Abstract

This project optimized the heat exchanger network of an acetone production process. The final optimized design meets all of the specifications given in the base case. Modifications of the base case resulted in the elimination of low-pressure steam in reboilers by extension of the Dowtherm G loop, and significant reductions in cooling water by more effective heat exchange. The final design was rigorously modeled and features detailed design parameters for all of the heat exchangers in the process. The final design has an 83% yearly reduction in utility costs and a 7% reduction in capital expenditures resulting in overall yearly savings of 53%. The net present cost of $1.2 million is a 75% savings over the base case cost of $5 million.

Introduction

The purpose of this project is to optimize the heat exchanger network of the acetone production process proposed by Group C. Williams, and Kavinetor, Inc. The main goals in this study involve minimizing capital and utility costs associated with heating and cooling operations without altering any thermodynamic or physical conditions in the reactions and separations process streams.

Pinch analyses techniques were used to determine minimum utilities required by the process. Methods heat transfer between process streams were explored in order to minimize utility usage. Several preliminary design cases were proposed throughout the course of this study. The design case with the greatest economic advantage and most efficient use of energy was then further analyzed.

Base Case

Process Description

The acetone production process as designed by Group C. Williams and Kavinetor uses cooling water, refrigerated water, low pressure steam, electricity, and gas as process utilities. The following units were included in the optimization:

* The cooling water unit cost used in Carl Williams' report is slightly higher than the cost used in Kavinetor, Inc.'s and our report.

Heating of the reactor feed is accomplished by two heat exchangers, E-104 and E-100. E-104 transfers heat from the reactor effluent which exits the reactor system to the reactor feed. Subsequently, the reactor effluent is cooled from 356.3°C to 75.41°C, while the reactor feed is heated from 31.19°C to 108°C. E-100 supplies additional heating of the reactor inlet with Dowtherm G. Dowtherm G also serves as the heat transfer medium used in the shell and tube reactor system. After performing various heat transfer duties, a gas heater supplies additional energy into the Dowtherm G stream. Before entering the separation absorber in the separations process, reactor effluent must be at cooled to 20°C. Cooling water cools the reactor effluent stream from 75°C to 40°C in E-105. Refrigerated water then cools the reactor effluent from 40°C to 20°C in HTXCH. Significant utility usage occurs in the separation towers where low-pressure steam and cooling water run the reboilers and the condensers. Since wastewater from the waste water column must be discharged at around 45°C, HTXCH2 uses cooling water to cool the waste water process stream from 97°C to 45°C.

Heat Integration

Methodology

Many technical and economic considerations must be kept in mind while designing a heat exchanger network. In this particular case, modifications to the heat exchange network must not impact reaction or separation conditions in any way. Product specifications must not diminish as a result of heat transfer optimization. A minimum temperature approach of 10°C was maintained in every heat exchanger to prevent any temperature crossovers. Heuristics were followed closely to insure technical integrity of the design.

Pinch Analysis

A pinch analysis was found using the Matlab module ENTHC. The results showed a cold stream pinch of about 70°C and a hot stream pinch of about 80°C. In addition the minimum heating and cooling utilities were each found to be 6 MMBtu/hr. After the pinch analysis had been performed the streams could be matched up to determine optimal heat exchange. The pinch analysis was not used explicitly in determining the final design, as there was little useful information. It was obvious that the only streams available for exchange were the reactor effluent and the wastewater stream. Everything else would require utilities, and the final design focuses more on reducing the cost of these utilities through more effective usage than on crossing process streams. For example it is much cheaper to replace the low-pressure steam in the reboilers with the Dowtherm loop, than it would be to reduce the amount of steam needed through stream crossing.

The enthalpy in each stream was taken from the data generated by the previous designs and the average was found using the equation:

Havg=(Hout-Hin)/(Tout-Tin)

This and the corresponding temperature intervals were plugged into the Matlab Program ENTHC1 and the output was used with ENTHC to generate heating and cooling curves for the overall process. These curves were shifted until the minimum approach temperature was satisfied. The resulting plot shows where the minimum pinch occurs as well as the minimum utilities.

Optimized Heat Exchange Network

E-104, E-100, HTXCH2 (Heating Reactor Inlet with Reactor Effluent, Dowtherm G, & Wastewater)

For reactions to occur at yields specified by the base case design, the reactor feed stream must enter the reactors at 250°C and 290 kPa. In the base case design, E-104 uses the reactor effluent to partially heat the reactor inlet, while E-100 brings the reactor inlet stream up to the required temperature. E-104 does not involve any external utilities to conduct heat transfer while E-100 harnesses energy provided by a gas heater in the Dowtherm G stream.

