PDF Archive

Easily share your PDF documents with your contacts, on the Web and Social Networks.

Share a file Manage my documents Convert Recover PDF Search Help Contact



Plant Limits By Flare .pdf



Original filename: Plant-Limits-By-Flare.pdf

This PDF 1.5 document has been generated by Microsoft® Office Word 2007, and has been sent on pdf-archive.com on 19/07/2013 at 12:56, from IP address 118.172.x.x. The current document download page has been viewed 670 times.
File size: 3.1 MB (25 pages).
Privacy: public file




Download original PDF file









Document preview


JL-18407: G80 Plant Capacity vs. Availability of LRGO Stages by Otis Armstrong, P.E. NATSD/SSU

Summary
LP Flare System:



The oil feed rates for G80 are given in Table 1 of page 2. These rates are based on available LP
Flare stages. There are substantial variations depending on which stage(s) are out of service.
The GOR’ used to develop the Table is not a true facility GOR; LPPT gas is processed thru HP
flare system. GOR total = [ GOR’+ GORLPPT ]. Capacity = (SCF of LP flare)/GOR’

HP Flare System:
With current operation, loss of any HP flare runner is cause for immediate plant shutdown to repair
failure. Additional HP flare capacity can be obtained by reduction of depressure valve Cv. The
justification for Cv reduction is to comply with SAES-B58 par.7.2. The B58 standard allows 15 minutes
to depressure by 50%. The current G54 arrangement is to depressure in 2.5 minutes. The present HP
system Vented Shutdown procedure does not conform to original OIM or to design. Operation in
present conditions allows pressure & flows that have exceeded HP flare capacity. Operations which
exceed OEM capacity results in flare heat levels that exceed SAES-F-007 allowable levels.
Table 2 presents a review of HP flare operation data from PI. This review found 4 excursions in past 5
years beyond stated capacity of HP Flare system. The stated capacity of HP Flare per OIM manual is
2.05MM lb/hr at 11.9 psig header pressure, not to be exceeded! Repair operations inside fenced LRGO
without either plant shutdown or complete sparing of both LP and HP flares is unsafe. This is based on
review of heat radiation values and potential for H2S release in LRGO area.
Items of concern that can be addressed Short Term:
 Reconfigure vented shut down to sequence NV valves as specified on the ISS; see report for
details, (the double valves on booster & pipeline compressors are not sized to be opened
simultaneously). Current OIM sequence does not conform to original OIM valve sequence.
 Reconfigure vented shut down to utilize Burn-Pit Condensate Drains during de-pressure
sequence, not presently part of sequence.
 Add NV valves to PM schedule of 1/yr stroke test. A field check found 2 NV not in working
order, 1 NV had field stroke device removed, 1 NV found to have broken positioner
 Correct Stroke time of NV valves. The NV’s are specified to stroke in 30 to 60 seconds. Field
review found: 17% un-responsive and 60% stroked open too fast, under 30 seconds
 Repair LP Flare Flow Meter, signal jumps from zero to full scale
 De-commission Jump-Over between HP & LP Flare headers. The operating pressures between
these systems are not compatible.
Items of concern that to be addressed Long Term:
 Either retrim or Stroke limit NV vent valves to extend depressurization time from the design
depressure time of 2min. and 30 sec. to 10 min depressure time. SAES B-058-7.4, allows
depressurization time of 15 minute to 50% of design pressure.




Not Conforming to B-058 standard, very short depressure time, eliminates excess capacity in
HP Flare system. This is shown by HP flare header data, see report body. Spare HP flare
capacity can be obtained by limiting flow thru depressure NV valves. Recommended limitation
on HP flare flow rate is 275 lbs/sec 310MMSCFD. This value is about the maximum recorded
flow rate during Jan/2010 shutdown. Also reducing depressurization flow will save spared
flare total cost; operating (fuel purge), maintenance cost, and installation costs.
Size and Locate new HP spared flare based on the Revised Flow Rate. (no changes required for
LP system. The LP system maximum rate is set by compression performance. .

9-Mar-10 final issue

Page 1

JL-18407: G80 Plant Capacity vs. Availability of LRGO Stages by Otis Armstrong, P.E. NATSD/SSU

Background
This document details effects of
LRGO Flare stages losses on plant
processing capacity. There are two
flare systems, High and Low
Pressure, HP & LP. Both LRGO
systems are staged flare systems.
The Low Pressure system has 5
stages of progressive increasing
relief capacity. The High Pressure
system has 6 stages of progressive
increasing relief capacity.
The HP Flare system performs the
following functions:
 Plant G54 for shutdown
venting
 Plant G54 blocked discharge
compressor flow
 Glycol System depressure or
venting via PCV-914
 Provides G80-D1/D2 over
pressure
control
and
depressure venting via PCV107
 Receive some RV discharge
gas

TABLE 1: LP Flare Stage vs. Capacity
Calculation of G80 Inlet Capacity vs # LP Runners Avail
Lost
avail #/hr
mBPD mol/hr mSCF/DY SCF/BBL' case
-5
1,2,3,4 220000 431 5729
52,113 121 -1runner
-4
1,2,3,5 235000 460 6120
55,666 121 -1runner
-3
1.2,4,5 353000 691 9193
83,617 121 -1runner
-2
1,3,4,5 322000 630 8385
76,274 121 -1runner
-1
2,3,4,5 310000 607 8073
73,431 121 -1runner
-4,-5
-3,-5
-2,-5
-1,-5
-4,-3
-2,-3
-1,-3

