HC Processing April 1995 PDF

Download as pdf or txt
Download as pdf or txt
You are on page 1of 11

April 1995

Amine plant troubleshooting


and optimization
Here is a practical operating guide
that uses control charting to maximize performance
R.G.F. Abry and R.S. DuPart
GAS/SPEC Technology Group

INEOS LLC
As part of a Federal Trade Commission mandated remedy to the merger of The Dow Chemical Company and the Union Carbide Corporation,
INEOS plc was able to purchase both Dows Ethanolamines and GAS/SPEC MDEA-based specialty amine businesses. This purchase became effective on
February 12, 2001.

INEOS LLC was set up as the newly acquired company, which includes the GAS/SPEC Technology Group. All the key Ethanolamines and GAS/SPEC personnel
were retained by INEOS LLC. All GAS/SPEC products, technology and know-how became the exclusive property of INEOS on a global basis.

Reprinted from HYDROCARBON PROCESSING, April 1995 issue, pages 41-50.


Copyright 1995 by Gulf Publishing Co., Houston, Texas. All rights reserved. Used with permission.
GAS PROCESSING DEVELOPMENTS

Amine plant troubleshooting


and optimization
Here is a practical operating guide tions when the plant is running smoothly. Without this
baseline information, future operational optimization
that uses control charting to maximize and troubleshooting will be difficult and time consum-
performance ing in addition to being shots in the dark. All upsets
can be traced back to a change in the process. Identify-
ing and responding quickly to these changes will result in
R. G. F. Abry, The Dow Chemical Co., fewer upsets and a more efficient plant operation.
Ft. Saskatchewan, Alberta, and M. S. DuPart, It is worthwhile for operators and engineers to trans-
The Dow Chemical Co., Freeport, Texas late the information gathered in the log to a graphical
format to easily track trends and identify problem areas.
systematic method for troubleshooting and These graphs visually describe to the operators and engi-

A optimization of amine plants, if properly used,


will result in fewer plant upsets, quick and
correct responses to changing conditions and long-term
neers how the plant responds to change. This is critical
information when either troubleshooting or optimizing
a facility. With a little extra effort, these trend graphs
profitable operation of any amine unit. It is important for will become the process control charts. They will serve
amine plants to maintain safe, continuous and optimized as a visual reminder of where the plant comfortably oper-
operations for short- and long-term success. Effective ates, if the process changes or drifts away from the norm.
and fast resolution of amine unit upsets plays a large part
in this success. These considerations are as important in TROUBLESHOOTING
plants using generic amines such as monoethanolamine This usually falls into four categories for amine plants:
(MEA), diethanolamine (DEA), methyldiethanolamine  Failure to meet specification
(MDEA) and specialty amines based on MDEA.  Foaming
The key to troubleshooting and opti-
mization is a systematic approach.
Developing and using control charts Sales ____MMscfd
____psig Acid
can also be used to monitor amine gas Surge
____F gas
Date ____MMscfd tank
plant operations. By using these tech- Company ____psig Reflux
niques collectively, a formal method Location ____F ____F condenser
Contact ____ppm H2S ____F ____F
for troubleshooting and optimization Phone ____% CO2 __F
Fax
can be established. This will ultimately ____F
____gpm
__psig
____F
result in a more trouble-free, contin- ____psig
Absorber ____# of ____F Reflux
uous operation. trays accum.
Inlet ____# of
gas
Data collection. Successful amine ____MMscfd ____P,
trays

plant troubleshooting and optimiza- ____psig psi Filters ____gpm


____F
____P, Stripper
tion relies on reliable data collection ____% H2S
____% CO2 ____F ____F psi
and interpretation. Because data col- ____psig ____F Reboiler
lection is so important when trou- ____F Amine
bleshooting or optimizing an amine cooler
____F
plant, listed in Fig. 1 and Table 1 is ____gpm ____F
____psig
the minimum amount of data that Inlet ____F ____F
____psig
should be collected on a routine basis. gas
____psig
____F
separator ____F Cross ____F
Frequently these items are collected exchanger
as part of the operator s daily log. Flash
____F
____gpm ____F
____gpm
tank
Items that are missing should be Source
____lb/hr
added to this log. Barometric pressure psia
It is critical that data be collected
during or at routine operating condi- Fig. 1. Amine plant data.

