Document X8kM5KReQz1B10ROko7BaJYgw
PPG Industries, Inc. Industrial Chemical Division Lake Charles, Louisiana
subiocT: pm ;
PPG INDUSTRIES
Of 1u4Jthh J.luuddicicial District ffiffimTT TT-1
No. 91-1145
LJ\-ZZ. I U
February 1, 1982
BAILEY, T. A. HANSON, D. B.
The Reliability Study of the Lake Charles Plant "B-ll" - OHC II Unit CONFIDENTIAL
Distribution G. C. Strickler (Abstract only) 1. R. P. Lynch/J.E. Fike 2. R. P. Byars/E. J. Tullier 3. D. K. Pearson 4. J. H. Morgan/R. E. Baker (w/o Appendix) 5. R. E. Sanders 6. & 7. Authors 8. C. R. West - G. O. Technical Files (Title Page Only)
Key Words OHC Reliability Study EDC HCI Release Proof-Testing Vapor Cloud Explosive Mixture Dowtherm Furnace Loss Prevention Review Toxic Release
CONFIDENTIAL: Subject to Protective Order of 14th Judicial District Court
No. 91-1145
SL 011929
LR--2215 February 1, 1982
ABSTRACT
A detailed Reliability Study (Level III) and a Loss Prevention Review was made on the OHC II Unit of Tri-Ethane II. Fault trees were drawn and quantified to determine the frequency of occurrence of potential hazards. The potential hazards identified and studied include blown rupture discs on the Reactor, DH Still, and Lights Still, formation of an explosive mixture in the Reactor vent, tube rupture in the Dowtherm furnace, and continuance of natural gas flow into a hot Dowtherm furnace after loss of flame.
Of the three pieces of equipment studied for hazards from a blown rupture disc only the Reactor posed a significant hazard - the release of HCI. The occurrence of this hazard can be reduced from once every 13.3 years to once every 104.2 years by proper proof-testing of the existing instrumentation as recommended in this study. The other hazards show a similar decline in frequence of occurrence when proper proof-testing is maintained. With the implementation of the proposed recommendations the OHC II Unit will be a safe unit from a reliability standpoint.
COTIFIDETSTIAL:
of
S1u4btjhecjtutdoicPi arol. t'eupi-ciirsti^vareii'c-Ot rCdeorurt
No. 91-1145
011930 SL
INTRODUCTION
CONFIDENTIAL Subject to Protect! ;.*7e2 3O*r*d*er Of 14th Judicial Disz rict Court
No. 91-1145
LR-2215/2
A Level III Reliability Study was made of the OHC II unit as part of the continuing effort to study all of the units in Plants "B-l" and "B-ll". This report will complete the study of the Tri-Ethane II and OHC II units.
DISCUSSION
To develop a working reliability knowledge of OHC-II, plant tours were made and the P & I drawings and OHC-II operating manual were studied. Next cause-consequence sheets were developed for all the equipment in OHC-II. These sheets were reviewed with an opera tions representative for their completeness and accuracy. The cause-consequence portion of the study indicated the areas of potential hazard. These areas included (1) the OHC Reactor, (2) the Dowtherm Furnace, (3) the DH Still, and (4) the Lights Still. Fault trees were developed for these areas assuming system integrity (no pipe leaks, pump leaks, sample point leaks, or failure due to corrosion). The remainder of the system was considered to be covered sufficiently by the cause-consequence sheets.
The fault trees were quantified using data from ICI's manual of failure rates by D. W. Heckle along with information gathered from previous studies, operations experience, and other pub lished sources. Some of the failure rate data obtained was modified to fit the conditions applicable to the OHC-II system.
The hazards or top events covered in the OHC-II fault trees are (1) the release of flammables and HCI from the reactor, (2) the formation of an explosive mixture in the reactor vent, (3) th overheating and rupture of a tube in the Dowtherm Heater, (4) the loss of flame with continuing introduction of gas into the Dowtherm Furnace, (5) the leakage of gas into the
Dowtherm Heater causing an explosive mixture prior to light-off, (6) the release of com
bustible material (EDC) from the DH and Lights Stills. The frequency of occurrence (occasions/year) was calculated for each hazard. The effect of proper proof-testing was
determined by calculating the frequency of each hazard (1) with no proof-test, (2) with a
one year proof-test interval, and (3) with a six month proof-test interval. The release of material from the OHC Reactor, the DH Still, and the Lights Still was the only hazard for which the severity was calculated. All three sources of a release are capable of releasing flammable materials. During a shutdown the reactor can also release a toxic (HCI).
The severity of the release of HCI from the reactor was determined by the use of an ICI cornput r program TOXIDISP. Using the results of the program the severity of the release can be broken down into the frequency of occurrence of four hazard classes (See Table I). The fre quency of occurrence of each class (for no proof test, one year, and six month proof test) are presented in Table II.