1,2,3
1,2,4
1,3,4
2,3,4
1,2,5
1,4,5
2,4,5

95000
213000
182000
170000
228000
315000
303000

186
417
356
333
446
617
593

2474
5547
4740
4427
5938
8203
7891

22,503
50,454
43,111
40,269
54,008
74,616
71,773

121
121
121
121
121
121
121

-2runner
-2runner
-2runner
-2runner
-2runner
-2runner
-2runner

-5,-4,-3 1,2
88000 172 2292
20,845 121 -3runner
-2,-4,-5 1,3
57000 112 1484
13,502 121 -3runner
-2,-3,-5 1,4
175000 343 4557
41,453 121 -3runner
-4,-3,-2 1,5
190000 372 4948
45,006 121 -3runner
-1,-4,-5 2,3
45000 88
1172
10,659 121 -3runner
-1,-3,-5 2,4
163000 319 4245
38,611 121 -3runner
-1,-3,-4 2,5
265000 519 6901
62,772 121 -3runner
-1,-2,-5 4,3
132000 258 3438
31,268 121 -3runner
-1,-2,-3 4,5
265000 519 6901
62,772 121 -3runner
-1,-2,-4 5,3
147000 288 3828
34,821 121 -3runner
GOR' based on inlet BPD, 38.4MW avg from PFD's
GOR' from Average of 5yr PI, 98 & PFD, 113 sum. use 121

The LP flare system performs
following functions:
 Accept discharge of relief
valves (A.K.A. PSV’s or PZV)
 Serves as scavenger system to remove residual pressure in shutdown operations
 Take compressor discharge in event of blocked discharge
 Take gas if atmospheric compressor trip from sources: D3, D4, D14, C1, C2

Plant System Capacity for LP Stages Out of Service
Table 1 and Graph 1 present an analysis of plant feed capacity vs. number of LP Runners
remaining in service. The basis was capacity from LP flare charts, Molecular Weight of gas
from PFD and GOR’ determined by average of PI data and PFD data. The GOR’ referenced in
this section is NOT TRUE GOR. It is simply the GOR’ which relate to gas sent to LP Flare
system. The true plant GOR would need to add LPPT gas to arrive at total GOR. The LPPT gas
when not processed by LPPT compressor is sent to HP flare via control of G80-PCV106/7.

9-Mar-10 final issue

Page 2

JL-18407: G80 Plant Capacity vs. Availability of LRGO Stages by Otis Armstrong, P.E. NATSD/SSU

GRAPH 1 Inlet Feed vs. LP Flare Stages in Service

Plant System Capacity for HP Stages Out of Service
Depressurization sequence requires revision to re-sequence and stroke limit NV opening.
Until such revisions are made, the HP system is safety concern. The concerns are heat
radiation from possible high gas flows and need to enlarge existing heat boundary limit
fencing.
At present, any failures in HP runner system should be repaired by shutdown. No work should
be conducted with any HP stage out of service.
The HP system is depressurizing in too short a time which has in-turn causes an exaggerated
emergency depressure flare rate (several times those benchmarked for other facility). There is
potential to overpressure and over load existing flare system as indicated by historical data.
The following changes are recommended:
 Re-trim or stroke-limit the de-pressurization NV valves. This will increase depressure
time to conform to SAES B58 “de-pressure by 50% in 15 min.
 “Equip NV’s with smart positioner to facilitate future works, allow valve diagnostics and
more importantly, manipulate stroke time and valve characteristics” field survey of NV
stroke times found majority of valves opening in times less than ISS specification.
Slower valve stroke time reduces system shock and maximum flare rate. An example of
such shock effects can be found in damaged contactor trays of C-101. It is a known fact
that pressure shock waves damage tower trays. Aramco DGA Gas Treat trays were
replaced with trays of stronger support design. UGP & SGP implemented this multimillion dollar re-tray to handle pressure shocks which exceeded initial design of tray
mechanical limits. Limiting valve opening time will mitigate potential for future tray
collapses in C-101, Glycol Contactor.
 Re-sequence the depressurization steps. Current steps do not conform to OIM 34-085 pg. 34
 During vented shutdown; utilize the burn pit (G80-F4) by sending liquid to pit.

9-Mar-10 final issue

Page 3

JL-18407: G80 Plant Capacity vs. Availability of LRGO Stages by Otis Armstrong, P.E. NATSD/SSU

As shown below by 5 year trend of HP flare data, there are were 4 incidents where excess
pressures exceeding 12 psig happened. At 12.3 psig, the extrapolated flare rate was 2.2 million
pounds per hour, while rated flare capacity is 2.05MM#/hr. The Zink Manual requires
maximum rate not be exceeded. The consequence of exceeding design flow is to exceed heat
radiation limits of SAES-F-007.
Table 2: Excursions of HP System & Graph 2: HP Flare Capacity

REMOVE Jump-over Between HP & LP Flares
Currently there is an eight inch jump-over between HP & LP Flares. This jump-over needs to
be decommissioned. The HP flare drum and flow to HP Flare K/O drum is based on operation
at 37psia. The LP knock-out drum operation is based on pressure of 21.9 psia. These 2
systems are not compatible and jump-over needs to be decommissioned.