Heating of the reactor inlet can be optimized by increasing heat transfer across E-104, and decreasing the duty in E-100. The wastewater stream in the separations process provided an additional heat source to raise the temperature of the reactor inlet. By using the reactor inlet to cool the exiting wastewater stream, heat was effectively transferred from the wastewater stream to the reactor inlet. Flows in E-104 & E-100 are shown below.

* We re-simulated HTXCH2 with the base case heat exchanger specifications in ASPEN PLUS and found that less water could have been used to cool the reactoreffluent. It is unclear why Kavinetor, Inc used the above water flow rate.

Cooling of the wastewater with reactor effluent successfully eliminated the use of cooling water in HTXCH2. The reactor inlet stream increased in temperature as desired without any additional utilities.

Heat transfer between the reactor effluent and reactor inlet was optimized by increasing the amount of heat transfer occurring in E-104. For a more detailed explanation of the heat exchanger sizing, refer to the section titled "Sizing of Heat Exchangers."

After optimizing E-104, the reactor inlet stream leaving the tube side enters E-100 66°C higher than the base case temperature.

An increase in the reactor inlet stream temperature to 175.3°C reduces the heat duty in E-100 required to keep the reactor feed at 250°C. A decrease in heat duty increases the Dowtherm G exit temperature from 350.2°C to 357.2°C, thereby reducing the amount of gas needed to maintain the Dowtherm G stream at specified conditions.

P-101, Acetone Column Reboiler, Wastewater Column Reboiler (Dowtherm G Heat Transfer Loop)

Dowtherm G served as the heating medium for most of the process's heating operations. Because of its efficient performance in the reactor design, our process extended the use of Dowtherm G for heating operations in the separation column reboilers. Kavinetor, Inc.'s design used low-pressure steam for energy in the reboilers. Although the use of Dowtherm G in the reboilers increased the amount of gas used, this process proved to be more economically advantageous than the base case design.

Using Dowtherm G to power column reboilers also resulted in an increase in electricity usage by P-101 (Dowtherm pump). More electricity was required to compensate for the increased pressure drop experienced. As mentioned before, the economic advantages of this design outweigh the increase in electricity and gas. For a more detailed breakdown of the costs, refer to the Economic Analysis section.

E-105, HTXCH (Reactor Effluent Cooling)

As mentioned before, reactor effluent cooling takes place over three separate heat exchangers: E-104, E-105, HTXCH. Heat transfer in E-104 does not involve any external utilities while E-105 and HTXCH make use of cooling water and refrigerated water respectively. Extensive optimization performed on E-104 was discussed in previous sections. E-104 optimization involved maximizing the amount of heat transfer occurring in E-104, which reduced the cooling required in E-105 and HTXCH. Since E-105 and HTXCH both use utilities to cool process streams, it is desirable to perform less heat transfer in these exchangers.

*Cooling water is normally at 30°C. It is not known why Carl Williams' group used 37.78°C.

In the optimized design, E-105 reuses refrigerated water from HTXCH. This eliminates the need for a separate cooling water source.

Optimizing HTXCH by increasing the amount of heat transfer occurring reduced the amount of refrigerated water needed to cool the reactor effluent from 40°C to 20°C. Our choice of simulation specifications also lowered our refrigerated water usage. Kavinetor Inc.'s simulation specified the exiting temperature of refrigerated water to be 15°C. Our design discharged refrigerated water at the warmer temperature of 23.53°C, therefore less refrigerated water was required in our simulation. Use of refrigerated water in HTXCH can not be avoided since cooling water enters at 30°C.

Exchanger Sizing and Specifications

The simplest exchanger type is a double pipe exchanger that consists of one stream flowing through an inner pipe while the other stream flows countercurrent in the area between the outer and inner pipe. This design has the advantage of being the cheapest, but it should not be used in situations where more than 100 ft2 of exchange area are needed. These exchangers are also not recommended for boiling or vaporization use and offer better performance at high pressures. Shell and Tube exchangers consist of an outer shell filled with many small diameter tubes. They can be used for situations requiring greater exchange area due to the fact that the large number of tubes increases the surface area available for exchange. However, in situations where the temperature difference between the two fluids is greater than 200 °C a special type of shell and tube exchanger called a floating head must be used. This design allows the tubes to move so that if the large temperature differences cause a differential expansion between the shell and tubes, there will not be any damage to the exchanger. In addition there are shell and tube exchangers designed exclusively for use in vaporization, which are known as kettle reboilers.

