were calculated, with the results plotted below.

The optimum configurations are as follows:

 Column Sizing

The number of actual trays is the number of theoretical trays, as determined in the Aspen simulation, plus a 10% safety factor. The height of columns was determined by assuming a tray spacing of 22 inches, plus 4 feet at the top of the column for vapor separation and 6 feet at the bottom of the column to allow room for liquid holdup and the reboiler.

The diameter of the columns were calculated using the following equations obtained from Svrcek and Monnery [4]:

where Q is the vapor flow rate, and U is found from a Sauders-Brown equation:

Where r L is the density of the liquid, V is the density of the vapor, and K can be found if the pressure is known:

Reboiler and Condenser Sizing

The reboilers and condensers were sized using the basic heat transfer equation: Q = U*A*D T_lm for countercurrent flow. Heat transfer coefficients for refineries were found in Perry’s Handbook [5], and ranged from 482-511 J/(m²*s*K) for the condensers, while 795 J/(m²*s*K) was used for the reboilers.

Final Economic Analysis

As discussed above, each of the three distillation towers was optimized independently, yielding a minimum discounted cost for each tower. The total 20 year cost for the optimized process is $2,532,000. The following charts indicate the cost distribution.

Cost Savings

Much of the cost savings over the base case arose from a decrease in capital costs. The improved design reduced the number of trays in all three towers from the initial specifications in the base case, yielding significant capital savings as shown in the bar chart below. For towers T-501 and T-503, utilities cost increased due to increased reflux ratios; however, the increase was exactly offset by a decrease in utilities expenses in column T-502, caused by the decrease in column pressure. Other costs, such as capital costs for the reboilers and condensers, remained essentially constant. Total cost savings, after discounting all utilities expenses for 20 years, came to $177,000, equal to 6.5% of the initial base case cost.

References

1. Turton, Richard et al. Analysis, Synthesis and Design of Chemical Processes, Upper Saddle River, New Jersey: Prentice Hall, 1998.

2. Presentation by Aspen Tech representative Brian Hart.

3. Laugier, Serge and Richon, Dominique. "High-Pressure Vapor-Liquid Equilibria for Ethylene + 4-Methyl-1-pentene and 1-butene + 1-hexene." J. Chem. Eng. Data, 41 (2), 1996, pp. 282-4.

4. Svrcek, W.Y. Monnery, W.D. "Design Two-Phase Separators Within the Right Limits".Chemical Engineering Progress. October 1993.

5. Perry, Robert, Ed. Perry’s Chemical Engineer’s Handbook. Sixth ed. 10-44.