9-Mar-10 final issue

Page 4

JL-18407: G80 Plant Capacity vs. Availability of LRGO Stages by Otis Armstrong, P.E. NATSD/SSU

Graph 2: Zink HP Flare Capacity Curves

Graph 3: HP FLOW EXCERSIONS: flows above 375 mmscfd recorded only as “over range”

9-Mar-10 final issue

Page 5

JL-18407: G80 Plant Capacity vs. Availability of LRGO Stages by Otis Armstrong, P.E. NATSD/SSU
Graph 4: HP Pressure Excursions: p’s above 11.9 exceed flare rated capacity

Graph 5, 1985 OIM Depressure Flow Rate, 500#/s (1.8MM#/hr) max rate (OEM dsn 1.614mm#/hr)

TABLE 3: BENCHMARK FLARE De-Pressure Rate to Continuous Rate between TANAJIB to MANIFA

9-Mar-10 final issue

Page 6

JL-18407: G80 Plant Capacity vs. Availability of LRGO Stages by Otis Armstrong, P.E. NATSD/SSU

Benchmark of Depressure Flow Rates to Similar SA Facility
This overdesign of HP de-pressure rate of Tanajib is further confirmed by benchmarking
emergency depressure flow to max continuous flow for SA facilities in NAOO. For example on
SG-4 this ratio is about 300/100 =3/1 and at Manifa the ratio is 2.3, while at Safaniya Plant the
ratio is about 1.9 but for Tanajib the ratio is 650SCFd/60SCFd ~11/1 or 13.5/1 on a mass rate.
On per barrel of MSC crude Tanajib HP flare rate is 2 times higher 82/44 than Manifa. Both
Manifa and Tanajib have similar processes and oil.
Safaniya Gas Plant Depressurizing Summary: During the Gas Plant ESD, the Spheroid
Compressor trains will depressurize the gas only to the spheroid flare in 9 minutes. Then, the
Trap Compressor trains will vent the gas to the high-pressure flare in 10 minutes. By that time
the differential pressure across the depressurizing valve will be 45 psig. At this switchover
point, the valve to the high-pressure valve will start to close and the valve to the spheroid flare
opens. (From OIM of R74) // Based on R74 OIM, the switch over between HP & LP flares
should happen at 45psid, not 0.5kPa as given in the Tanajib OIM)

Depressure (NV) Valves Review
A review of the HP depressure NV valves was made by review of ISS sheets in conjunction with
CSD calve specialist. Their recommendation is as follows:
Per our phone discussion, and since these valves are globe valves, we suggest that you change their internal
trims to smaller ones to make these valves inherently safe. All of our major CV suppliers have other reduced
trim options to tackle this situation. We would like to revert to travel stops when absolutely nothing else
works. Furthermore, if they are the only solution, we want their installation to be in accordance with the
following:
1. They shall be marked in the P&IDs, and in the field
2. Their DCS signal shall have signal limit to the same valve stop values.
3. They shall be permanently welded type not screwed. Screwed types can be accepted if followed
by a seal weld.
We also highly recommend when you do any repair work for these valves to replace or equip them with
smart positioner to facilitate future works, allow valve diagnostics and more importantly, manipulate stroke
time and valve characteristics. This shall cause minimum cost impact to the project and will defiantly result
in better payback. Regards, Mohammed K. Al-Juaib, RSA for Instrumentation, Chairman for ISC
P&CSD/PID/Instrumentation Unit, Tel: +966 3 873 0222 Fax: +966 3 873 0037

The current ESD /emergency depressure sequence/ have violated original design intention of
the design. Present sequence has dual NV de-pressure valves on pipeline and booster
compressors opening at same time. A time delay between low Cv valve and high Cv valve is
correct sequence to mitigate potential for excess flow. Additionally valve should either be
stroke limited or re-trimmed to not exceed intended capacity.
Booster valve data sheets were based on NV1097/1098 opening 1st with stroke time of 30
seconds and valves 1050/1079 opening after 254psi drop, again with 30 sec opening time.
Additionally, the full open Cv compared to design Cv is 58% and 63% respectively. The valves
should be stroke limited to achieve specified flow, by allowing valve to full open the flow
capacity will be much greater than target. Field inspection determined valve to stroke to 100%
open, 1097 had 5 sec stoke time to full open, 1098 was not functioning, 1079 had 10 sec stroke
time to full open, and 1050 was found to not function.
9-Mar-10 final issue

Page 7

JL-18407: G80 Plant Capacity vs. Availability of LRGO Stages by Otis Armstrong, P.E. NATSD/SSU

The valve data sheets were based on NV809A/846A opening 1st with stroke time of 30 seconds
and valves 809B/846B opening after 458psi drop, again with 30 sec opening time.
Additionally, the full open Cv compared to design Cv is 65% and 52% respectively. The valves
should be stroke limited to achieve specified flow, by allowing valve to full open the flow
capacity will be much greater than target. Field inspection found NV’s to stroke 100% open.
Field inspection also found NV-809A was not functioning, 809B was not functioning, 846A had
10 sec stroke time to full open, and 846B was found to stroke 100% open in 32 sec. The stoke
times of other NV valve were as follows: NV.675-11sec open, NV.710-6sec, NV.747-9sec,
NV.777-45sec. All NV valves tended to stroke closed in same time as stroke open time. All
tests were conducted on Feb.1 with assistance of G54 operations.
Furthermore, review of depressure time with the large valves is very rapid. The B 58 standard
allows 15 minutes to depressure to 50% of pressure.