Th release of flammabl s was handled according to C. R. W st's pap r "Fire/Explosion Methodology". Under the guidelines in this pap ronlyth r actor and th Lights Still released enough flammable mat rial to pose a hazard from a vapor cloud explosion. These two releas s were further studied using the TOXIDISP computer program to det rmine the
SL 011931
CONFIDENTIAL: Subject to Protective Order o lth Judicial District Court
Mo. 91-1143
LR-2215/3
probability that a vapor cloud of sufficient concentration (1/2 the lower explosive limit)
to pose a hazard would encounter a known source of ignition. The computer program rsvaalec that the cloud was well below the concentration to be considered hazardous (1/2 LEL) by the time it reached the ground. Therefore there is no hazard from the release of flammable material from the reactor, DH Still, or Lights Still other than a local fire hazard at the point of release.
In an effort to expand the scope of the reliability studies it was suggested by the Chief Loss Prevention Engineer that a Loss Prevention Review be undertaken for OHC-II. It was felt that the additional examination of the process with respect to codes of practice, fire pro tection, and other items not normally embraced by a Reliability Study could strengthen the overall process safety posture.
The Loss Prevention Review was completed with OHC-II using the General Office Loss Prevention Manual. After over six hours in several meetings it was determined that OHC-II met most of the requirements. Two questions required additional information and were handled as follows.
1. Is the vent header adequately protected by overpressure devices to prevent rupture of the line?
The section of line in question runs from the vent MLS (manual loading station) to the vent scrubber. This line is 10" FRP (Furan) with a 10 inch graphite rupture disc made to relieve at 10 psi. Several times during shutdowns this line has ruptured at a foint while simultaneously blowing the rupture disc. This is caused by hydraulic shock forces from a standleg of condensibles accumulated fust above the MLS. During shutdowns the non-condensible flow is increased and if the MLS is opened suddenly the hydraulic surge has caused the line to move and sometimes crack in the joints.
Expansion joints have been installed but with no appreciable effects. The stroke of the MLS is being closely monitored and along with a procedure for opening the MLS slowly, has prevented the line from rupturing on recent shutdowns. A drain line will be installed to control this accumulation of condensibles above the MLS. The combination of these measures should eliminate the problem of line rupture during emergency shutdowns.
2. Should the catalyst dump container be grounded to prevent ignition of catalyst
dust and reactor vent gases (ethylene) from static charge buildup?
The dump container is a cardboard drum with a metal rim. The drum cover Is fabricated steel connected to a short section of rubber hose connected to the piping from the spent catalyst collection bin. There have been no problems to date concerning possible ignition but the potential does seem to exist. The good record thus far might b attributable to th fact that static charges will not accumulate where the relative humidity is greater than 60%. Lake Charles weather usually meets this criteria. This question will be address d further by the ORC in hop of rendering a definitive plan of action.
a>
o
CD CO N>
CONFIDENTIAL: Subject to Protective Order of 14th Judicial District Court
No. 91-1J45
LR-2215/4
As an extension of the Loss Prevention Review the Loss Prevention Job (LPJ) was studied to determine the status of fobs that the ORC or previous safety studies had requested but had not yet been completed.
Three fobs which were listed under the category "Awaiting ACT Approval" are completed or will be during the next scheduled outage. These jobs are listed as follows:
LPJ-839 Remote Isolation Valves at Dowtherm Heater (To be installed on next S/D)
LPJ-936 Dowtherm Heater Snuffing Steam Inadequate
The other pending fob is LPJ-52, Check Valves in Air Utility Drops (Review need). There is no special need beyond normal Lake Charles Complex consideration for these check valves in OHC-II. It will remain part of the plant wide study under P-978.
RESULTS
The severity and frequency of occurrence of a hazard due to release of HCl from the OHC Reactor is presented in Table II. This table only shows HCl releases from the rupture discs. The chance of a HCl release due to vessel failure caused by overpressure was so slight (see Table III) it wasn't considered to be a hazard. Table II shows that yearly proof-testing reduc s the total hazard from .07489 occasions/year to .0296 occasions/year. Six month proof-testing further reduces the frequency to .0096 occasions/year. This represents a reduction in the frequency of a blown rupture disc due to overpressure of once every 13.4 years with no proof-test to once every 104 years with a six month proof-test interval.
The other hazard associated with the reactor is the formation of an explosive in the vent. The occurrence of this hazard is reduced from once every 1.34 years with no proof-test to once every 46.3 years with a six month proof-test interval. This hazard can be further
reduced to once every 73.5 years by installing a redundant O2 temperature transmitter.