HYDROCARBON PROCESSING / APRIL 1995 3


Table 1. Heat and material balance data
Dates
Contactor Table 2. Troubleshooting guide for failure to meet
Feed gas flowrate, MMscfd specification
Feed gas temperature, F
Feed gas pressure, psig Acid gas concentrations in inlet gas have changed
Feed gas H2S content, mol% Feed gas temperatures have changed
Feed gas CO2 content, mol% >120F for H2S
Treated sales gas flowrate, MMscfd <75F for CO2
Treated sales gas temperature, F Lean amine solution is too hot
Treated sales gas pressure, psig Fouled cross exchanger and/or lean amine cooler
Treated sales gas H2S content, ppm Undersized cross exchanger and/or lean amine cooler
Treated sales gas CO2 content, mol% High ambient temperature
Differential pressure, psig Amine concentration too high or too low
Lean solvent flowrate, gpm Lack of equilibrium driving force (too high)
Lean solvent temperature, F Inadequate moles of amine for moles of CO2 and/or H2S (too low)
Lean solution concentration, wt% Inadequate lean loading
Rich solvent temperature, F Poorly stripped solvent
Flash tank Insufficient stripping steam in stripper
L.P. contractor solvent flowrate, gpm Low/high stripper pressure
L.P. contractor solvent temperature, F Foaming in stripper
Bulk temperature, F Contaminated solvent
Bulk pressure, psig Circulation rate below minimum turndown
Loss of trays or plugged packing
Lean/rich cross-exchanger Other sulfur species causing high partial pressure
Rich solvent in, F Leak in lean/rich cross exchanger
Rich solvent out, F Amine feed tray too low in absorber
Lean solvent in, F Not enough stages for required removal
Lean solvent out, F Loss of contact trays/plugged packing in absorber
Exchanger approach temperature, F No stages for removal
Regenerator Poor distribution/lack of contact between gas and amine
Bottom temperature, F Amine circulation rate too low
Bottom pressure, psig Not enough moles of amine for moles of CO2 and/or H2S
Overhead temperature, F Gas pressure has dropped
Overhead pressure, psig Driving force for removal reduced
Lean solution H2S loading, mol/mol Contactor operating below minimum turndown
Lean solution CO2 loading, mol/mol Poor hydraulics/loss of efficiency
Steam flowrate, lb/hr Maximum gas flow exceeded
Steam pressure, psig Jet flooding/poor hydraulics/loss of efficiency
Reboiler heat medium flowrate, gpm Degraded solvent (heat stable salt buildup)
Reboiler heat medium density, lb/gal Less moles of active amine available for acid gas removal
Reboiler heat medium specific heat, Btu/lb F Oxygen in feed gas
Direct-fired reboiler fuel feed rate, MMscfd Thermal degradation
Direct-fired reboiler efficiency,% Carboxylic acids in feed gas
Reflux section Mechanical damage
Reflux condenser inlet temperature, F Collapsed trays/broken packing
Reflux condenser outlet temperature, F Plugged and/or broken distributors
Reflux accumulator temperature, F Corrosion
Reflux accumulator pressure, psig Contaminated/dirty solvent
Reflux accumulator return to regenerator, gpm Well treating chemicals
Acid gas temperature, F FeS
Acid gas pressure, psig Corrosion products
Barometric pressure, psia Foaming (see Table 3 for a more detailed outline)

 Amine losses ately before they become a serious problem.


 Corrosion. As suggested in Table 2, key process parameters that
The first two categories, failure to meet specification should be monitored and graphed are:
and foaming, are normally immediate concerns requiring Lean amine temperature and concentration
quick action. The other categories, amine losses and cor- Lean H2S and/or CO2 loading
rosion, are long term in their manifestation. Amine circulation rate
Stripper overhead temperature
Failure to meet specification. Failure to meet the Absorber P
treated gas specifications can be the result of a single Inlet-gas flowrate
factor or several factors working in conjunction. These Inlet-gas H2S and/or CO2 concentration
factors are outlined in Table 2. Generally, final identifi- Treated gas composition.
cation of the cause(s) is a process of elimination. The A prior knowledge of anticipated conditions, operat-
cause(s) can be identified more quickly and corrected the ing ranges and their effects on treated gas specifications
sooner the possible contributing factors can be ruled out. are helpful in setting control points and appropriately
The problem can sometimes be avoided altogether by reacting to changes as observed on control charts. A proac-
monitoring key process parameters that could result in tive approach to plant operations is always better than a
off-specification treated gas and responding appropri- reactive one.

4 HYDROCARBON PROCESSING / APRIL 1995


ting specific raw data against time.
5 3.0 The only difference when creating
50 wt% MDEA @ 110 F control charts is that these data
500 psia absorber pressure
4 should be plotted on the graph papers
Treated gas, ppm H2S