The depressurization sequence from the OIM of G54 are detailed below.
Page 29 of G54 OIM 34.085 in compliance with SAES B-058 7.4 Process vessels shall be designed with systems to
de-pressure to 50% of the vessel's design gauge pressure within 15 minutes if:... and..b) The vessel is designed for
pressures equal to or greater than a gauge pressure of 1725 kPa (250 psig).
 1. at time t=0 •
o Atmospheric Compressor K-101 HP Flare Valve NV-675 opens •
o LPPT Compressor K-102 HP Flare Valve NV-747 opens •
o Pipeline Compressor K-103 HP Flare Valves NV-809A and NV-809B open
 2. At time 60 seconds •
o Atmospheric Compressor K-201 HP Flare Valve NV-710 opens, if train 2 is operational
o LPPT Compressor K-202 HP Flare Valve NV-777 opens, if train 2 is operational
o Pipeline Compressor K-203 HP Flare Valves NV-846A and NV-846B open, if train 2 is if operational
 3. At time 250 seconds
o Booster Compressor K-105 HP Flare Valves NV-1097 and NV-1050 open
o Booster Compressor K-205 HP Flare Valves NV-1098 and NV-1079 open if train 2 is operational.
The rapid depressurization sequence was confirmed by review of data obtained during the
January 25, 2010 shutdown. Sample charts are included in report. Additional charts and data
are given in two additional files: Charts.pdf and NV_DePr.Evl.xls.
Rapid depressurization can damage equipment, such as collapse of TEG Contactor Trays.
Furthermore the above depressurization sequence does not follow the original OIM sequence.
Inclusion of entire OIM sequence is excessive long and is included in supplemental document,
charts.pdf. In brief, valve data sheets for 809AB and 846AB are clear valves not to be opened
simultaneously. Original OIM has no mention to open 809AB at same time. Instead, valve
809A open 1st followed by valve 809B after 60sec. The original OIM does not open K105/205
vent valves at same time.
The proposed valve sequent is presented below:

9-Mar-10 final issue

Page 8

JL-18407: G80 Plant Capacity vs. Availability of LRGO Stages by Otis Armstrong, P.E. NATSD/SSU

Tenative NV Sequence Times
time
0
60
90
140
200
230
action TR1 NV809A NV809B NV747
NV675
NV1050 NV1097
NV820, burn pit
NV850, burn pit
time
360
420
450
500
560
590
action TR2 NV846A NV846B NV777
NV710
NV1079 NV1098
Graph 6, G80 Depressure Flow Rate Jan/2010 Shutdown PI Data Trend

Typical Only, for all charts & development data, see file “Charts” which if added to this report file would
make it too large for server processing. Server Limit=10MB, this 4MB, charts 9MB total =13MB

9-Mar-10 final issue

Page 9

JL-18407: G80 Plant Capacity vs. Availability of LRGO Stages by Otis Armstrong, P.E. NATSD/SSU
Graph 7, G80 Pressure Decline Jan/2010 Shutdown PI Data Trend, Atmospheric Compressor

Atmospheric Compressor depressured in 1 minute.
Graph 8, G80 HP Flare flow & Pressure Jan/2010 Shutdown DCS Data Trend,

9-Mar-10 final issue

Page 10

JL-18407: G80 Plant Capacity vs. Availability of LRGO Stages by Otis Armstrong, P.E. NATSD/SSU
Graph 9, G80 HP Flare Stage Valves Flow & Pressure Jan/2010 Shutdown DCS Data Trend

Since pressure at flows greater than 600,000 #/hr, never exceeded 10psig during this
shutdown, the 6th runner stage was never used. The shutdown of Jan 25 should be termed a
near “ideal shutdown”. Maximum achieved flow was recorded at 225mmscfd or 1050000#/hr.
This plant shutdown indicates rate to HP flare never exceeded 1.05MM#/HR. Also capacity of
5 stages was never exceeded. An extrapolated rate of 1.3mm#/hr is required to activate stage
6.
By either retrim or stroke limit NV valves, it can be assured rate will never exceed 1.1MM#/hr.
The assurance of HP flows to rates that can be handled by stages 1 to 5 will provide sufficient
HP flare capacity to allow isolation of any single stage. Limiting HP flare rate to 1.1MM#/hr
will provide additional capacity to allow routine maintiance without plant shut down.
However for T&I of LRGO flares will require addition of spared flares. The addition of spared
flares is required because work on one elevated LRGO with other LRGO in service will exceed
heat levels specified in SAES-F-007.