For the design, exchangers were chosen to be the ones, which provided all the necessary qualities at the lowest cost. E-104 was chosen to be a floating head condenser due to the large temperature difference between the two streams. In the distillation columns kettle reboilers were used for the reboilers and shell and tube exchangers were used for the condensers due to the large area needed. The wastewater cooling exchanger was chosen to be a double pipe as it was found to have a small area, and there was no phase change or large temperature difference. The wastewater cooler HTXCH2 was chosen to be a double pipe due to the fact that it required such a low transfer area. The design of the heat exchange with the reactor was not changed from that chosen by Carl William's project. The heat exchange area in the fired heater was not found, as only the duty was required in order to determine its price in CAPCOST. Heuristics were used to determine which process stream would go through the shell, and which would go through the tube. There were no corrosive or fouling streams in the process so this was not a concern. The design has condensing and vaporizing streams on the shell side, and liquid streams on the tube side. As there were no corrosive materials all exchangers can be constructed of carbon steel on both shell and tube sides. Detailed design considerations such as tube size and length were studies for all of the proposed exchangers. The recommended final design specifications are listed in Appendix 1. As can be seen in Table 7 the largest exchanger is E104, which does the majority of the preheating for the reactor inlet stream. The other large exchanger is Condenser 2. This condenser is on the final separation column, which is more difficult and thus requires higher reflux. It must have a large area in order to condense the necessary amounts of the process streams. The rest of the exchangers all have very modest areas. A comparison to the areas of the exchangers in the base case was not made since they were sized using a different approach, however the optimized design does have a reduction in capital expenditures of 6.75%.

The method suggested by Seider et al, p.333-334 was followed for the sizing of various exchangers. Sample calculations for E104 are listed in Appendix 2, One of the problems found when sizing the exchangers is that all of the parameters such as pressure drop and heat transfer coefficients depend on many factors and so must be found using an iterative process. Therefore various heuristics were used to find certain parameters so that complicated iterations could be avoided. It was initially assumed that all of the heat exchangers were shell and tube countercurrent exchangers. This was specified in HYSYS along with necessary temperature changes and pressure drops. The pressure drops due to friction were not calculated explicitly, but rather were taken to be the same as those given in Seider on p. 312, which depended mainly on the phase change of the streams. It is assumed that these pressure drops will result in turbulent flow allowing effective exchange of heat. With these specifications, the HYSYS simulation was run, and checked to make sure that the exchanger had a minimum approach temperature of at least 10 °C. With this condition satisfied, the second law of thermodynamics assures that all necessary heat exchange will take place.

Once the simulation was complete, the exchanger areas had to be calculated. This was done using the equation:

Q=UA*FTDTLM

Where,

Q=Exchanger Duty as calculated by HYSYS

U= overall heat transfer coefficient

A=Exchanger Area

FT= Fouling Factor which accounts for the regions with cocurrent flow found in shell and tube arrangements with multiple tube or shell pass arrangements as found by HYSYS

DTLM=Log mean temperature Difference as calculated by HYSYS

In order to solve this equation for the exchanger area, an overall heat transfer coefficient had to be found. This heat transfer coefficient was estimated using Table 8.5 in Seider, which provides overall heat transfer coefficients for a wide variety of fluids and gases. The heat transfer coefficient given in this table is for a dirty exchanger, so initially the coefficient will be somewhat higher. Experiments were made on each exchanger trying multiple pass configurations, but it was found that in general this resulted in a fouling factor lower than the recommended value of .85. This causes the exchanger area to be increased, and so all exchangers were designed to be simply countercurrent with a fouling factor equal to 1.

The first step in calculating the detailed design of shell and tube exchangers was to find the necessary cross sectional area in the tubes. In order to do this, a flow rate of 10 ft/s as suggested by Seider was assumed. The volumetric flow rate on the tube side was used to find the cross sectional area. With this area, Table 8.4 was used to find the necessary outer diameter and Birmingham Wire Gauge (BWG) to use for the tubes. The heat exchange area was taken to be the area on the outside surface of the tubes. Assuming a tube length of 16 ft, allows one to calculate the necessary number of tubes in the exchanger. Table 8.6 was used to determine the shell inner diameter to hold that number of tubes. The square layout was chosen as this makes cleaning the exchangers easier, and the necessary tube pitch depends on the diameter of the tubes. Tube pitch is the center-to-center distance between two tubes. For the double pipe exchanger, the diameter of the inner pipe was again found assuming a flow rate of 10 ft/s and the volumetric flow rate through the tube. Table 8.3 was used to determine the necessary pipe size to supply this diameter.