top half instead of the entire page. As

Treated gas, ppm H2S


2.5 previously indicated, plotting raw
3
data versus time is a run chart.
Once constructed (Fig. 8), the run
2 chart should be examined for any
2.0 trends, cycles or spikes. Trends or
1 cycles may reflect recent changes to
50 wt% MDEA @ 0.003 process operating conditions or
mole/mole lean loading
0 500 psia absorber pressure
repeatable events such as daytime
0 0.001 0.002 0.003 0.004 0.005 0.006 1.5 heating cycles. Spikes may reflect bad
95 100 105 110 115 120 125 130
Lean H2S loading, mol/mol
Lean amine temperature, F
data points, but they may also corre-
spond to conditions that are quite far
Fig. 2. H2S in treated gas vs. lean amine Fig. 3. H2S in treated gas vs. lean amine tem- away from the normal set of oper-
loading. perature.
ating conditions. The run chart is
Example: A plant is designed to treat 15 MMscfd of intended to provide a feel for how the system is operating,
gas at 500 psia, 100F, and contains 3% CO2 and 5% H2S. but it does not provide any guidelines for when the data
The product has <4 ppm H2S and 2% maximum CO2. The should be acted upon.
plant operates with a 50 wt% MDEA solution circulat- Step 2: range charts. This will help the operator and
ing at 100 gpm. In this case, knowledge of trends that engineer determine whether the process is in control and
would be useful in setting up control charts and ranges whether further action can be taken with the raw data
might include but are not limited to: run chart. It is important to collect a minimum of 20 data
 H2S in treated gas vs. lean H2S loading (Fig. 2) points, and preferably 50 data points, before proceeding.
 H2S in treated gas vs. lean amine temp. (Fig. 3) First, the range between individual data points is
 Stripper overhead temp. vs. reflux ratio (Fig. 4) obtained by calculating the absolute difference between
 Lean H2S loading vs. stripper reflux ratio (Fig. 5) consecutive raw data points. This will result in a table
 Turnup and turndown characteristics at different of data with one less data point than the raw data. This
circulation rates (Fig. 6). is called moving range data or MR data. These are plot-
Knowing directionally which way the plant will ted in a second graph using the bottom half of the same
respond under different conditions allows operators to chart paper where the run chart was plotted. The first
respond quickly and appropriately to upsets or chang- moving range data point will correspond to the second
ing conditions. Similar curves can easily be generated time increment since there is no moving range data point
for different amines and process conditions. available for the first time increment.
Fig. 7 is an example of a run chart that has raw data Once the MR data have been plotted, a center line and
for several of the critical parameters plotted versus time. control limits must be added. The center line is the MR
These data show that the plant is performing adequately data average value. By definition, there is no lower con-
and meeting specifications until the H2S goes off speci- trol limit because this is a moving range chart. There is,
fication at 4 a.m. Without creating a run chart, an oper- however, an upper control limit. This value is calculated
ator would not realize the plant was heading for trouble. by multiplying the moving range average by 3.27. If the
When the sales gas finally went off spec, the operator process is under control, 99.7% of all points measured
would be in a panic situation. However, by plotting will fall below the upper control limit (UCLmr). This
these data, it is obvious that the H2S
starts to creep up after 5 p.m. The run
chart also shows that there is an 3.2 3.2
increase in lean amine H2S loading 3.0 3.0
Reflux ratio, mol H2O/mol acid gas

8 psig stripper
Reflux ratio, mol H2O/mol acid gas

occurring at the same time. This is 8 psig stripper


2.8 overhead pressure overhead pressure
2.8
probably due to a stripper problem as 2.6 2.6
indicated by a corresponding decrease
2.4 2.4
in stripper reflux ratio.
2.2 2.2
Although run charts show raw
2.0
data as a function of time, they still 2.0
do not tell the operators when to start 1.8 1.8
reacting. If sufficient data is collected, 1.6 1.6
these run charts can easily be trans- 1.4 1.4
lated into control charts. 1.2 1.2
1.0 1.0
Control charting. 200 204 208 212 216 220 0 0.001 0.002 0.003 0.004 0.005 0.006
Stripper overhead temperature, F Lean amine H2S loading, mol/mol
Step 1: run charts. The initial
stage in developing a control chart Fig. 4. Stripper overhead temperature vs. Fig. 5. Lean amine H2S loading vs. stripper
has already been illustrated by plot- reflux ratio. reflux ratio.

HYDROCARBON PROCESSING / APRIL 1995 5


graph is known as a moving range chart or MR-chart
40
(Fig. 9).
36-in contactor with 40,60, and 100 gpm on MR-charts tell something about the process stability
35 130 gpm trays, 100 F inlet gas, 0.65 sp. gr. and are analyzed using two simple rules. The process
is in control if all data points fall below UCLmr and
30
the Rules of Seven are not violated. The Rules of
Seven are that no seven consecutive points are above or
Inlet gas flowrate, MMscfd

below the center line, trending upward or trending


25 downward.
If any data points lie above the UCLmr, the operator
20 or engineer must check whether this point was properly
Turndown limit measured or whether the measuring device was work-
Turnup limit
Turndown limit ing properly. If measured properly, then the operator or
15
Turnup limit engineer must determine the plant operating conditions
Turndown limit
Turnup limit during the sampling, if possible. If the plant operated
10 better around the sampling time, then the operators and
engineers must decide whether the plant should be oper-
5
ated at these conditions as a matter of course. If the plant
operated poorly around the sample point then the plant
should not be operated at this condition again. This avoids
0
0 100 200 300 400 500 600 700 800 900 1,000 an out-of-control process.
Inlet gas pressure, psig Step 3: variables control chart. For this step, the
MR-chart must be in control as previously defined. Fig.
Fig. 6. Turnup/turndown vs. inlet pressure. 9 shows that the MR-chart is indeed in control. There
appears to be some cycling at the later times. This may be
due to daytime heating interrupted by a rain shower, but
6
the cycles do not violate the Rules of Seven.
Outlet H2S, ppm Since the MR-chart is under control, the run chart can
5 Lean flow, gpm/100
Lean temperature, F/100 now be revised. By adding a central line as well as an
Lean loading, m/m  1,000 upper and a lower control limit, this run chart can be
4 Reflux ratio, mol/mol
translated into a variables control chart, also called an X-
Parameter