9-Mar-10 final issue

Page 11

JL-18407: G80 Plant Capacity vs. Availability of LRGO Stages by Otis Armstrong, P.E. NATSD/SSU
Appendix 1

Scope of Work for Flare Upgrades

1. Make accounting of all relief valves in all plants G82, G80, G54, approximately 80 RV’s.
A list of relief valves may be obtained from the relief valve coordinator of NATSD.
2. Make accounting table in excel of where each relief valve dumps. The following
drawings will serve as reference for further investigation: G80-NA-J55021 through G80NA-J55030 all sheets.
3. Review and revise as necessary RV data sheets.
4. Develop Aspen-Hysys and Flare net model for facility, including depressurization model
which comply with SAES-B58 para.7.4. Some Volume and Pipe Isometric summary data
are available in Tanajib Library.
5. Review and revise as necessary the flare material balance sheets: G80-NA-J-54992sheets 1&2.
6. Make new drawing entitled: G80 “Thermal Radiation Map” for the flare area of all flares,
LRGO, Ground Flare F4 and any new additions. One map needing revision is included in
this report. Standards for flare maps are found in SAES-F-007.
7. Make preliminary sizing of new relief line from G-80-D3/D4 to flare area. Refer to study
SAMIR 6000001800 Recommendation 032 and RELIEF LOADS EVALUATION FOR
SPHERES G80-D-3 AND G80-D-4 IN MARJAN ONSHORE, P&CSD-T 0382/09 of Oct
19/2009. Approximate length for the new line is 7,500.0 feet, lay out of which is
depicted in Flow Diagram Figure 3
8. Make Layout & cost estimate for LP flare looping in spheres area, as recommended by
Energy Conservation study. A diagram is provided in the attachments.
9. Provide Cost estimate for NV Cv
cost $
d"
L or H, ft
correction to comply with valve ISS Item
$ 6,185,813 42
7500
specification. The ISS documents for New Sphere Header
550
NV valves are provided in the charts New Loop LP, C1/2+D14 $ 259,215 24
attachment.
LP Flare
$ 2,200,000 36
170
Reference Letters:
P&CSD: T-106/10,
T081/08
LP/TSU-005/10

Engi + OverHead

T382/09,

T357/09,

54
$ 1,686,754 n/a
$12,931,782 n/a

170
n/a
n/a

$ 2,600,000

HP Flare

Total BI Cost

Pipe cost basis

COST Basis Flares Work
Proposed for Tanajib G80
Comprehensive Flare

#/SF/in

40

$/# installed

5.00

pipe w/t in

0.375

Cost Estimate, All Flare Jobs

9-Mar-10 final issue

Page 12

JL-18407: G80 Plant Capacity vs. Availability of LRGO Stages by Otis Armstrong, P.E. NATSD/SSU

Appendix 2, Review of P&CSD Recommendation P&CSD-T021/10 Jan 17/2010

Review of Vendor Quotes based on P&CSD
Specification
The adjacent summary values of spare flare
quotes are based on P&CSD Specification.
It appears the Zeeco quote is superior because
 Zeeco flare tip requires less space to
disperse heat, dist to 500 BTUH/SF
 Zeeco Flare height is less, 160 feet vs.
Zink at 170 feet, meaning lower
installation & transport cost
 PSI required at base is less. This
especially important for the LP flare case
as back pressure on spheres will be
minimized by the Zeeco design.
 Zeeco delivery time is 4 weeks less than
Zink.

Criteria
Zink
Flare Diam inch
54
Flare Ht. Feet
170
Guy Wire Diam Ft.
no info
Pilot scfh
200
Purge Gas scfh
2247
Psi at base
3.8
Dist to 500BTUH/SF
1000
HP Gas Dsn Temp F
65
Delivery, week ARO
36
N/R = No Requirement
Wind Speed, fps
LP Min Flow Tip V fps
HP Min Flow Tip V fps

32
19
9

Zeeco
54
160
250
260
2250
3.0
840
60
32

Spec
N/R
N/R
N/R
N/R
N/R
3
N/R
60
N/R

32
19
9

if 36" tip /33inch ID tip used
LP Min Flow Tip V fps

45
23

HP Min Flow Tip V fps
Concerns with existing quote:
Use of spared flare leaving one LRGO in service while making maintenance activity on other
LRGO is not a feasible alternate. The entire basis of quotation needs revision.
Neither quote provides back pressure control valves with buckle pins. Such valves are
specialty to flare vendors
The 54 inch tip will have flare tip velocity less than the wind velocity, under minimum
continuous flow condition listed on drawing G80-A-NA-J-J54992 sheet 1. Such low flow will
accelerate flare tip deterioration. Such deterioration has been noticed on SG-4. Safaniya. GOSP
4 was originally provided with an oversized flare, which required frequent rebuilds due to tip
burn out. The original 36 inch HP tip was replaced with a 17” tip because original design basis
called for 2 minute depressure time.
It is recommended to take necessary steps to increase depressure time, re-rate flare flow to
1.1MM#/HR and resize tip to more closely match normal process flow requirements.

The requote should include buckling pin valves, requirement to re-rate LRGO heat radiation
map to lower HP flow, (a concern when working on LP LRGO system), provide detailed layout
considering flare K/O drums and LRGO heat map.

