Unformatted text preview: Water Management
Water
in Refineries M. Colin Arnold
M.
1. Image  http://www.ge.com/research/grc_2_6.html & Joshi Samuel
Joshi Overview
Overview
Background
Background
Project Goals
Project
Unit Operations
Unit
Water Treatments
Water
Results
Results
Conclusion
Conclusion 1. http://pubs.usgs.gov/circ/2004/circ1268/htdocs/textin.html Background
Background
Some water uses in refinery 1 Caustic treatment
Caustic
Distillation
Distillation
Sweetening
Sweetening
Desalting
Desalting 1. http://pubs.usgs.gov/circ/2004/circ1268/htdocs/textin.html Background
Background
Traditionally
• Only fresh water feed sources
• No recycle
• Collected into a sink
• Disposed after clean up 1. Koppol, A.P., et al. Adv. in Env. Res., V(8), 2003, 151171. Background
Recently
• Water reuse
• Minimal or zero discharge
• Minimizing Cost Reasons
Reasons
• Stricter EPA regulations
• Water scarcity
• Purchase Cost 1.
2. Koppol, A.P., et al. Adv. in Env. Res., V(8), 2003, 151171.
Image  http://en.wikipedia.org/wiki/Image:Water_pollution.jpg Project Goals
Project
Reduce
1. Fresh water intake
2. Total operation cost
• Optimizing Waste water treatment • Minimizing Total discharge • Maximizing Water reuse Project Goals
Project
• In other words, we want to change this: Project Goals
Project
• In other words, we want to change this: Project Goals
Project
• In other words, we want to change this:
to this: Unit1
Unit2
Unit3 Unit Operations
Six Units:
Six
1:: Caustic Treating
1
2:: Distillation
2
3:: Amine Sweetening
3
4:: Merox I Sweetening
4
5: Hydrotreating
Hydrotreating
6:: Desalting
6 Unit Operations
Unit
Six Units:
1:: Caustic Treating
1 Figure  http://www.pall.com/chemical_5582.asp Unit Operations
Unit
Six Units:
2:: Distillation
2 1. http://www.cluin.org/download/toolkit/petrefsn.pdf Unit Operations
Six Units:
3:: Amine Sweetening
3
4:: Merox I Sweetening
4
Amine or
Merox I Refined Fuel Steam Amine or Merox 1
With Absorbed Acid gases
Absorption Unit Sour Feed Sweetening PFD
1  http://www.newpointgas.com/amine_treating.php Steam Stripping Unit Amine or
Merox I Unit Operations
Unit
Six Units:
5: Hydrotreating
Hydrotreating 1 http://www.hghouston.com/refining.html Unit Operations
Unit
Six Units:
6:: Desalting
6 1 http://www.hghouston.com/refining.html Unit Operations
Unit
Process Contaminant Salts Each unit has an
Each
inherent values
inherent 300 Cout,max
(ppm)
500 Mass Load
(kg/h)
0.18 50 500 1.2 H2S 5000 11000 0.75 Ammonia (1) Caustic Treating Organics Cin,max
(ppm) 1500 3000 0.1 10 200 3.61 1 4000 100 Salts
(2) Distillation Organics 0 500 0.25 Ammonia • Cin,max H2S 0 1000 0.8
0.6 Salts • Cout,max 1000 1 3500 30 H2S 0 2000 1.5 Ammonia (3) Amine Sweetening 10 Organics 0 3500 1 Salts • Mass Load (4) Merox I Sweetening 100 400 2 Organics 200 6000 60 50 2000 0.8 1000 3500 1
3.8 H2S
Ammonia
Salts 350 200 1800 45 H2S 300 6500 1.1 Ammonia (5) Hydrotreating 85 Organics 200 1000 2 Salts
(6) Desalter From: Koppol, A.P., et al. Adv. in
Env. Res., V(8), 2003, 151171. 1000 9500 120 Organics 1000 6500 480 H2S 150 450 1.5 Ammonia 200 400 0 Unit Operations
Unit
Each unit has an
Each
inherent values
inherent FPROCESS FPROCESS Process Stream Process Stream • Cin,max
• Cout,max FWATER FWATER Extraction Stream Extraction Stream • Mass Load Cin, max Cout, max Water Treatment
Water Tables  Cartwright, Peter. Process Water Treatment – Challenges and Solutions. Chemical Engineering Magazine. March 2006. Water Treatment
Water
Three options for treatment1
1. Option 1: API separator followed by ACA 2. Option 2: Reverse Osmosis Treatment 3. Option 3: Chevron waste water treatment 1. Koppol, A.P., et al. Adv. in Env. Res., V(8), 2003, 151171. Water Treatment
Water
[2] Option 1: API separator
Option
followed by ACA
followed
Reduces Organics to 50 ppm
ppm
$0.12 per ton
$0.12 1.