Safety and Health

Acetone is an extremely flammable and toxic liquid. When exposed to acetone, eye irritation and many other harmful symptoms can occur. Therefore, like all other processes involving volatile and flammable chemicals, plants manufacturing acetone should be well ventilated. Ventilation techniques applicable to the plant addressed in this report include dilution and local exhaust techniques. The exhaust system built should be separate from all other exhaust ventilation systems on site. Equipment design should minimize the release of acetone vapors. During disposal procedures, acetone should not be exposed to any strong oxidizing substances. Protective equipment should be prescribed for those operating equipment involved in the manufacturing of acetone.

As mentioned before, acetone poses as a serious fire hazard if handled improperly. Therefore, it is imperative that proper engineering takes place to ensure that this and all other acetone production plant designs take precautions to ensure the safe handling of acetone.

Environmental

The United States Environmental Protection Agency (U.S.EPA) recently exempt acetone from a few air pollution regulations because of its "negligible ground-level ozone forming properties." Acetone is also exempt from the Toxic Release Inventory (TRI) and VOC classification. These regulations make acetone a very attractive alternative for other solvents that are air pollutants and ozone-depleting agents.

Although acetone is not vigorously regulated in many instances, its release into wastewater should be minimized. Therefore, the wastewater column in the separations process of this plant should separate most or all of the acetone from the wastewater discharge stream.

Heat Transfer Fluid

Dowtherm G is a specialty chemical produced by the Dow Chemical Company, and is a part of the Dowtherm high temperature heat transfer fluid product line. It demonstrates thermal stability, improves heat transfer, and results in optimum operating economics. Dowtherm G was chosen in the reactor process because of its recommended temperature range of 29° C to 380° C. Dowtherm G was also used in the optimized heat exchanger network to supply heat to reboilers because it was a cheaper alternative than low-pressure steam.

Economic Analysis

Capital Costs

The economic conclusions of the optimized case proves the design's superior capabilities. The economic analysis focuses on the improved utilization of utilities and changes in capital cost. Capital cost changes resulted in 7% savings. The table below compares the various heat exchanger, fired heater, and pump costs for the base case and the optimized case. CAPCOST was used to determine the capital expenditures. The type of heat exchanger was also optimized.

The following graphs summarize the base case and optimized case capital costs.

The graphs and table shown above gives a more detailed cost breakdown of the base and optimized case capital expenditures. Our design used the same number of heat exchangers as before. Capital cost savings originate from more efficient heat transfer in each heat exchanger. For example, many of our heat exchangers perform the same duty as the base case with less heat transfer area.

Utility Costs

The various costs for utilities are shown in the table below.

The optimized case design resulted in an 83% decrease in yearly utility expenditures. Cooling water was reduced significantly, low-pressure steam was eliminated, the refrigerated water costs were cut nearly in half, and the gas energy was reduced. These reductions far outweigh the increase in the electricity costs.

The elimination of low-pressure steam and a significant reduction in cooling water and refrigerated water accounted for the utility cost savings. Improving heat transfer in E-104 (Reactor Inlet Heater/Reactor Effluent Cooler) resulted in an overall reduction in gas and CW/RW usage. The gas energy values include the fixed heat required for the reactor to operate at 350°C.

Conclusions

The optimized design reduced utility costs by 83% per year and overall cost savings by 53%. Heat transfer between utility and process streams was minimized because the optimized design relied on heat sinks/sources within the process. The optimize design did not alter any process stream and equipment operating conditions such as reactor temperatures, separation column designs, and reactor inlet/outlet temperatures. In addition to the optimized design presented in this report, other HEN scenarios were also studied. The following paragraph lists various networks considered during the course of this study:

1) Using reactor effluent to supply heat to the second reboiler while using Dowtherm G for the first reboiler.

2) Creation of a "cooling loop" with DowFrost or other heat transfer media in a design similar to the Dowtherm heating loop used in the base case design.

The optimized design presented in this report proved to be the most economically advantageous and feasible model. Our design met all specified objectives in a cost-effective and technically sound manner.

References

1. Nyalakonda, K. et.al. Kavinetor Inc.: Optimized Acetone Separations Design. 1999

2. Seader, J.D. et.al. Process Design Priniciples: Synthesis, Analysis, and Evaluation. John Wiley & Sons, New York, 1999

3. Williams, C. et.al. Acetone Production Via Isopropanol Dehydrogenation: Reactor Design. 1999.

4. http://www.ccohs.com/oshanswers/chemicals/chem_profiles/acetone/

5. http://www.epa.ohio.gov/opp/solvents/fact34.html