9-Mar-10 final issue

Page 13

B, LP

A, HP

New Spare Flare

To LP LRGO 54”RL94

To HP LRGO 48”RL90

New Spare Flare PFD

1.9psig

PIC
LP K/O D-25

H/S

10.0 psig

HP K/O D-26

Flare purge gas

Valve A to hold HP back pressure & -O2 in
Valve B to prevent O2 ingress to LP system

NEW

Hdr. Purge gas bypass

JL-18407: G80 Plant Capacity vs. Availability of LRGO Stages by Otis Armstrong, P.E. NATSD/SSU

Appendix 2, Review of P&CSD Recommendation P&CSD-T021/10 Jan 17/2010
Proposed Configuration of P&CSD-T021/10 Jan 17/2010

9-Mar-10 final issue

Page 14

JL-18407: G80 Plant Capacity vs. Availability of LRGO Stages by Otis Armstrong, P.E. NATSD/SSU
Appendix 2, Review of P&CSD Recommendation P&CSD-T021/10 Jan 17/2010 SPARED FLARE LAYOUT
A budget cost estimate for P&CSD proposal can be based on layout of EDR 017523 (Tanajib Spare Flare)
design as presented below. The plant data from Jan.25, 2010 shutdown shows it is possible to reconfigure
NV valves to ensure this rate is never exceeded. Alternatives to reconfigure valves are detail in CSD reply. It
will be necessary to then revise PFD’s and PID’s.
The precedence for such action is on offshore GOSP #4 and onshore GOSP1. SG-4 was originally sized with
flare of 36” to handle 312mmscfd from 2 minute depressure time. The depressure time was extended by
limiting gas rate to 114 mmscfd and using 17” flare tips. The 36” tips of S.GOSP-4 were burning out because
normal flare rate allowed flame to burn back into 36” tip. This caused chronic premature failure of 36”
oversized flare tip. This will happen at Tanajib if flare tip not reduced from 56” to 36”. Therefore it is
imperative to reduce design gas rate to more closely match actual processed gas rate. Below is my
calculation of a 56” Tip 160Ft high Heat Flux Curve for 1.1mm#/hr M#=0.26

36” Tip 160Ft high Heat Flux Curve for 1.1mm#/hr, M#=0.50, heat flux of 500btuh obtained at 576 feet.

Appendix 2, Review of P&CSD Recommendation P&CSD-T021/10 Jan 17/2010
9-Mar-10 final issue

Page 15

JL-18407: G80 Plant Capacity vs. Availability of LRGO Stages by Otis Armstrong, P.E. NATSD/SSU
Appendix 2, SPARE FLARE LAYOUT based on P&CSD Proposal
Using 1.1mm#/HR design rate and 36” tip will save about 60 feet of pipe run and allow pipe diameter to be
reduced from 56” to 42”. Suggest use of previous Job log layout as presented below for cost estimation
purposes. Final design should be based on process package from flare vendor to detail pipe diameter and
backpressure control valves.

The proposed heat rates are confirmed by Zeeco calc on a 56” tip, which agrees with my heat curves for 56
inch diameter tip.

9-Mar-10 final issue

Page 16

JL-18407: G80 Plant Capacity vs. Availability of LRGO Stages by Otis Armstrong, P.E. NATSD/SSU

Appendix 2, Review of P&CSD Recommendation P&CSD-T021/10 Jan 17/2010
SPARE FLARE LAYOUT based on P&CSD Proposal
This information is provided only for purpose of providing engineering contractor some background
information about the proposed location for spared flares. It in no way is an endorsement of a proposed
solution.

9-Mar-10 final issue

Page 17

JL-18407: G80 Plant Capacity vs. Availability of LRGO Stages by Otis Armstrong, P.E. NATSD/SSU

Economic Comparison – Spare Flare Options, Effect of Flare Size on Total Cost
Description
Cv 1
Cv 2
X-500 btu/hr/sf
Flare size, inch D
Flare size, -ft H
Purge Gas, #/hr
Purge Gas, mscfd
Pilot Gas mscfd
Total Gas/yr at 19dy/yr
$/mscf fuel gas
m$/Yr gas
Mean years/failure
Cost to overhaul m$
Maint. Cost m$/yr
total opern m$/yr
m$ NPV @ 10% dcfr

56" Flare 2.03MM#/hr 56" Flare 1.1MM#/hr

36" Flare 1.1MM#/hr

38556
2800
840
56
160
172.2
54.0
5.5
1131
$3
$3
3
$100
$33
$37
$331

21174
2800
615
56
160
172.2
54.0
5.5
1131
$3
$3
3
$100
$33
$37
$331

20308
2800
590
36
160
64.3
20.2
5.5
488
$3
$1
8
$67
$8
$10
$88

$514
$100
$100
$98
$103
$915
$1,830

$514
$90
$100
$52
$103
$859
$1,719

$484
$90
$100
$50
$97
$821
$1,642

retrim or stroke limit valves
sub valve work

130
325
455

130
325
455

130
325
455

Total Process Cost ESTIMATE
Total PROCESS NPV m$

$2,285
$2,615

$2,174
$2,504

$2,097
$2,186

FLARE Cap cost delivered m$
Buckle Pin Valves, 2pc m$
Tie-in cost (1-HP+1-LP)
Pipe
Instruments
Installation
Total Flare
NV & PCV valves work 13V's
Repair/Refubish actuator 45sec open t

Cost factors based on vendor Quotes & USEPA Doc. Ch7.
Costs of Pollution Control Equipment, Ch.7 Flares, 1995