2. http://www.monroeenvironmental.com/clarifierapiseparator.htm
Koppol, A.P., et al. Adv. in Env. Res., V(8), 2003, 151171. [1] Water Treatment
Water
[2] Option 1: API separator
Option
followed by ACA
followed Figures  Cartwright, Peter. Process Water Treatment – Challenges and Solutions. Chemical Engineering Magazine. March 2006. J= − kA ∆P
× µ δ Water Treatment
Water
Option 2: Reverse Osmosis
Option
Treatment
Treatment [2] Reduces Salts to 20 ppm
ppm
$0.56 per ton
$0.56
[3] 1.
2. http://www.aquatechnology.net/commercialro.html
http://ag.arizona.edu/region9wq/pdf/nv_ROhow.pdf J= − kA ∆P
× µ δ Water Treatment
Water
[2] Option 2: Reverse Osmosis
Option
Treatment
Treatment
Reduces Salts to 20 ppm
ppm
$0.56 per ton
$0.56 [3] − kA ∆P
J=
× µ 1.
2. δ http://www.aquatechnology.net/commercialro.html
http://ag.arizona.edu/region9wq/pdf/nv_ROhow.pdf where
“J”  Volumetric flux across membrane
“k” – permeability
“A” – flux area
“∆P” – Pressure drop
“ ” – viscosity
“ ” – membrane thickness Water Treatment
Water
Option 3: Chevron waste
Option
water Treatment
water
Reduces H2S
Reduces
to 5 ppm
ppm
Reduces
Ammonia to
30 ppm
ppm
$1.00 per ton
$1.00 http://www.chevron.com/products/prodserv/refiningtechnology/waste_wtr_treat_6a.shtm Unit Operations
Unit
Assumptions
1. Parallel Operation 2. Outlets from a unit may be split and
Outlets
fed to any unit
fed 3 a) Outlets can be combined, treated,
Outlets
and recycled OR
and b) Outlets can be treated separately
Outlets
and recycled
and 4. No water loss during treatment GAMS Model
GAMS
• The Backbone of the Program Set u water using units
w freshwater source
s wastewater sink
c Contaminant
Alias(u,ua); / 1*6 /
/1/
/1/
/ 1*4 /; Parameters
CFW(w) Cost of freshwater in $ per ton
/ 1 .32 /
CWW(s) Cost of wastewater treatment $ per ton
/ 1 1.68 /;
Table ConFW(w,c) Freshwater source concentration in ppm
Table ConFW(w,c Freshwater
ppm
1
2
3
4
10
0
0
0; GAMS Model
GAMS
• The Backbone of the Program
Table Cinmax(u,c)
Table Cinmax(u,c
1
2
1 300 50
2 10
1
3 10
1
4 100 200
5 85
200
6 1000 1000 maximum inlet concentration in the units in ppm
maximum
ppm
3
4
5000 1500
0
0
0
0
50
1000
300 200
150 200; Table Coutmax(u,c) maximum outlet concentration in the units in ppm
Table Coutmax(u,c maximum
ppm
1
2
3
4
1 500 500 11000 3000
2 200 4000 500
1000
3 1000 3500 2000 3500
4 400 6000 2000 3500
5 350 1800 6500 1000
6 9500 6500 450
400;
Table ConWW(s,c) Concentration limits at the sink in ppm
Table ConWW(s,c Concentration
ppm
1
2
3
4
1 10000000000 10000000000 10000000000 10000000000; GAMS Model
GAMS
• The Backbone of the Program
Table ML(u,c)
1
1 180
2 3610
3 600
4 2000
5 3800
6 120000 Variable
FW(w,u)
FW(w,u)
F(u,u)
F(u,u)
FS(u,s)
FS(u,s)
Cout(u,c)
Cost
Cost
Consu Mass Load (transfered to water at unit) in g per hour
2
3
4
1200
750 100
100000 250 800
30000 1500 1000
60000 800 1000
45000 1100 2000
480000 1500 0; Flowrates between freshwater sources and units in ton per hour
Flowrates
Flowrates between units in ton per hour
Flowrates
Flowrates between units and sinks in ton per hour
Flowrates
Outlet concentration in the units in ppm
ppm
Cost in $Mil per year
Cost