Table IV shows the hazard frequencies for the Dowtherm Furnace. A proof-test interval of six months reduces the hazard from a tube failure from once every 11.3 years to once every 1,429 years, and the hazard from natural gas continuing into a hot furnace after flame failure decreases from every 27.6 years to every 526 years.
The hazard associated with the DH Still was the release of material caused by overpressure. The frequency of occurrence of these events are shown In Table III. While the release of material from the DH Still did not prove to be a vapor cloud hazard. It is felt that it would be advantageous (operationally) to have its instruments on a one year proof-test interval. This reduces the frequency of loss of material and down-time from once every 17.1 years to once every 244 years.
The Lights Still also proved not to pose a vapor cloud hazard. The study showed that proof testing did little to reduce the frequency of blown rupture discs. Ther fore proof-testing is not necessary for reliability reasons.
SL 011933
CONFIDENTIAL: Subject to Protective Order
Of 14th Judicial District Court No. 91-1145
LR--2215/5
RECOMMENDATIONS
With the addition of one back-up transmitter the existing instrumentation is sufficient to make the OHC-II unit reliable. The following is a list of recommendations that resulted from the OHC-II stud/. The first three are most important, while the rest are made to Improve information available on the unit.
1. Install a redundant temperature transmitter and alarm on the oxygen flow to the reactor. This will back up TT-664.
2. Provide proper proof-testing at the intervals listed in Table V.
3. The use of the master key to by-pass the shutdown system should be kept to a minimum. For this study use of the by-pass for 2 hours a week (historic informa tion from an operations representative) was assumed. Any significant increase in the frequency or duration of use of the by-pass will significantly increase the probability of a hazardous situation.
4. Revise the operating manual. The manual has not been updated since start-up. Several areas needing attention are as follows:
a. Jet compressor controls
b. Recycle treatment
c. More detailed and updated equipment data
d. Operating parameters
5. Update the P & I drawings to show any changes made to the process and equipment.
The Loss Prevention Job (LPJ) list was reviewed. The only job not in progress was LPJ-52, which called for installation of check valves on all plant air drops. This was looked at and no hazards unique to OHC-II were seen. Therefore it is felt that this problem is better handled on a plant wide basis under the existing project (P-978) for this purpose.
SL 011934
Date Date
Date z/zz/%z~ Date -XX, / XX f V' <?-
CONFIDENTIAL:
_,.
protective Order
Subject to
strict Court
of 14th Judicial D*s|-ict
No. 91 ~ * "*
TABLE I
TOXIC GAS HAZARD CLASSIFICATIONS
LR-2215/6
Toxic gas emission hazard calculations were made using the I.C.I. prepared computer program TOXIDISP to determine the probability for obtaining various hazard categories at the plant boundary based on multiple conditions such as release rates, location, weather conditions, etc.
Hazard Class
0
II
III
Definition
No nuisance to the public at the plant boundary.
May be smelled and thereby cause a nuisance to the public at the plant boundary - unlikely to cause alarm.
May cause some degree of distress to people, or damage to vegetation at the plant boundary which would cause a nuisance and which could lead to claims for compensation.
Could be dangerous or distressing to personal safety, or could be possible risk of life at the plant boundary.
SL 011935
v Confidential:
6f l4thCU Protective Order th Judicial District Court
No. 91-1145
LR-2215/7
TABLE II
HAZARD SUMMARY - REACTOR HCl RELEASE
Category
No Proof Test Occ./Yr.
1 Yr. Proof Test Occ./Yr. ,
6 Mo. Proof
Occ./Yi
III .0201
(49.8)*
.0079 (127)*
.0026 (385)*
II .0235 (42.6)*
.0093 (108)*
.0030 (333)*
1 .0109
(91.7)*
.0043 (233)*
.0014 (714)*
0 .0204
(49)*
.0081 (123)*
.0026 (385)*
Total
.0749 (13.4)*
.0296 (33.8)*
.0096 (104)*
* Yea re Per Occurence
SL 011936
Subject to Protective Order of 14th Judicial District Court
No. 91-1145
LR--227 5/8
TABLE III RELEASE FREQUENCIES OF VESSEL CONTENTS - REACTOR, DH STILL, AND LIGHTS STILL
Cause Reactor Over Pressure RD Release
Vessel Rupture
No Proof Test Occ./Yr,
1 Yr. Proof Test Occ ./Yr.
6 Mo. Proof Test
Occ./Yr.