Line item
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31

opn dy/yr
19

General: costs expressed as 1000’s of dollars, m$; LI-8, average operation taken at 19 days/year for fuel consumption;
LI-11 Mean time between failure estimates based on SG-4 experience with oversized flare tip; LI-18 Items A & B on PFD,
for prevention of Oxygen ingress prevention and keep flare header Mach # to 0.4 as per design., LI-15 the 10%dcfr
gives PV of 10 times annual cost. LI-23 is process engineering estimate, Design Unit to make final cost estimate. LI-25
to 28 is work required to prevent pressure excursion & required even if no spared flare installed. LI-31 is the total NPV,
not the total project cost. It shows Option C to be less expensive due to lower purge. LI26/27 pending vendor review,
the vendor review pending access

9-Mar-10 final issue

Page 18

JL-18407: G80 Plant Capacity vs. Availability of LRGO Stages by Otis Armstrong, P.E. NATSD/SSU
Appendix 3, Review of LRGO Area Heat Radiation Map
ORIGINAL RADIATION MAP & COORDINATES, with update:
Listed below are original heat radiation map coordinates and on following page is the Zink sketch submitted
to LPD in 1985. Basis was 1.614MM#/hr. No evidence is found of upgrade of heat map when flare rate uprated to 2.05MM#/HR. However since the mass flow rate is being de-rated to 1.1MM#/HR the value of
radiation lines will change to lower values. The value of radiation lines projected to 1.1MM#/HR are 1500 is
de-rated to (1.1/1.6 *1000 = 681Btu/hr/sf) and 1500 becomes 0.681*1500 = 1022. While if upgraded to the
higher relief rate, 1000 becomes 2.05/1.614*1000 = 1220 and 1500 becomes: 1900btu/hr/sf.

1500 btuh/sf
N
E
3078015 289731
3078015 289502
3078434 289502
3078434 289731
na-53023 Heat Map
Fence 289547

1000 btuh/sf
1500 btuh/sf
1000 btuh/sf
N
E
dist from
fence, M
3077985 289761
35n
60n
3077985 289472
30e
60e
3078457 289472
30s
60s
3078457 289761
45w
75w
MONC-2593/85 Zink 03/05/85 to ASC -1985
A-G80-NA-J55787
basis: 1.614E6 #/hr gas

The flare designer should use ‘Flarenet’ or similar certified program to update the radiation maps and input
the map as a drawing number for project IFC.

9-Mar-10 final issue

Page 19

JL-18407: G80 Plant Capacity vs. Availability of LRGO Stages by Otis Armstrong, P.E. NATSD/SSU

Appendix 3, Review of LRGO Area Heat Radiation Map
1985 LRGO Flare Radiation Map, reviewed for work hazard by Loss Prevention
The concerns are: 1) this is not valid heat radiation map 2) T&I work cannot be safely conducted in the flare
area if either HP or LP LRGO flare is operational because heat radiation values exceed 1500BTU/HR, as
indicated by this heat map. 3) Heat Radiation levels are understated in the pre 1990 LRGO models.

9-Mar-10 final issue

Page 20

JL-18407: G80 Plant Capacity vs. Availability of LRGO Stages by Otis Armstrong, P.E. NATSD/SSU

Appendix 4

Effect of Depressure Valve Cv upon Depressurization Time

Relationship Between Depressure Time, Vent Volume, and Valve Cv.
A determination of vent volumes can be determined by three methods, 1) calculation by
physical dimensions of vessels and pipes, 2) pressure verses time decline curves and
depressure equations; 3) integration of Venting rate and flow time; [scfm*(dt, min)] =scf; the
SCF are then corrected to actual volume by the gas law rules.

Appendix 4:

Graph 10, Equation for Gas Depressurization Time of closed vessel

Method of Half Lifes

The above de-pressure methods are somewhat complex. Comparison data from M.R. Beychok
was used to calibrate the spreadsheet and calculation method coded to an excel spreadsheet.
It can be seen that initial pressure
Method of Half Lifes
can be eliminated by the pressure
Halfs
1/2
1/4 1/8 1/16 1/32 1/64
term of gas density. The prior
2
3
4
5
6
above can be approximated by use Number of Half Lifes 1
i f 1.5 i s ti me for 1 ha l f l i fe 1.5
3
4.5 6
7.5 9
of half life method, time to 50%
initial
value
800,
V(n)=
400
200
100
50
25
12.5
reduction.
if value to be 20, then time estimated as 7.5 + 5/12*1.5=8.1

Time in sec to 50% is

t=[√(MW/(Z*oR))](V/(C*A))/ξ

where ξ is 190 @ k=1.3 & 192 @ k=1.2

The below table compares half life by example to actual calculation, the approximation agrees
to within 5% of actual formula values for given gas and V/CA.
t50%, s
45.1
61.4

V/CA
52376
52376

k
1.307
1.200

9-Mar-10 final issue

R
520
520

MW
16.04
29.00

Z
1
1

~t1/2
48.4
64.4

t25% s
95.3
127.2

~t1/4
96.8
128.8

t12.5% s t1/8
151.0 145.2
197.7 193.3

t 6.3%
213.0
273.0

~t1/16
193.7
257.7

Page 21

JL-18407: G80 Plant Capacity vs. Availability of LRGO Stages by Otis Armstrong, P.E. NATSD/SSU