Consumption in ton per hour; Positive variable FW,F,FS,Cout; GAMS Model
GAMS
• The Backbone of the Program
Equations
waterbalance(u) Balance of water
Balance
iinlet(u,c)
nlet(u,c)
Limit of inlet concentration of the units
Limit
outlet(u,c)
Calculation of outlet concentration of the units
outlet(u,c)
Calculation
maxout(u,c)
Limit of outlet concentration of the units
Limit
sink(s,c)
Limit of inlet concentration of the sinks
sink(s,c)
Limit
ObjCost
Objective function that minimizes cost
ObjConsu
Objective function that minimizes cosumption
Objective
cosumption
**Linear  Starting points
**Linear
inletl(u,c)
Limit of inlet concentration of the units
Limit
outletl(u,c)
Calculation of outlet concentration of the units
Calculation
maxoutl(u,c)
Limit of outlet concentration of the units;
Limit
waterbalance(u)
inlet(u,c)
outlet(u,c)
maxout(u,c)
(u,c)
sink(s,c)
sink(s,c) ..sum(w,FW(w,u))+sum(ua,F(ua,u))=e=sum(ua,F(u,ua))+sum(s,FS(u,s));
..sum(w,FW(w,u)*ConFW(w,c))+sum(ua,F(ua,u)*Cout(ua,c))=L=(sum(ua,F(u,ua))+sum(s,FS(u,s)))*Cinmax(u,c);
..sum(w,FW(w,u)*ConFW(w,c))+sum(ua,F(ua,u)*Cout(ua,c))+ML(u,c)=E=(sum(ua,F(u,ua))+sum(s,FS(u,s)))*Cout(u,c);
..Cout(u,c)=L=Coutmax(u,c);
..Cout(u,c)=L=
..sum(u,FS(u,s)*(Cout(u,c)ConWW(s,c)))=L=0;
..sum(u,FS(u,s)*( GAMS Model
GAMS
• The Backbone of the Program
ObjCost
..Cost=e=(sum(w,sum(u,FW(w,u))*cfw(w))+sum(s,sum(u,FS(u,s))*cww(s)))*0.008760;
**This cost assumes constant operation
**The "*0.008760" term comes from 8760 hours opperated per year and divided by 1E6 to get units
**The
opperated per
of millions of dollars
of
ObjConsu
..Consu=e=sum(w,sum(u,FW(w,u)));
inletl(u,c)
outletl(u,c)
maxoutl(u,c)
(u,c) ..sum(w,FW(w,u)*ConFW(w,c))+sum(ua,F(ua,u)*Coutmax(ua,c))=L=(sum(ua,F(u,ua))+sum(s,FS(u,s)))*Cinmax(u,c);
..sum(w,FW(w,u)*ConFW(w,c))+sum(ua,F(ua,u)*Coutmax(ua,c))+ML(u,c)=E=(sum(ua,F(u,ua))+sum(s,FS(u,s)))*Coutmax(u,c);
..Cout(u,c)=e=Coutmax(u,c);
..Cout(u,c)=e= MODEL Reuse / waterbalance,inlet,outlet,maxout,sink,ObjCost,ObjConsu/;
MODEL
waterbalance,inlet,outlet,maxout,sink,ObjCost,ObjConsu
MODEL Start / waterbalance,inletl,outletl,maxoutl,ObjCost,ObjConsu/;
MODEL
waterbalance,inletl,outletl,maxoutl,ObjCost,ObjConsu
*SOLVE Start using MIP minimizing cost;
SOLVE Start using MIP minimizing consu;
SOLVE
consu
*SOLVE Reuse using MINLP minimizing cost;
SOLVE Reuse using MINLP minimizing consu;
SOLVE
consu
DISPLAY Cost.l,Consu.l,FW.l,F.l,FS.l,Cout.l;
DISPLAY Cost.l,Consu.l,FW.l,F.l,FS.l,Cout.l GAMS Model
GAMS
In other words…
• Amount of freshwater was calculated
Amount of waste water was calculated and checked to assure
a zero mass balance
zero
• The cost of fresh water and treatment was calculated as the total
cost
cost
• The program minimizes either the total cost or freshwater
The
required
required
First estimates solution by solving the problem linearly
First
Then uses a nonlinear algorithm to find a solution
Then
The initial linear guess is necessary because of nature of
nonlinear systems
nonlinear GAMS Model
GAMS
Important Equations
• waterbalance(u) ..sum(w,FW(w,u))+sum(ua,F(ua,u))=e=sum(ua,F(u,ua))+sum(s,FS(u,s)); ∑ FW w, u w + ∑ Fu , u = ∑ Fu , u + ∑ FSu , s
j uj i i uj j s Water balance around each unit