chart (Fig. 10). To calculate the central line, the raw data
3
is averaged. The upper and lower control limits are then
calculated in two steps.
2
First, the estimated standard deviation is calculated
by dividing the moving range average value by 2.128.
1
This standard deviation estimate is then multiplied by 3.
The resulting product is added to and subtracted from
0
5:00 9:00 13:00 17:00 21:00 1:00 5:00 9:00 the central line value to become the X-chart upper control
7:00 11:00 15:00 19:00 23:00 3:00 7:00 limit (UCLx) and X-chart lower control limit (LCLx).
Actual time, hr
Again, 99.7% of the data points will fall between these
Fig. 7. Raw data run chart. limits if the data is in control.
The Rules of Seven are then applied to the X-chart.
Furthermore, the data points must fall between the two
2.5 control limits to be considered in control. Fig. 10 satis-
2.4
fies all these criteria, so the data are in control. An exam-
ple of what the overall X-MR control chart would look
2.3 like for this data set is in Fig. 11.
2.2
Since these data represent stable plant operation, con-
trol limits and center lines will be used as new data is
2.1 collected. This way the new data will be compared with a
ppm H2S

data set generated when the system was in control. As the


2.0
operators and engineers advance to the next level of con-
1.9 trol charting, they will learn when it is appropriate to
recalculate the center lines and control limits. This will
1.8
not be discussed here.
1.7 Fig. 12 has some new data added to the X-chart. The
H2S outlet concentration gradually increases overnight.
1.6
Incremental changes in the outlet H2S concentration
1.5 are such that no indication is given that the process is
5:00 7:00 9:00 11:00 13:00 15:00 17:00
6:00 8:00 10:00 12:00 14:00 16:00
out of control. The variability between individual data
Actual time, hr points is not large enough nor continuous enough to
trigger an out-of-control point or a trend on the extended
Fig. 8. Run chart. MR-chart (Fig. 13).

6 HYDROCARBON PROCESSING / APRIL 1995


Under normal operations, it is conceivable to have the
operator not respond to this gradual change until the 0.30
H2S has nearly reached the treated pipeline specifica-
Moving range values
tion of 4 ppm. At this point, the activity becomes a reac- AVGmr
tive exercise, not a proactive one. By using the control 0.25 UCLmr
chart limits and the Rules of Seven, the operator would
have responded within two hours of the H2S going out
0.20
of control. Fig. 14 shows the control chart sensitivity.
After 17:45, there are seven consecutive data points above

Moving range
the center line. Also, there are seven data points trend-
0.15
ing upwards after 18:30. Either of these violations would
trigger an operator response.
By recognizing a potential problem early enough,
0.10
the operator has time to collect and review other data.
The operator would look at all the run and control
charts and check for positive or negative correlations.
0.05
Fig. 15 shows a combined run chart of six critical oper-
ating parameters. There is a positive correlation
between outlet H2S content and differential pressure 0.00
across the contactor. By a careful data review, the usual 5:00 7:00 9:00 11:00 13:00 15:00 17:00
6:00 8:00 10:00 12:00 14:00 16:00
solution of increasing the reboiler temperature can be Actual time, hr
avoided. This is because the run chart data indicates the
problem is in the systems front end and not the sys- Fig. 9. Moving range control chart (MR-chart).
tems regeneration section. Increasing heat to the
reboiler may appear to temporarily solve the problem by
giving a leaner solution entering the contactor and, Table 3. Troubleshooting guide for foaming
hence, a lower outlet H2S content. But this will only
Symptoms
delay proper response to the true problemfoaming Fluctuating absorber or stripper P
in the contactor. Amine carryover from absorber, stripper or flash tank
Even though this example is simplistic, it does point Swinging liquid levels in any vessel
out the power of control charts combined with a little Decrease in H2S removal with increase in CO2 removal
Off-specification treated gas
process know-how. Positive results from foam test
Step 4: implementing control charts. This Poorly stripped solvent
methodology is not difficult to put into place. However, Causes
it takes a commitment by all members associated with Hydrocarbons + organic acids = > amine soaps
a particular gas treating facility. The process can be Poor inlet gas separation
Lean amine temperature too low
implemented only with some graph paper, a pencil and Well-treating fluids and corrosion inhibitors
a calculator. If fancy output is desired, there are various Inadequate inlet gas separation
programs that will collect data and create elegant pro- Oil, inhibitors and welding fluxes
cess control charts. Improperly cleaned equipment
Suspended solids* (FeS, carbon fines, colloidal iron oxide)
The key is to first have the people that are directly Poor inlet gas separation
involved with the process decide which data are critical. Inadequate solution filtration
Next, this same group of people will be given responsi- Raw cotton filters
bility for constructing and maintaining the control charts. Improper selection of filter material
Makeup water
It does no good to have the data go into head office, have Corrosion inhibitors from steam condensate system
it manipulated by office professionals and sent back out. Bacteria
The time delay is too great and the person who sees imme- Preventionmaintaining a clean system is the key!
diate changes in the process is not at the plant and may Maintain lean amine temperature 10F above inlet gas temperature to min-
not appreciate the significance of the change. imize hydrocarbon condensation
Adequate inlet gas separation to minimize liquid hydrocarbons, iron sul-
As more data is collected, it is most likely the facility fide/suspended particles and well-treating chemicals from entering the
will move away from X-MR charts and advance to Mbar- amine system
R charts and then to Xs charts. The major difference Proper cleanout of new equipment prior to startup
between these chart types is the amount of data the Adequate carbon and mechanical filtration
Foam test with proposed makeup or initial fill water
charts are based on, as well as how the data is manipu- Proper mechanical filter material selection
lated. We will not go into these control charts. Their pro- Temporary use of antifoams
cedures are clearly outlined in other references.16 Either silicone, polyglycol or high molecular-weight alcohols
We strongly recommend that personnel involved in Antifoam usage in most cases should be considered a temporary treat-
ment while the root cause(s) is identified and corrected
the implementation and maintenance of control charts
*Suspended solids do not cause foaming, but rather stabilize the foam.
be given the proper tools to efficiently handle chart devel-
opment. This can be through a team leader reading the
appropriate background material and summarizing the
critical issues during a couple of training sessions. Or it Benefits of using control charts will soon be seen in
can be as simple as an operator doing some background the daily operation of a facility as well as during opti-
reading and trying the charting. mization activities.

