Connie Hou, Lincoln Armstrong, Glenda Allum
Dr. Davis
CENG 403
December 10th, 1997
TABLE OF CONTENTS
FIGURES AND TABLES |
The focus of this project is to implement a heat exchanger network
for the ethylene oxide reactor and separator designs and to design
shell-and-tube exchangers for the network. Feed modifications to the
reactor project were made to create a continuous flow process
producing 30,000 lb/hr of ethylene oxide from the separation column.
Composite curve analysis revealed that five exchangers are needed for
this ethylene oxide process. Three exchangers require cooling water
at a total rate of 6,926 gal/min, one exchanger utilizes 55,000 lb/hr
of 150 psi superheated steam, and the final exchanger is a
process-process exchanger. Approximately $1.6 million are saved by
this integration. Total yearly estimated utility cost is $3.5
million.
The final exchanger and the steam exchanger connected in parallel
comprise the double reboiler for the distillation column. The reactor
effluent is used to supply heat to approximately 1/3 of the column
bottoms in the process-process exchanger. The design of this process
reboiler supported by rigorous ASPEN simulations consists of 250 1.5"
tubes 16 feet long encased by a 32" diameter shell for a total
surface area for heat transfer of 1570 ft2. The reboiler
is a BEM exchanger with an expansion joint. ASPEN generated cost
analysis priced the double reboiler at a total purchase cost of
$93,000: $31,900 for the process reboiler and $61,100 for the steam
reboiler.
Heat optimization and integration is one of the most important
aspects in the design of any chemical process. Utility costs are
indeed the largest of all operational costs and can save a company
millions of dollars if it is designed carefully and correctly. The
main objective is to utilize as many of the process streams as
possible to provide the required heat exchanges for different areas
of the plant and then implement steam and cooling water when extra
utilities are needed.
In the ethylene oxide plant, there are five main streams of
concern, four of which need cooling and one heating. The final system
is determined through heat network synthesis. This system consists of
five heat exchangers, three of which are associated with final
separation column. The design of the column's reboiler is modified to
optimize heat exchange by creating a parallel system of two heat
exchangers, one using a process stream to provide heat and one using
steam utility.
Both short-cut and rigorous methods are performed in ASPEN in
order to design safe, cost effective and reliable heat exchangers.
Important factors such fluid allocation, pressure drop, tube size,
and total surface area for heat transfer are taken into account.
The composite curve analysis is an effective way of
determining exactly how many utility heat exchangers are
required [1]. The composite curves for the four hot
streams and the cold stream are generated in MATLAB. This is
done by first calculating all the FCp's (flow rate*specific
heat capacity) for each temperature interval in order to
find the required or available heats. Phase changes are
treated separately because there is no temperature
change. The subroutine entH2 generates the information necessary
to create the s from the calculated heats and given
temperatures [2]. Since the curves produced may not
intersect, a shift analysis is performed on the cold
composite curve to find out how much utilities are needed.
The point at which the two curves cross is called the pinch
point. In this design, the streams above the pinch point do
not need utilities and streams below the pinch point do.
|
|
The existing layout in Figure 2 calls for five heat exchangers,
four which are to be cooled with on-site cooling water and the one
using steam as a heating medium. The four cooling exchangers are the
reactor feed cooler, the reactor effluent cooler, the column feed
cooler, and the condenser. The remaining exchanger is the
reboiler.
Table 2 dictates the original requirements for heating and cooling solely using utilities. Cooling water is available on-site at ambient pressure and feeds at 85 F. The cooling tower can process the water from 115 F to 85 F [5]. Steam is supplied from cogeneration units at 150 psi superheated to 400 F. The results are obtained from ASPEN using the short-cut method for heat exchangers. Since both the reactor design project and the separations project select NRTL as the thermodynamic package of choice, NRTL is also used for this project [3,4].
|
|
|
|
(lb/hr) |
|
|
|
|
|
|
|
|
|
|
|
|
|
feed cooler |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
From the information generated in constructing composite curves,
only one of the cooling streams has high enough temperatures to be
incorporated for efficient heat exchange in the reboiler. The design
in Figure 3 takes advantage of the only usable heat source, the
reactor effluent, in order to reduce the total amount of steam needed
to maintain the 72.2 MMBtu/hr heat duty for the column. This
integrated plan utilizes a double reboiler consisting of two heat
exchangers, one using the reactant effluent to vaporize a fraction of
the bottoms of the column and a second using steam to provide the
remaining heat for the rest of the bottoms flow.