Water
“FW” Flow rate of streams from fresh water to units
Flow
“F” Flow rate of streams between units
Flow
“FS” Flow rate of streams from units to sinks
Flow
“w” Fresh water source
Fresh
“s” Waste water sink
Waste
“u”, “ui”, and “uj” Any unit
and
Any GAMS Model
GAMS
Important Equations
• inlet(u,c)
..sum(w,FW(w,u)*ConFW(w,c))+sum(ua,F(ua,u)*Cout(ua,c))=L=(sum(ua,F(u,ua))+sum(s,
FS(u,s)))*Cinmax(u,c); in
( FWw, u ×Cw, c ) + ∑ ( Fu , u ×Cu , c ) ≤ (∑ Fu , u + ∑ FSui , s ) × Cu ,max
∑
j w uj i j i uj j i s Sets the mixed inlet concentration of a contaminant less than
its allowed maximum
its
“C” Concentration
Concentration GAMS Model
GAMS
Important Equations
• outlet(u,c)
..sum(w,FW(w,u)*ConFW(w,c))+sum(ua,F(ua,u)*Cout(ua,c))+ML(u,c)=E=(sum(ua,F(u,ua
))+sum(s,FS(u,s)))*Cout(u,c); out
(FWw, u ×Cw, c) + ∑(Fu , u ×Cu , c) + MLu, c = (∑ Fu , u + ∑ FSui, s) × Cu
∑
j w uj i j i uj j i s Finds the outlet concentration of a contaminant once it has
picked up a mass load in a unit
picked
“ML” Mass Load
Mass GAMS Model
GAMS
Important Equations
• maxout(u,c)
(u,c) ..Cout(u,c)=L=Coutmax(u,c);
..Cout(u,c)=L= out
u C ≤C out , max
u Sets each outlet concentration from a unit to less than or
equal to the allowed maximum
equal GAMS Model
GAMS
Important Equations
• sink(s,c) ...sum(u,FS(u,s)*(Cout(u,c)ConWW(s,c)))=L=0;
. ∑ (F u, s ×(C u , c − C s , c )) ≤ 0 s Sets the concentration of contaminants going to a sink to not
exceed the limits allowed at the sink
exceed GAMS Model
GAMS
Important Equations
• ObjCost
..Cost=e=(sum(w,sum(u,FW(w,u))*cfw(w))+sum(s,sum(u,FS(u,s))*cww(s)))*0.0
08760; Cost = ∑ (∑ ( FWw, u × Pw) + ∑ ( FWu , s × Ps )) × 0.008760
w u s Determines the total cost of a given setup
Determines
Pw Price of purchasing fresh water
Price
Ps Price of treating waste water
Price
Multiplied by 0.008760 to convert from $/hr to $Mil/yr
Multiplied GAMS Model
GAMS
Important Equations
• ObjConsu ..Consu=e=sum(w,sum(u,FW(w,u))); Consu = ∑ ∑ FWw, u
w u Calculates the fresh water requirement of a setup
Calculates GAMS Model
• The maximum number of
streams available are
shown schematically as: GAMS Model
2.616 • And the results of the
And
GAMS model minimizing
consumption are:
consumption 25.0 8.571 8.239 24.959 49.947 Caustic
Treating
System 21.718 Distillation
System Amine
Sweetening
System MeroxI
Sweetening
System 10.345 Hydrotreating
System Desalting
System 87.273 EndofPipe
Treatment GAMS Model
• Result (all flowrates in ton/hour)
Result
flowrates Origin Destination
Fresh Water Source
Caustic Treating
Distillation
Amine Sweetening
Merox I Sweetening
Hydrotreating
Desalting
Sum to Destination Caust. Dist.
Am.Sw. M1S
Hydr.
Desalt Sink
Sum from Source
2.616
25
8.571
8.239 24.959 49.947 n/a
119.332
0
0
0
1.815
0.057
0.794
0
2.666
0
0
0
0.29
0
2.995
21.715
25
0.025
0
0
0
0
8.547
0
8.572
0
0
0
0
0
0
10.345
10.345
0.026
0
0
0
3.109
24.99
0
28.125
0
0
0
0
0
0
87.273
87.273
2.667 25 8.571 10.344 28.125 87.273 119.333 • Can compare to published results1 Origin Destination
Fresh Water Source
Caustic Treating
Distillation
Amine Sweetening
Merox I Sweetening
Hydrotreating
Desalting
Sum to Destination Caust. Dist.
Am.Sw. M1S
Hydr.