HYDROCARBON PROCESSING / APRIL 1995 7


2.5
Raw H2S data
2.4 AVGx
UCLx 2.7
2.3 LCLx
2.5 Raw H2S data UCLx
AVGx LCLx
2.3

X-chart
2.2
2.1
2.1 1.9
ppm H2S

1.7
2.0 1.5
0.35
1.9
0.30 Moving range values
AVGmr
1.8 0.25 UCLmr

MR-chart
0.20
1.7 0.15
0.10
1.6
0.05
1.5 0.00
5:00 7:00 9:00 11:00 13:00 15:00 17:00 5:00 7:00 9:00 11:00 13:00 15:00 17:00
6:00 8:00 10:00 12:00 14:00 16:00 6:00 8:00 10:00 12:00 14:00 16:00
Actual time, hr Actual time, hr

Fig. 10. Variables control chart (X-chart). Fig. 11. Control charts for variables data.

Table 4. Troubleshooting guide for losses


Mechanical
Both gas/liquid & liquid/liquid applications:
ANY permanent, unwanted or unscheduled physical removal of solvent from the plant
Leaks at pipe flange/gasket connections Pump seal flushes or leaks
Pressure gauge/sample line purges Frequent filter changes
Unrinsed cartridge filter elements Unrinsed carbon bed
Amine sump contamination/disposal Excessive daily sample volumes
Heat exchanger leaks
Entrainment
Gas/liquid applications only:
Unwanted or uncontrolled liquid in gas dispersions (mists/sprays) formed
Tower diameter too small for gas flow Tower pressure below design
Trays at or above flooding Trays damaged or plugged
Amine distributors undersized or plugged Damaged or missing mist eliminator pads
Damaged knock-out vessels
Unwanted or uncontrolled gas in liquid dispersions (foams) formed
Refer to Table 3
Liquid/liquid applications only:
Unwanted or uncontrolled liquid in liquid dispersions (emulsions) formed
Fluid velocity greater than 15 gpm/ft2 Packing material not steel or ceramic construction
Amine distributor orifice velocity > 170 ft/min Amine superficial velocity > 60 ft/hr
Hydrocarbon superficial velocity > 130 ft/hr Damaged tower internals
H.C. disperser orifice velocity > 1.25 ft/sec
Solubility
Liquid/liquid applications only:
Amine concentrations greater than 40 wt% High system temperatures
Low system pressures MEA > DEA > MDEA
Vaporization
Both gas/liquid & liquid/liquid applications:
High amine concentrations High system temperatures
Low system pressures MEA > DGA > DIPA > MDEA > DEA
Degradation
Both gas/liquid & liquid/liquid applications:
Chemical degradation
Reaction with carbon dioxide (MEA & DEA only)
Reaction with oxygen (all amines)
Thermal decomposition
Skin temperatures greater than 350 F
Heat stable salt formation
Inlet gas contaminants
Oxygen degradation
Thermal degradation byproducts

8 HYDROCARBON PROCESSING / APRIL 1995


4.5

0.30

4.0 Moving range values


AVGmr
0.25 UCLmr
Raw H2S data
AVGx
3.5 UCLx
LCLx 0.20
H2S sales spec
ppm H2S

Moving range
3.0
0.15

2.5
0.10

2.0
0.05

1.5 0.00
5:00 9:00 13:00 17:00 21:00 1:00 5:00 9:00 5:00 9:00 13:00 17:00 21:00 1:00 5:00 9:00
7:00 11:00 15:00 19:00 23:00 3:00 7:00 7:00 11:00 15:00 19:00 23:00 3:00 7:00
Actual time, hr Actual time, hr

Fig. 12. Extended X-chart with new data. Fig. 13. Moving range control chart (MR-chart).