The exchangers are connected in parallel for flexibility if the
need for maintenance or expansion arises.
5.0 OPERATING CONDITIONS AND
SPECIFICATIONS
The integrated layout still contains the same operating conditions specified in Table 1 for the reactor feed cooler, the column feed cooler, and the condenser. The need for cooling water to cool the reactor effluent stream is eliminated as is 33% of the steam supplied to the reboiler. Table 3 gives the process flows for each exchanger. Table 4 supplies the new process flow conditions for the double reboiler.
|
|
|
|
|
(liquid) |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
The two-exchanger reboiler for the ethylene oxide/water column
consists of a steam exchanger and a two-stream process exchanger. In
order to maintain the molar reboiler ratio of 1.07, the total duty
72.2 MMBtu/hr is required. The bottoms fluid enters the exchangers as
8430 lbmol/hr of liquid at 292.8 F where approximately 51.8 mol% of
the liquid feed is vaporized before returning to the last stage of
the column. The reactor effluent is able to provide enough heat to
32.4 wt% of the total feed. The remaining 67% of the bottoms feed is
partially vaporized in the steam exchanger.
6.0 SHORT-CUT EXCHANGER
DESIGN
ASPEN is implemented for short-cut design of all the exchangers excluding the process reboiler. Table 5 includes the total surface area for heat transfer and the overall heat transfer coefficient for each exchanger as produced by ASPEN. By inspection, the condenser has an enormous surface area required. This is attributed to the small temperature difference between the cooling water (85-105 F) and the condenser (114 F), thus producing a low driving force for heat transfer. Possible improvements are to implement an air-cooled heat exchanger, a chiller with refrigerant , or the use another process stream available in the preliminary separation section at a lower temperature for heat removal.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
7.0 RIGOROUS DESIGN OF
PROCESS REBOILER
ASPEN is used for the rigorous design of the process reboiler. A full specification sheet is detailed in Table 6. The exchanger has a total surface area of 1570 ft2 for heat transfer with an overall heat coefficient of 109.3 Btu/hr/ft2/F (dirty) and 131.6 Btu/hr/ft2/F (clean).
Service of unit Process reboiler Item No. 3 |
||
Size 30-192 Type BEM Position Horizontal |
||
Surface per unit 1570 sq. ft. Shells per unit 1 Surface per shell 1570 sq. ft. |
||
No. of units 1 Shell arrangement Engrs. GAA, LTA, CYH |
||
|
||
|
|
|
Fluid circulated |
|
|
Total fluid entering lb/hr |
|
|
Vapor lb/hr |
|
|
Liquid lb/hr |
|
|
Steam lb/hr |
||
Non-condensables lb/hr |
||
Fluid vaporized of condensed lb/hr |
|
|
Steam condensed lb/hr |
||
Gravity-liquid S.G. oAPI |
||
Viscosity-liquid CP.@avg. T. |
||
Molecular Weight-Vapors |
|
|
Enthalpy-Btu/lb |
|
|
Temperature in F |
|
|
Temperature out F |
|
|
Operating pressure lb/in2 abs. |
|
|
Number of passes Per shell |
|
|
Velocity ft/sec |
||
Pressure drop (max. allow.) lb/in2 |
|
|
Fouling resistance |
|
|
Heat exchanged-Btu/hr 23,400,000 MTD (corrected) 136 F |
||
Transfer rate-service 109.3 Btu/hr/sq.ft./F |
||
|
||
Design pressure lb/in2 |
|
|
Test pressure lb/in2 |
|
|
Design temperature F |
|
|
Tubes TP-304-SS No. 250 O.D. 1.5" BWG 13 Length 16 ft. Pitch 30o triang |
||
Shell C. Steel Floating head cover |
||
Channel: Channel cover |
||
Tube sheets-stationary S.S. Floating |
||
Baffles-segmental 8 Tube supports |
||
Type joints-shell Gasket and Packed Tubes |
||
Gaskets-shell Fltg. hd. Channel |
||
Connections-shell-in 3" Out 10" Series |
||
channels-in 20" Out 20" Series |
||
Corrosion allowance-shell side Tube side |