Desalt Sink
Sum from Source
2.4
25
8.571
8.388 24.445 50.518 n/a
119.322
0
0
0
1.645
0.775
0
0
2.42
0
0
0
0.312
0
2.97
21.718
25
0
0
0
0
0
8.571
0
8.571
0
0
0
0
0
0
10.345
10.345
0
0
0
0
0
25.21
0
25.21
0
0
0
0
0
0
87.269
87.269
2.4 25 8.571 10.345 25.22 87.269 119.332 • Very similar for the most part • Only significant differences are that F1,6, F3,1, F5,1, and F5,5 are all
5,5
zero in the published results
zero 1. Koppol, A.P., et al. Adv. in Env. Res., V(8), 2003, 151171. GAMS Model Problems
GAMS
• Minimum flowrates
Minimum flowrates
Very low flowrates may be physically unrealizable
flowrates
As an example, a minimum of 0.1 ton per hour is set
As • One possible solution is use of a binary marker Yi,i
One • Set so that Yi,i*Fmin ≤ Fi,i≤Yi,i*Fmax
If Fi,i ≤ Fmin, the model should automatically set Yi,i to
to
zero and reset Fi,i to zero
Maximum must be included as well so that Yi,i isn’t
always zero
Fmax is an arbitrarily large number GAMS Model Problems
GAMS
Binary Marker, Yi,i
• Using this marker works fundamentally Origin Destination
Fresh Water Source
Caustic Treating
Distillation
Amine Sweetening
Merox I Sweetening
Hydrotreating
Desalting
Sum to Destination Caust. Dist.
Am.Sw. M1S
Hydr.
Desalt Sink
Sum from Source
79.343 3.E+01
8.571
9.828
0
0 n/a
122.742
0
0
0
0 80.478
0
0
80.478
0.1
0
0
0.517
2.847
0
21.536
25
1.035
0
0
0
0
0
7.536
8.571
0
0
0
0
0
0
10.345
10.345
0
0
0
0
0.1 83.325
0
83.425
0
0
0
0
0
0
83.325
83.325
80.478 25 8.571 10.345 83.425 83.325 122.742 • However, the solver did not guarantee the solution to be the
However,
absolute optimum in this case
absolute
• Solution still is very close to previous results
Consumption = 122.761 ton/hr, Cost = $Mil 2.151/yr
Consumption GAMS Model Problems
GAMS
Binary Marker, Yi,i
• Results can be compared to published results Origin Destination
Fresh Water Source
Caustic Treating
Distillation
Amine Sweetening
Merox I Sweetening
Hydrotreating
Desalting
Sum to Destination Caust. Dist.
Am.Sw. M1S
Hydr.
Desalt Sink
Sum from Source
79.343 3.E+01
8.571
9.828
0
0 n/a
122.742
0
0
0
0 80.478
0
0
80.478
0.1
0
0
0.517
2.847
0
21.536
25
1.035
0
0
0
0
0
7.536
8.571
0
0
0
0
0
0
10.345
10.345
0
0
0
0
0.1 83.325
0
83.425
0
0
0
0
0
0
83.325
83.325
80.478 25 8.571 10.345 83.425 83.325 122.742 • Published Results: Origin Destination
Fresh Water Source
Caustic Treating
Distillation
Amine Sweetening
Merox I Sweetening
Hydrotreating
Desalting
Sum to Destination Caust. Dist.
Am.Sw. M1S
Hydr.
Desalt Sink
Sum from Source
2.4
25
8.571
8.388 24.445 50.518 n/a
119.322
0
0
0
1.645
0.775
0
0
2.42
0
0
0
0.312
0
2.97
21.718
25
0
0
0
0
0
8.571
0
8.571
0
0
0
0
0
0
10.345
10.345
0
0
0
0
0
25.21
0
25.21
0
0
0
0
0
0
87.269
87.269
2.4 25 8.571 10.345 25.22 87.269 119.332 GAMS Model Problems
GAMS
• Minimum flowrates
Minimum flowrates
Using a binary marker uses very many resources
Using
Thus, a different process may be desired
Thus,
• An alternate solution is to set individual flowrates lless than the
An
flowrates ess
minimum to zero
minimum
F3,1 = 0.025 → F3,1 = 0
0.025
• If another flowrate iis less than the minimum after doing this, it is
If
flowrate s
set to zero also
set
• This process is repeated until none are below the minimum
• Six different combinations of streams set to zero meet the
Six
minimum flowrate standards
flowrate GAMS Model Problems
2.616 • New Results (flowrates in ton per hour)
25.0 Before:
Before:
8.571 8.239 24.959 Origin Destination
Fresh Water Source
Caustic Treating
Distillation
Amine Sweetening
Merox I Sweetening
Hydrotreating
Desalting Caust. Dist.