Foaming. Table 3 lists some common symptoms of treating plants can, however, be controlled and mini-
foaming along with likely causes. Foaming is the most mized with proper plant design and operating param-
common cause of upsets in an amine plant, usually eters. The key is to minimize acid-gas flashing and to
resulting in excessive amine losses, off-specification understand other contributing factors that can enhance
treated gas and a reduction in treating capacity. All corrosivity of amine solutions. Like the three previ-
of these affect operating costs and revenue. Monitor- ous categories, monitoring and an understanding of
ing key parameters such as absorber or stripper pres- key process parameters can help minimize corrosion in
sure drop, vessel liquid levels and treated gas outlet amine units. Listed in Table 5 is a troubleshooting
composition can alert the operators to potential foam- guide for corrosion based on a method from DuPart,
ing problems. Bacon and Edwards.8
Fig. 14 shows the use of a combination of run charts
to alert the operator to a foaming problem by moni- OPTIMIZATION
toring absorber P as well as outlet H 2 S content. It is almost imperative that amine plant operations
Again, not only is it important to monitor key process be optimized to the maximum. It is no longer acceptable
parameters, but to understand what they are indi- to operate plants with huge safety margins on circula-
cating so appropriate corrective or preventative action tion rate and energy input. Plants are required to func-
can be taken. tion closer to the edge of the operating envelope. This
Operating philosophy when dealing with foaming new operating discipline can only be achieved with an
should be to address the root cause as opposed to con- understanding of the operating envelope and penalties
tinually treating symptoms. Using antifoams for tempo- associated with huge safety margins.
rary relief is acceptable. But for long-term, more reliable To successfully optimize an amine unit and main-
operation, identifying and correcting the root cause(s) tain an optimum process, a good baseline data set and
should be the norm. an understanding of, and commitment to, optimized
target conditions is required. If one or the other is
Amine losses. Every amine plant can expect to have missing, the optimization will be haphazard or will
some amine losses associated with its operation. Accept- not occur. Again, continuous gathering of plant data,
able losses vary with different applications. However, graphical representation and an understanding of the
excessive amine losses can have a negative economic cause-and-effect relationship between different pro-
impact on total plant operating cost, and it is important cess parameters is a key aspect of successful plant
to understand how losses can occur and the appropriate optimization.
corrective action to take. Losses usually fall into five Optimization of a plant entails the correct balance
major categories: mechanical, entrainment, solubility, between capacity, energy and corrosion. Ideally, acid gas
vaporization and degradation. Listed in Table 4 is a trou- in the treated-gas stream should be slightly less than
bleshooting guide for amine losses based on Stewart and the maximum specified limit, heat input to the reboiler
Lannings method.7 should be at a minimum and corrosion should be within
tolerable and accepted limits.
Corrosion. This can result in unscheduled downtime, An important, but frequently overlooked, aspect of
loss of production, reduction in equipment life and plant optimization is to get everybody involved in the
even injury or death. Corrosion in alkanolamine gas process. Teamwork and buy-in are very important and

HYDROCARBON PROCESSING / APRIL 1995 9


Table 5. Troubleshooting guide for corrosion
Raw H2S data Types of corrosion
2.4 AVGx
UCLx General corrosion
LCLx Uniform material deterioration
Galvanic corrosion
Dissimilar metallurgy coupled together in the presence of an electrolyte
2.2
Crevice corrosion
Concentration cells are allowed to form in stagnant areas of the system
Pitting corrosion
ppm H2S

Stagnant areas with high halide ion content accelerate this condition
2.0
Intergranular corrosion
Most often found in stainless steels in heat affected zones
Selective leaching corrosion
Removal of one element from a metallic alloy
1.8
Erosion corrosion
High flowrates and turbulence can accelerate attack on a metal
Aggravated by the presence of solid particles or entrained gas bubbles
Stress corrosion cracking
1.6 Simultaneous presence of tensile stress and corrosive media
Presence of chloride ion can induce this phenomenon
Hydrogen damage corrosion
14:00 15:30 17:00 18:30 20:00 21:30 23:00 Mechanical damage caused by the presence of, or the interaction of the
Actual time, hr metal with, hydrogen
Symptoms
Fig. 14. Detailed look at extended X-chart. High solution metals (Fe, Cr, Ni, etc.) content Vessel failure
Wall thinning using ultrasonic monitoring High rates on coupons
6
or probes
Physical leaks Positive solution corro-
ppm H2S sivity test
5 Lean gpm/100 Process considerations
Lean temp/100
Lean load  1,000 Rich acid gas loading too high due to low amine concentrations or under-
Reflux ratio circulation of amine
4 Delta pressure There is potential for corrosion at total acid gas loadings greater than:
0.35 mols acid gas/mol MEA
Parameter

