Upload
others
View
39
Download
1
Embed Size (px)
Citation preview
Petrocontrol 1
Coker advanced control application at Petronas Melaka
Azura Bt. Azahar Y. Zak Friedman Seung-Yun Nam
Petrocontrol 2
Why delayed coking unit is a good APC candidate?
• Unit with operational difficulties – Drum switch disturbances – High temperature coking environment – Many constraints
• APC can – Improve response to drum switch
disturbances – Run the unit closer to constraints – Maximize middle distillates yield
Petrocontrol 3
Melaka coker design overview
Naphtha
FEED
COKE DRUMS
FRACTIONATOR
LCGO
COKER FURNACE
Fuel oil
STRIPPER
STABILIZER
LPG
QUENCH
HCGO RECYCLE
HHCGO
Petrocontrol 4
Coker fractionator configuration
FC
FC TC FC LC
FEED FC
LC
HHCGO
FC
FC
RECYCLE
FUELOIL LC FC
LC FC
FC
LCGO FC
QUENCH TC FC
FC TC
PC
NAPHTHA
LC FC
FC
HCGO
LC
Petrocontrol 5
Coker drums, drum A working
DRUM A
To FRACTIONATOR
Quench Quench
Blowdown Blowdown
Steam Steam
FROM FURNACE
DRUM B
Blowdown Blowdown
LI
FC
FC
FC
FC
LI
TI
TI TI
TI
TI
Petrocontrol 6
Drum B worming up
DRUM A
To FRACTIONATOR
Quench Quench
Blowdown Blowdown
Steam Steam
FROM FURNACE
DRUM B
Blowdown Blowdown
LI
FC
FC
FC
FC
LI
TI
TI TI
TI
TI FLOW AND HEAT DISTURBANCE
Petrocontrol 7
Drum B starting, drum A stripped
DRUM A
To FRACTIONATOR
Quench Quench
Blowdown Blowdown
Steam Steam
FROM FURNACE
DRUM B
Blowdown Blowdown
LI
FC
FC
FC
FC
LI
TI
TI TI
TI
TI 20% OF NORMAL FLOW, MUCH STEAM
Petrocontrol 8
Drum B heating, no drum A steam
DRUM A
To FRACTIONATOR
Quench Quench
Blowdown Blowdown
Steam Steam
FROM FURNACE
DRUM B
Blowdown Blowdown
LI
FC
FC
FC
FC
LI
TI
TI TI
TI
TI 30% OF NORMAL FLOW, NO STEAM
Petrocontrol 9
APC economics
• Maximize coker distillate, subject to naphtha and distillate specifications
• Control coker recycle ratio at target • Maximize fresh feed • Maintain property control, fractionator
inventory and heat balance control during drum switching
• Avoid excessive fractionator coking
Petrocontrol 10
Robust inferential model
FC
LC FC
FC Pumpdown
Pumparound
Pot
TI PI
Reconstructed TBP cutpoint
Dewpoint
TBP IBP
GCC inferential model
Petrocontrol 11
Naphtha 90% point inference
120
130
140
150
160
170
180
190May/04/2012
May/14/2012
May/24/2012
Jun/03
/201
2
Jun/13
/201
2
Jun/23
/201
2
Jul/0
3/20
12
Jul/1
3/20
12
Jul/2
3/20
12
Aug/02/2012
Aug/12/2012
Aug/22/2012
Sep/01
/201
2
Sep/11
/201
2
Sep/21
/201
2
Oct/01/2012
Oct/11/2012
Oct/21/2012
Oct/31/2012
Nov/10/2012
Nov/20/2012
Cocker nap
htha
90%
point, d
eg C
Lab Inferential
Petrocontrol 12
LCGO 90% point inference
330
340
350
360
370
380
390
400Ap
r/02
/201
2
Apr/12
/201
2
Apr/22
/201
2
May/02/2012
May/12/2012
May/22/2012
Jun/01
/201
2
Jun/11
/201
2
Jun/21
/201
2
Jul/0
1/20
12
Jul/1
1/20
12
Jul/2
1/20
12
Jul/3
1/20
12
Aug/10/2012
Aug/20/2012
Aug/30/2012
Sep/09
/201
2
Sep/19
/201
2
Sep/29
/201
2
Oct/09/2012
Oct/19/2012
Oct/29/2012
Nov/08/2012
Nov/18/2012
Cocker LCG
O 90%
point, d
eg C
Lab 90 Inferential
Petrocontrol 13
Two RMPCT controllers • Furnace RMPCT
– Maximize feed to furnace constraints – Balance furnace passes – Increase COT after drum switch to speed up
the return to normal operation • Column RMPCT
– Control product specifications, heat balance, mass balance, column constraints
– Accept the given feed, relax product specs if needed
Petrocontrol 14
Main column control variables • Product inferential calculations
– Naphtha and LCGO draw 90% points • Column inventory constraints that
become active during drum switches – Draw pans and overhead drum levels
• Temperature constraints in the lower section of the column
• Coker recycle ratio • Hydraulic constraints
– Maximum valve positions, flows or sometimes pump currents
• Coking furnace constraints – Most active are skin temperatures
Petrocontrol 15
Main column manipulated variables
• Column top temperature – Controls naphtha 90% point
• Pump-down flows – Control side draw 90% points – HCGO pump-down controls coker recycle
ratio • Other secondary MVs
Petrocontrol 16
Evidence of CIT increase
CIT
Steam generator valve
Petrocontrol 17
No throughput maximization currently CIT
Throughput
LCGO
Naphtha
After
Before
Petrocontrol 18
Evidence of naphtha shift to LCGO
Petrocontrol 19
Conclusion: delayed coking unit is a good APC candidate
• Conditions of successful implementation – Knowledge of the process and economics – Good quality inferential models – Much attention to detail – Operator and process engineer participation
at the design stage
Petrocontrol 20
Example of throughput increase
150
155
160
165
170
175
180
185
190
195
200
1 14 27 40 53 66 79 92 105 118 131 144 157 170 183 196 209 222 235 248 261 274 287 300 313 326 339 352 365 378
Minutes
M3
/ Hr
Throughput before APC Throughput after APC
Drum switch
Petrocontrol 21
Naphtha 90% point inference
130
135
140
145
150
155
160
165
170
175
180
24-May-11 29-May-11 03-Jun-11 08-Jun-11 13-Jun-11 18-Jun-11 23-Jun-11
NAP90_M NAP90_L
Petrocontrol 22
LCGO 90% point inference
340
350
360
370
380
390
400
410
420
24-May-11 29-May-11 03-Jun-11 08-Jun-11 13-Jun-11 18-Jun-11 23-Jun-11
LCG90_M2 LCGO90_L