Am.Sw. M1S
Hydr.
Desalt Sink
Sum from Source
2.616
25
8.571
8.239 24.959 49.947 n/a
119.332
0
0
0
1.815
0.057
0.794
0
2.666
0
0
0
0.29
0
2.995
21.715
25
0.025
0
0
0
0
8.547
0
8.572
0
0
0
0
0
0
10.345
10.345
0.026
0
0
0
3.109
24.99
0
28.125
0
0
0
0
0
0
87.273
87.273 • Published Results (flowrates in ton per hour) Sum to Destination 2.667 25 8.571 10.344 28.125 87.273 119.333 49.947 Caustic
Treating
System 21.718 Distillation
System Amine
Sweetening
System MeroxI
Sweetening
System 10.345 Hydrotreating
System Desalting
System 87.273 EndofPipe
Treatment GAMS Model Problems
2.616 • New Results (flowrates in ton per hour)
25.0 After:
After:
8.571 • All flowrates are now greater than the
All flowrates are
arbitrary minimum (0.1)
arbitrary 24.959 Origin Destination
Fresh Water Source
Caustic Treating
Distillation
Amine Sweetening
Merox I Sweetening
Hydrotreating
Desalting
Sum to Destination Caust. Dist.
Am.Sw. M1S
Hydr.
Desalt Sink
Sum from Source
2.4
25
8.571
8.388 24.445 50.518 n/a
119.322
0.267
0
0
1.645
0.775
0
0
2.687
0
0
0
0.312
0
2.974
21.715
25.001
0
0
0
0
0
8.571
0
8.571
0
0
0
0
0
0
10.345
10.345
0
0
0
0
2.905
25.21
0
28.115
0
0
0
0
0
0
87.273
87.273
2.667 25 8.571 10.345 8.239 28.125 87.273 119.333 • Fresh water requirements do not change
Fresh
by use of minimum flowrates
flowrates 49.947 Caustic
Treating
System 21.718 Distillation
System Amine
Sweetening
System MeroxI
Sweetening
System 10.345 Hydrotreating
System Desalting
System 87.273 EndofPipe
Treatment GAMS Model Problems
2.616 • New Results (flowrates in ton per hour)
25.0 Caustic
Treating
System 21.718 Distillation
System Origin Destination
Fresh Water Source
Caustic Treating
Distillation
Amine Sweetening
Merox I Sweetening
Hydrotreating
Desalting
Sum to Destination Caust. Dist.
Am.Sw. M1S
Hydr.
Desalt Sink
Sum from Source
2.4
25
8.571
8.388 24.445 50.518 n/a
119.322
0.267
0
0
1.645
0.775
0
0
2.687
0
0
0
0.312
0
2.974
21.715
25.001
0
0
0
0
0
8.571
0
8.571
0
0
0
0
0
0
10.345
10.345
0
0
0
0
2.905
25.21
0
28.115
0
0
0
0
0
0
87.273
87.273
2.667 25 8.571 10.345 28.125 87.273 119.333 8.571 8.239 24.959 • Published Results (flowrates in ton per hour) 49.947 Origin Destination
Fresh Water Source
Caustic Treating
Distillation
Amine Sweetening
Merox I Sweetening
Hydrotreating
Desalting
Sum to Destination Caust. Dist.
Am.Sw. M1S
Hydr.
Desalt Sink
Sum from Source
2.4
25
8.571
8.388 24.445 50.518 n/a
119.322
0
0
0
1.645
0.775
0
0
2.42
0
0
0
0.312
0
2.97
21.718
25
0
0
0
0
0
8.571
0
8.571
0
0
0
0
0
0
10.345
10.345
0
0
0
0
0
25.21
0
25.21
0
0
0
0
0
0
87.269
87.269
2.4 25 8.571 10.345 25.22 87.269 119.332 Amine
Sweetening
System MeroxI
Sweetening
System 10.345 Hydrotreating
System Desalting
System 87.273 EndofPipe
Treatment GAMS Model Robustness
GAMS
• The GAMS model can be used to test
The
other scenarios
other
• Scenario 1:
The Amine sweetening unit (unit 3) is
aging early and another has been ordered
to replace it; unfortunately, the unit will
not arrive and be in operation for another
year. A consequence of the early aging of
the unit is that it can only handle Cin,max
one fifth of its previous operation
one
capacity, the Cout,max are cut in half and the
are
mass loads triples.
mass Origin Destination
Fresh Water Source
Caustic Treating
Distillation
Amine Sweetening
Merox I Sweetening
Hydrotreating
Desalting
Sum to Destination Caust. Dist.