||
Code requirements ASME Code Stamped TEMA Class C |
||
Weights-each shell and bundle Bundle only Full of water |
||
Unit Price |
Amount |
|
Remarks: |
|
|
7.1 DETERMINATION OF HEAT
EXCHANGER TYPE
A shell-and-tube heat exchanger was chosen as the design basis for
the process reboiler because these types of heat exchanger are most
commonly used for two fluid heat transfer. Generally, reboilers are
designed as kettles or thermosiphons [6]. However, because
ASPEN does not have these designs available, a BEM one-pass heat
exchanger with an expansion joint was opted. The exchanger is a fixed
tubesheet exchanger with removable heads. Because of the temperature
difference desired in the reactor effluent of 110 F, an expansion
joint is necessary to avoid cracking of the exchanger due to stress
and strain.
A one-pass arrangement was selected because of the high flow rate
encountered in the reactor effluent. For reasons discussed in the
next section, the reactor effluent is allocated to the tube-side of
the exchanger. If a two-pass exchanger were selected, the velocities
in the bend of the tubes could accelerate corrosion and cracking of
the tubes that would result in tube leakage. Replacing tubes,
especially those tubes inside the bundle, in a U-tube exchanger can
be difficult if not impossible. Also, bundle replacement can be
costly. One-pass exchangers are more easily cleaned and repaired.
As mentioned earlier, the reactor effluent is allocated to the
tube-side for a variety of reasons. The effluent is 100-200 degrees
higher than the column bottoms. By placing the effluent in the tubes
as opposed to the shell, less insulation will be needed around the
exchanger, thus resulting in a safer environment for operators. In
addition, ethylene oxide is corrosive and should be restricted to the
tube-side. Finally, the high flow rates encountered would cause
larger pressure drops if the fluid is placed in the shell of the
exchanger.
Placing the column bottoms in the shell also has advantages.
Composed of mostly pure water, shell-side allocation of the bottoms
fluid eliminates the need to have a stainless steel heat exchanger.
Instead, only stainless steel tubes are needed, thus reducing the
base cost of the exchanger. The column bottoms runs at a considerably
lower pressure (60 psi) than the reactor effluent (230 psi). This
will allow the pressure rating for the shell to be lower than the
tube-side pressure, thus saving additional cost in shell thickness.
Also, since the bottoms is undergoing a partial phase change, the
shell provides more room for phase change.
The main problem to eliminate in any heat exchanger design is
pressure drop. Distillation columns are extremely pressure sensitive
so it is vital to minimize the pressure drop in the bottoms stream. A
0.4 psi pressure drop was achieved by two adjustments. The baffle cut
was enlarged to 45% from 40%, a design guideline. In addition,
correctly sized inlet and outlet nozzle diameters of 3" and 10"
respectively allowed for the partial phase change inside the
exchanger. Eight single segmental baffles support the tubes for a
baffle-to-baffle length of 21". This length will provide enough
support for the tubes as well as minimize turbulent buffeting.
On the tube-side, the reactor effluent is not as pressure drop
limited. This effluent is upstream of preliminary separation
processes that occur at lower pressures. A 1.8 psi pressure drop was
achieved by increasing the nominal tube size from commonly used 1.0"
tubes to 1.5" tubes to allow more room for flow. Although a larger
tube size will require a higher cost per tubes, the number of tubes
needed were reduced by using 1.5" tubes from 433 1.0" tubes to 250
1.5" tubes.