0.40 mols acid gas/mol DEA


3 0.45 mols acid gas/mol MDEA
Residual lean amine acid loadings too high due to insufficient stripping.
Lean side acid gas flashing can occur if lean loadings exceed:
2 0.15 mols acid gas/mol MEA
0.07 mols acid gas/mol DEA
0.01 mols acid gas/mol MDEA
1
Poor quality makeup water adds contaminants to amine system that en-
hance corrosion
0 Water quality to amine plants should not exceed:
5:00 9:00 13:00 17:00 21:00 1:00 5:00 9:00 100 ppm for total dissolved solids
7:00 11:00 15:00 19:00 23:00 3:00 7:00 50 ppm for total hardness as calcium ion
Actual time, hr 2 ppm for chloride ion
3 ppm for sodium ion
Fig. 15. Combined run charts. 3 ppm for potassium ion
10 ppm for dissolved iron
Presence of heat stable salts will enhance corrosion
will contribute to short- and long-term success. The steps Incorrect amine chosen for the intended use
necessary in this concept are: Improper metallurgy in all or part of the system
1. Gather and examine baseline data set. Solution flow velocities exceeding 3 ft/sec. in carbon steel and 8 ft/sec. in
2. Generate the appropriate control charts. stainless steel
3. Identify potential areas for improvement. Large step changes, either increases or decreases, in piping internal
4. Set target process parameters. diameters
5. Adjust target process parameters. Short radius elbows
6. Monitor changes in process as result of adjustments. Flange gasket material protruding into pipe flow area causing a restriction
or turbulence
7. Determine leading edge process parameters. High suction pressure drop across pumps
8. Update the control charts. Insufficient available net pump suction head
9. Calculate cost savings and communicate to everyone. Control valves improperly located
Example: A plant originally designed for 50 MMscfd Inadequate mechanical and/or carbon filtration
is only treating 30 MMscfd due to declining flows from the Non-stress relieved or insufficiently stress relieved hardware being used
field. Design and baseline data is in Table 6. Inlet gas knock-out underdesigned, allowing contaminants to enter amine
Potential areas for improvement in this case include: solution
Reducing circulation rate Suction side leaks and no pad gas on tanks, sumps, etc., thereby allow-
Increasing amine concentration ing oxygen into the system
Lowering stripper reflux ratio Bulk solution temperatures exceeding 260F
Reducing overall reboiler duty.