Am.Sw. M1S
Hydr.
Desalt Sink
Sum from Source
2.4
25
8.571
8.091 24.371 19.394 n/a
87.827
0.267
0
0
1.529
0.871
0
0
2.667
0
0
0
0
0
0
25
25
0
0
0
0
0
0
8.571
8.571
0
0
0
0
0
0
10.345
10.345
0
0
0
0.725
1.349 24.242
0.275
26.591
0
0
0
0
0
0
43.636
43.636
2.667 25 8.571 10.345 26.591 43.636 87.827 GAMS Model Robustness
GAMS
• A more valuable check of robustness
more
is to fix the streams between units as
either existent or nonexistent once an
either
existent
initial minimization is run
initial
• The mass load can then be changed to
The
test that new solutions are reasonable
test GAMS Model Robustness
GAMS
Caustic
Treating
System 33.417
0.484 • New Scenario:
Mass loads double, network fixed
Mass 25 Amine
Sweetening
System 8.571 MeroxI
Sweetening
System 9.828 8.571 10.345 83.516 83.316 122.357 Cost = $Mil 2.144 / yr
Consumption = 122.357 ton/hr
Consumption 83.316 0.2 25 45.541 33.901 33.901 3.674 Caust. Dist.
Am.Sw. M1S
Hydr.
Desalt Sink
Sum from Source
33.417
25
8.571
9.828 45.541
0 n/a
122.357
0
0
0
0 33.901
0
0
33.901
0
0
0
0.517
0.2
0
24.283
25
0.484
0
0
0
3.674
0
4.413
8.571
0
0
0
0
0
0
10.345
10.345
0
0
0
0
0.2 83.316
0
83.516
0
0
0
0
0
0
83.316
83.316 0.2 Origin Destination Sum to Destination 4.413 0.517 • Initial Results:
Fresh Water Source
Caustic Treating
Distillation
Amine Sweetening
Merox I Sweetening
Hydrotreating
Desalting 24.283 Distillation
System 10.345 Hydrotreating
System Desalting
System 83.316 EndofPipe
Treatment GAMS Model Robustness
GAMS
Caustic
Treating
System 13.8
0.2 • New Scenario:
Mass loads double, network fixed
Mass 50 Amine
Sweetening
System 17.143 MeroxI
Sweetening
System 19.655 17.143 20.689 166.85 166.65 244.671 Cost = $Mil 4.287 / yr
Consumption = 244.67 ton/hr
Consumption 166.65 14 50 144.072 0.2 14 8.378 Caust. Dist.
Am.Sw. M1S
Hydr.
Desalt Sink
Sum from Source
13.8
50 17.143 19.655 144.072
0 n/a
244.67
0
0
0
0
14
0
0
14
0
0
0
1.034
0.2
0
48.766
50
0.2
0
0
0
8.378
0
8.565
17.143
0
0
0
0
0
0
20.69
20.69
0
0
0
0
0.2 166.65
0
166.85
0
0
0
0
0
0
166.65
166.65 0.2 Origin Destination Sum to Destination 8.565 1.034 • Final Results:
Fresh Water Source
Caustic Treating
Distillation
Amine Sweetening
Merox I Sweetening
Hydrotreating
Desalting 48.766 Distillation
System 20.69 Hydrotreating
System Desalting
System 166.65 EndofPipe
Treatment GAMS Model Robustness
GAMS
• Other modes of the GAMS program can be used find
Other
solutions that:
solutions GAMS Model Robustness
GAMS
• Other modes of the GAMS program can be used find
Other
solutions that:
solutions
Minimize cost
Minimize Origin Destination
Fresh Water Source
Caustic Treating
Distillation
Amine Sweetening
Merox I Sweetening
Hydrotreating
Desalting
Sum to Destination Caust. Dist.
Am.Sw. M1S
Hydr.
Desalt Sink
Sum from Source
16.68
25
8.571
0 24.774 44.307 n/a
119.332
0
0
0
9.852
0.243
7.062
0
17.157
0
0
0
0
0
3.285
21.715
25
0
0
0
0
0
8.571
0
8.571
0
0
0
0
0
0
10.345
10.345
0.477
0
0
0.493
3.109 24.047
0
28.126
0
0
0
0
0
0
87.273
87.273
17.157 25 8.571 10.345 28.126 87.272 119.333 GAMS Model Robustness
GAMS
• Other modes of the GAMS program can be used find
Other
solutions that:
solutions
Minimize cost, then minimize the consumption with the
initial solutions fixed
initial Origin Destination
Fresh Water Source
Caustic Treating
Distillation
Amine Sweetening
Merox I Sweetening
Hydrotreating
Desalting
Sum to Destination Caust. Dist.