A majority of the base cost of an exchanger comes from the type of shell used, namely the shell diameter. Hence, a major point in heat exchanger design comes with using the smallest shell diameter possible, without having an infinitely long exchanger, and with preserving a desirable pressure drop and acceptable tube-side and shell-side velocities. This exchanger has a 32" diameter, less than three feet, and has a tube length of 16 feet. The smaller 30" exchanger was eliminated based on high tube-side velocities from having a fewer number of tubes to accommodate the required reactor effluent flow.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
$ 38,700 |
$ 11,000 |
|
|
|
$ 140,000 |
$ 44,500 |
|
|
|
$ 242,000 |
$ 394,000 |
|
|
|
-- |
$ 31,900 |
|
|
|
$ 3,086,000 |
$ 60,100 |
Total |
$ 3,506,700 |
$ 541,500 |
* cooling water cost based on $0.12/Kgal
** steam cost based on $6.68/Klb, 150#
*** calculations based on a 350 day operating year
The estimated yearly utility cost with heat integration is
approximately $3.5 million. Using the original design layout without
heat integration resulted in a yearly utility cost of $5.12 million.
Thus, $1.6 million are saved per year. The majority of these savings
is due to the 33% decrease in the use of steam.
The total cost generated by ASPEN for the double reboiler system
is $93,000, $31,900 for the reactor effluent exchanger and $61,100
for the steam exchanger. The biggest capital cost is the condenser.
As explained earlier, other alternatives for the condenser may reduce
capital and utility costs.
Due to the nature of the operating temperatures of each hot or
cold stream, the only heat exchanger network possible is the one
provided in this study. However, the actual heat exchanger design for
the double reboiler should not be limited to a BEM exchanger. As
mentioned earlier, the use of ASPEN for heat exchanger design is
limited since ASPEN does not provide options for the kettle, 'K',
reboiler.
The use of a more rigorous heat exchanger rating and design
program is recommended, such as HTRI's ST (sensible heat transfer),
CST (latent heat transfer), or RK1 (reboiler design) packages. HTRI
is the most commonly used design package for heat exchangers in
industry. It has an excellent physical property databank and provides
a more in depth analysis of heat exchanger fluid flow. HTRI programs
will consider low-induced vibration, flow fractions, turbulent
buffeting, and thermal resistances not available in ASPEN. HTRI can
also perform rigorous film calculations to better predict phase
change scenarios [7].
The heat integration for the four hot streams and one cold stream
in this ethylene oxide facility involves using the hot reactor
effluent stream to partially supply heat to the reboiler. The
reboiler of the column is a double exchanger reboiler, one using
reactor effluent as the heating medium and the other using 150 psi
steam superheated at 400 F. A BEM one-pass exchanger with an
expansion joint is used for the process reboiler. The total surface
area of 1570 ft2 is achieved with a 32" shell diameter
with 250 tubes of 1.5" nominal diameter.
The total utility cost is $3.5 million per year, $420,000 for
cooling water and $3.1 million for steam. Approximately $1.6 million
per year is saved by integrating the reactor effluent to supply heat
to the reboiler. Total purchase cost for the double reboiler is
$93,000.
[1]. Biegler, L., et. al. Systematic Methods of
Chemical Process Design. Upper Saddle River, New Jersey: Prentice
Hall PTR, 1997.
[2]. http://www.owlnet.rice.edu/~ceng403/
[3]. Betton, S. "Ethylene Oxide Purification." CENG 403
Project 2 Report, Fall 1997.
[4]. Coombs, J. "Celanese Clear Lake Ethylene Oxide
Reactor Revamp." CENG 403 Project 1 Report, Fall 1997.
[5]. GPSA Engineering Data Book. Gas Processors
Suppliers Association, 1987.
[6]. Branan, C. Rules of Thumb for Chemical
Engineers. Houston, Texas: Gulf Publishing Company, 1994.
[7]. Sensible Heat Transfer and Condensation,
Training Course on the Design and Rating of Heat Exchangers, Amoco
Corporation.
[8]. Gerhartz. W. et. al. Ullmann's Encyclopedia of
Industrial Chemistry. 5th edition. Federal Republic of Germany:
VCH, 1985.