10 HYDROCARBON PROCESSING / APRIL 1995


values were determined by using this
Table 6. Data for optimization example
four-step process:
Design Baseline Optimized 1. Ratio the current gas rate to
Inlet gas flow, MMscfd 50 30 30 design gas rates.
Inlet gas pressure, psia 1,000 850 850
2. Multiply this gas ratio with
Inlet gas temperature, F 100 100 100
Inlet H2S, mol% 7.0 7.0 7.0 design circulation to get the target
Inlet CO2, mol% 3.0 3.0 3.0 lean-solvent circulation rate.
CO2 specification, mol% 2.0 max 2.0 max 2.0 max 3. Ratio this target circulation rate
CO2 in treated gas, mol% 1.47 1.13 1.51 to the design circulation rate.
H2S specification, ppm <4 <4 <4 4. Multiply the solvent circulation
H2S in treated gas, ppm 2.40 0.54 2.90
MDEA, wt% 50.0 40.0 50.0 ratio by the design reboiler duty to
Amine circulation, gpm 440 350 264 get the target reboiler duty.
Lean H2S loading, mol/mol 0.005 0.002 0.005 The final optimized operating vari-
Rich H2S loading, mol/mol 0.409 0.385 0.409 ables are in Table 6.
Lean CO2 loading, mol/mol 0.005 0.002 0.005 A review of optimized process
Rich CO2 loading, mol/mol 0.101 0.110 0.098
Lean amine temperature, F 110 110 110
parameters compared to the baseline
Rich amine temperature, F 150 141 149 conditions shows several immediate
Rich amine entering stripper, F 217 217 226 savings and benefits:
Stripper overhead temperature, F 206 213 206 Reduced amine pumping require-
Stripper reflux ratio, mol/mol 1.20 2.00 1.20 ments resulting in lower power or
Stripper bottoms temperature, F 247 245 247
steam consumption depending on the
Stripper bottoms pressure, psig 10.0 10.0 10.0
Inlet temperature to LA cooler, F 178 167 168 type of pump
Reboiler duty, MMBtu/h 24.93 20.94 13.83 Reduced air and/or water require-
Cross-exchanger duty, MMBtu/h 13.06 11.99 8.91 ments for the lean amine cooler and
Lean amine cooler duty, MMBtu/h 12.71 8.83 6.55 acid gas condenser
Acid gas condenser duty, MMBtu/h 10.54 11.09 6.29 Increased CO2 slip resulting in a
higher concentration of H2S to down-
stream sulfur recovery equipment
Typically, a reduction in circulation rate and heat input Reboiler energy savings of $74,655/yr, based on a
to the reboiler are the obvious areas for improvement. How- 350-day operating year and $1.25/MMBtu.
ever, some plants have been optimized by increasing cir- When a systematic approach is taken to amine plant
culation rate to minimize rich acid-gas loadings and cor- optimization, the results are usually immediate and sus-
rosion associated with acid-gas flashing in the rich amine. tainable money savers.
Optimization is not limited to adjustment of process
LITERATURE CITED
parameters. Optimization could include installation of 1 Shewart, W., Economic control of quality manufactured product, Milwaukee, Wisconsin,
an online, lean-amine-loading analyzer, reconfiguration ASQC Quality Press, 1932, 1980.
2 Ishikawa, K., Guide to quality control, Tokyo, Japan, Nordica International Ltd., 1968.
or installation of mechanical and carbon filtration and 3 ASTM manual on presentation of data and control chart analysis, Special Technical
using a different amine. Pub. 15D, American Society for Testing and Materials, Philadelphia, Pennsylvania, 1976.
4 Juran, J. M., Quality control handbook, 4th Ed., McGraw-Hill, New York, 1988.
The sequence that should be followed when trying to 5 Krishnamoorthi, K. S., Quality control for operators & foremen, Milwaukee, Wisconsin,
reduce circulation rate and plant energy input is to first 6
ASQC Quality Press, 1989.
Montgomery, D. C., Introduction to statistical quality control, 2nd Ed., John Wiley
reduce circulation rate while maintaining heat input into & Sons, New York, 1991.
7 Stewart, E. J., and R. A. Lanning, Reduce amine plant solvent losses, Hydrocarbon
the reboiler. This will reduce the sensible heat load in the Processing, May June 1994.
stripper and result in a higher stripper reflux rate with 8 DuPart, M. S., T. R. Bacon and D. J. Edwards, Understanding corrosion in alkanolamine
gas treating plants, Hydrocarbon Processing, AprilMay 1993.
about the same or lower lean acid-gas loadings. This will
keep treated gas at current levels. Process parameters to
watch when reducing circulation rate include rich amine The authors
temperature, treated-gas temperature, rich amine load- Raymond G. F. Abry has spent the last 17 years
ings and treated gas H2S and/or CO2 composition. Depend- in various nuclear, petroleum and process chem-
istry research activities. He is currently a devel-
ing on the amine used, there are guidelines for maximum opment leader for The Dow Chemical Co. and is
rich loadings and rich amine temperatures (Table 5). The responsible for providing technical service sup-
plant should not be allowed to go off specification. Treated- port to GAS/SPEC accounts in Canada as well as
gas temperature increases will result in a higher load on supporting global development activities. Mr.
Abry received a BSc degree in chemistry from the
downstream dehydration equipment. After the optimum University of Victoria in 1976 and a Geological Certificate from the
circulation rate has been determined, heat to the reboiler Southern Alberta Institute of Technology in 1983. He is a member of
can be reduced while closely monitoring treated H2S and/or the CIM, SPE and CGPSA.
CO2 in conjunction with lean amine loadings to make sure Mike DuPart is a development leader in the
they do not exceed specifications and maximum recom- GAS/SPEC Technology Group of The Dow Chem-
mended guidelines as outlined in Table 5. ical Co. He has a wide range of experience in gas
Optimum reboiler heat input is established when the processing from corrosion inhibitor research to
process optimization of amine plants. He holds
maximum lean loading is exceeded or the treated outlet four U.S. patents and has coauthored several
composition is bumping the specification. papers on gas processing. Mr. DuPart holds a BS
Target values for this example are 264 gpm and 14.96 degree in metallurgical engineering from the Uni-
MMBtu/h heat input to the reboiler. These two target versity of Texas at El Paso.

HYDROCARBON PROCESSING / APRIL 1995 11


INEOS LLC
Head Office
2925 Briarpark Drive, Suite 870, Houston, TX 77042
713.243.6200 main 866.865.4747 customer service
www.ineosllc.com

NOTICE: No freedom from any patent owned by Seller or others is to be inferred. Because use conditions and applicable laws may differ from one location to another and may change
with time, Customer is responsible for determining whether products and the information in this document are appropriate for Customers use and for ensuring that Customers workplace
and disposal practices are in compliance with applicable laws and other governmental enactments. Seller assumes no obligation or liability for the information in this document.
Products, Technology and Service from INEOS NO WARRANTIES ARE GIVEN; ALL IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE ARE EXPRESSLY EXCLUDED.

GSMR002-8/01

You might also like