Appendix 4 (continued)

Effect of Depressure Valve Cv upon Depressurization Time

Next relate CA to Cv. The relationship between (aC) and Cv is developed by equating Sonic
Control Valve flow to Relief Valve flow. This yields: 45(aC)=Cv, aC, square inch, ie Cv=950 gives
aC of 950/45=21.1sq”=0.147sqFT.
An alternative is the Crane TP-410 equation, Cv=29.9d2/√Kv and since d2=1.273a, so Cv
=38.1a/√Kv. Both methods are similar in result...
Irrespective or precise constants, involved in relation between C*A and Cv, for same vessel
with same gas at same temperature depressure time is: t50 =κ/Cv.
Application of Half Life Method to Revise NV Valve Cv or Alternatively NV Valve Stroke

Thus (Cvt50)1= κ=(Cvt50)2 . From this it can be seen that doubling Cv will half depressure time.
The total valve Cv is known from valve data sheets. Depressure times are know by initial plant
OIM’s and collaborated by actual plant data. Both data sets indicate 3 min depressure times.
Thus to increase depressure time to 9 min (allowing still better than 50% safety factor to SAES
–B058 of 15 minutes to 50% of pressure the new Cv may be determined as 1/3 of original
valve datasheet Cv.
Recommended G54 NV Valve Stroke Length or Cv reduction
name

NV-675 NV-747 NV-809A NV-809B NV-710 NV-777 NV-846A NV-846B NV-1097 NV-1050 NV-1098 NV-1079 Notes

NV# Cv ISS 1134 3097 600
Cv new
378 1032 200
Stroke
33% 33% 33%

1953
651
33%

1134 3097 600
378 1032 200
33% 33% 33%

1953
651
33%

262
87
33%

160
53
33%

262
87
33%

160
53
33%

exist valve total Cv
if retrim exist linear valve
stroke limit exist Vlvs

Adding new positioner and stroke limit existing valves is most expedient and cost effective
method since valves need not be removed from service. Add new smart transmitters with new
positioner and physical stroke stop can be achieved by work on valve positioner only.
REVIEW of G54 NV Valve Stroke TIMES
Valve Number

NV-675 NV-747 NV-809A

NV-809B NV-710 NV-777

NV-846A

t,s,open ISS

30
11
858
1134
0.76
132
40

60
out
1009
1953
0.52
194
122

std
60
10
32
387
1009
600
1953
0.65
0.52
155
194
580
122
over design %

t,s,open field
NV ds n Cv
NV Cv s el
Cv Ds n/s el

%Over Cap
Dsn Inlet Psi

30
9
1331
3097
0.43
233
140

std
out
387
600
0.65
155
580

out = not service nor stroke by either field or DCS

9-Mar-10 final issue

30
6
858
1134
0.76
132
40
avg=>

30
45
1331
3097
0.43
233
140
171

NV-846B NV-1097 NV-1050 NV-1098 NV-1079

nd
5
153
262
0.58
171
1755

30
30
99.9
160
0.62
160
1496

nd
out
153
262
0.58
171
1755

30
10
99.9
160
0.62
160
1496

diff btwn A&B valve P=diff open t

Page 22

JL-18407: G80 Plant Capacity vs. Availability of LRGO Stages by Otis Armstrong, P.E. NATSD/SSU

Appendix 5, Review of Tanajib Flare Design History

DBSP/PP 82'Braun
335 MBOD
508 MBOD
59mbd winter

IFC/ '85OIM

CURRENT OIM

508 MBOD
32 MBOD

600 MB/D
40 MB/D

Flare HP MMSCFD
Flare LP MMSCFD

1500
65

1500
65

Flare HP Diameter
Flare LP Diameter

66inch
60inch

48 inch
54 inch

48 inch
54 inch

1.614mm#/hr
0.683mm#/hr

1.614/2.05mm#/hr

2.05mm#/hr

60000 bpd

100mbd/15900 m3/cd

Man.Crude bpd
Marj Crude bpd
Condy bpd

HP Flare Max Flow #/hr
LP Flare Max Flow #/hr
burn pit bpd

Operating

various

0.392mm#/hr

Review of design Data found conflicting design values for HP flare. The spec to vendor, heat map, and
depressurization was based on Project Proposal, while an upward revision of flow to 2.05MM#/hr was
added to PFD without supporting documentation. Dwg: G80-F-NE-J-57108, F-G80-ND-J57123 rev 2 reflects
the lower mass rate while drawing G80-A-NA-J54992 states the higher rate. Appreciation is given to Ahmed
M Dhafeeri, (NATSD) for his research into this matter.

9-Mar-10 final issue

Page 23

JL-18407: G80 Plant Capacity vs. Availability of LRGO Stages by Otis Armstrong, P.E. NATSD/SSU

Proposed BI Work in Spheres & Flares Area

9-Mar-10 final issue

Page 24

JL-18407: G80 Plant Capacity vs. Availability of LRGO Stages by Otis Armstrong, P.E. NATSD/SSU

Proposed BI Work in LP Pipe Rack

9-Mar-10 final issue

Page 25


Related documents


plant limits by flare
flare parameters considerations
subsurface pump working principle
1 andy n sommeng aziz masykur lubad heri hermansyah
flare gas recovery
flare gas recovery


Related keywords