Am.Sw. M1S
Hydr.
Desalt Sink
Sum from Source
2.4
25
8.571
8.422 25.572 49.367 n/a
119.332
0.265
0
0
1.606
0.794
0
0
2.665
0
0
0
0.316
0.937
2.032
21.715
25
0
0
0
0
0
8.571
0
8.571
0
0
0
0
0
0
10.344
10.344
0
0
0
0
0.822 27.303
0
28.125
0
0
0
0
0
0
87.273
87.273
2.665 25 8.571 10.344 28.125 87.273 119.332 GAMS Model Robustness
GAMS
• Other modes of the GAMS program can be used find
Other
solutions that:
solutions
Minimize consumption, then minimize the cost with the
initial solutions fixed
initial Origin Destination
Fresh Water Source
Caustic Treating
Distillation
Amine Sweetening
Merox I Sweetening
Hydrotreating
Desalting
Sum to Destination Caust. Dist.
Am.Sw. M1S
Hydr.
Desalt Sink
Sum from Source
2.4
25
8.571
8.388 26.058 48.914 n/a
119.331
0.267
0
0
1.645
0.755
0
0
2.667
0
0
0
0.312
1.312
1.662
21.715
25.001
0
0
0
0
0
8.571
0
8.571
0
0
0
0
0
0
10.345
10.345
0
0
0
0
0 28.125
0
28.125
0
0
0
0
0
0
87.273
87.273
2.667 25 8.571 10.345 28.125 87.272 119.333 GAMS Model & Regeneration
GAMS
• Treatment units can be added midprocess to regenerate
process
streams
streams
• Equations must be updated to include flow from units to
Equations
regenerators and cost of regeneration
regenerators
• Calculation of outlet concentrations
Mixing Streams: Cmix = Σi(Cin,uniti * Fin,uniti)/Fout,total
Outlet Concentration: Cout = Cmix * X + Cregen
If contaminant is not reduced in regenerator, Cregen = 0 and
and
X = 1;;
1
If contaminant is reduced in regenerator Cregen iis its
s
is reduced
specified outlet concentration and X = 0
specified GAMS Model & Regeneration
GAMS
• Results:
Destination
Caust. Dist. 0
0
0
0
0
0
0
0
1.333
1.333 25
0
0
0
0
0
0
0
0
0 Sum to Destination Origin Fresh Water Source
Caustic Treating
Distillation
Amine Sweetening
Merox I Sweetening
Hydrotreating
Desalting
API and ACA
Reverse Osmosis
Chevron WWT 2.666 25 Am.Sw. M1S
Hydr.
Desalt API/ACA
8.571
0
0
0 n/a
0
0.014
0
0
2.55
0
0
0.016
0.013 19.814
0
0
0
0
8.471
0
0
0
0
0
0
0
0
0 29.893
0
0
0
0 56.396
0
0
0
0 n/a
0
2.187 10.905 74.413 n/a
0
7.885 19.072
0
0
8.571 10.086 29.993 74.426 117.124 RO
n/a
0
0
0
0
0
0
88.838
n/a
n/a
88.838 CWWT Sink
Sum from Source
n/a
n/a
33.571
0
0.1
2.664
0
5.156
24.999
0
0.1
8.571
0
10.085
10.085
0
0.1
29.993
0
18.03
74.426
28.289 n/a
117.127
0 n/a
88.838
n/a
n/a
28.29
28.289 33.571 • Consumption = 33.571 ton/hr, Cost: = $Mil 1.301/yr
• Published consumption is the same:
Consumption = 33.571 ton/hr, Cost = $Mil 1.110/yr
Cost is somewhat higher
Cost GAMS Summary
GAMS
• Conclusions
GAMS model values are very close to published results
GAMS
Model maintains effectiveness with use of regeneration
Model
Model is robust enough to predict results in other cases
Model Future Work
Future
• Include price of piping in GAMS model
Extensive piping networks may save in water cost
compared to simpler networks, but cost more to construct
compared
Cost of network, length and type of pipe required could all
be variables in model
be
• Include maximum flowrates
Include
flowrates
Need to include minimum flowrates previously explained
flowrates
Reasoning for using maximum flowrates is similar
flowrates
• Continue study with fixed initial setup and changing mass
Continue
loads
loads
This setup better models the cost over time, so extending
time period and increasing number of periods makes model
more useful
more Thank You!
Thank Questions? ...
View
Full Document
 Spring '10
 staff
 Chemical Engineering, Amine Sweetening

Click to edit the document details