.07489 (13.4)*
7xl0"6
(142,857)*
.0296 (33.8)*
2.96xl0-6 (337,838)*
.0096 (104)*
9.6x 10"7
(1,041,666)
DH Still Overpressure RD Release
Vessel Rupture
.0586 (17.1)*
5.87xl0"6 (170,358)*
.0041 (244)*
4.1 xlO-7
(2,439,024)*
.0033 (303)*
3.3x 10-7
(3,030,303)
Lights Still Overpressure RD Release
Vessel Rupture
.0491 (20.4)*
4.87xl0-6 (205,339)*
.0487 (20.5)*
4.87xl0"6 (205,339)*
.0487 (20.5)*
4.9xl0'6
(204,081)*
*Years per Occurence
SL 011937
CONFIDENTIAL:
Subject to Prf
of
14th
Jjuuddicial Di^r
No. 91-114!
Order ct Court
LR--2215/9
TABLE IV DOWTHERM FURNACE HAZARD FREQUENCIES
Cause
No Proof Test Occ ./Yr.
1 Yr. Proof Test Occ./Yr.
6 Mo. Proof Test
Occ ./Yr.
Tube Failure
Natural Gas Continues into Hot Furnace Explosive Mixture on Startup
.0883 (11 .3)*
.0362 (27.6)*
.009
(1111)*
.0012
(833)*
.0037 (270)*
-
.0007 (1,429)*
.0019 (526)*
-
*Years Per Occurence
Sl_ 0119^8
COKir -T T>'r.K' TI ?v i
Subject rr, T-.'cOrftor
Af, 14th auQJCir.j. -b 1 h'>
Court
TABLE V
LR-2215/10
FREQUENCY OF EXPLOSIVE MIXTURE (02/C?HA) IN REACTOR VENT
Cause
High O2 Flow (T)
No Proof Test Oce./Yr.
1 Yr. Proof Text Occ./Yr.
6 Mo. Proof T<
Occ./Yr.
.2284 (4.4)*
.0108 (92.6)*
.0095 (105.3)*
High O2 Flow
with redundant temp, trans.
.0027 (370)*
.00)5 (670)*
HCl Flow Change
.3152 (3.2)*
.0132 (75.8)*
.0097 (103.1)*
Low C2H4 Flow
.0905
(11.1)*
.0019 (516)*
.00094 (1064)*
Loss of Reaction
.0560 (17.9)*
.0008 (1250)*
.000557 (1795)*
Total Due from All Causes Using (j) Before Recommendations
.7443 (1.3)*
.0279 (35.8)*
.021597 (46.3)*
Total Due from All Causes Using (?) After Installing
Redundant O2 Temperature
T ransmitter
.0198 (50.5)*
.0)36 (73.5)*
* Years Per Occurence
SL 011939
CONFIDENTIAL? Subject to Protective Order of 14th Judicial District Court
No. 91-11*^
TABLE VI
LR--2215/71
PROOF TEST SCHEDULE OHC REACTOR #4 (EXPLOSIVE MIXTURE)
O2 Pressure Transmitter {for compensation) PT--661
02 Pressure Alarm PT--661
O2 Pressure Alarm PT-670 O2 Pressure Compensation PLC
O2 Flow Transmitter FT-662 O2 Flow Transmitter FT 622
02 Flow Deviation Alarm
O2 Flow Deviation Shutdown
O2 Flow Alarm FT-622, FT-662
02:HC1 Ratio Alarm 02:HCl Ratio Shutdown
O2 Temperature Alarm TT-664 2nd O2 Temperature Transmitter and Alarm when installed O2 Analyzer O2 Analyzer Alarm O2 Analyzer Shutdown PLC O2 Analyzer Shutdown Valves
Gas Chromatograph Gas Chromatograph Alarm
HCI Pressure Transmitter PT-657 HCI Pressure Alarm PT-657 HCI Pressure Compensation PLC
HCI Flow Transmitter FT-658 HCI Flow Transmitter FT-628 HCI Flow Deviation Alarm HCI Flow Deviation Shutdown HCI Flow Alarm FT-658, FT-628
C2H4 Pressure Transmitter (for compensation) PT-653 C2H4 Pressure Alarm PT-653 C2H4 Pressure Alarm PT-640 C2H4 Pressure Compensation PLC C2H4 Flow Transmitter FT-624 C2H4 Flow Transmitter FT-654
6 Months 6 Months 6 Months 6 Months
6 Months 6 Months 6 Months 6 Months 6 Months
6 Months 6 Months
6 Months 6 Months 1 /Week 6 Months 6 Months 6 Months 1 /Week 6 Months
6 Months 6 Months 6 Months
6 Months 6 Months 6 Months 6 Months 6 Months
6 Months 6 Months 6 Months 6 Months 6 Months 6 Months
SL 011940
CONFIDENTIAL! Subject to Protective Order of 14th Judicial District Court
No. 91-1145 TABLE VI (continued)
LR--2215/12
OHC REACTOR *4 (EXPLOSIVE MIXTURE) (continued)
C2H4 Flow Deviation Alarm C2H4 Flow Deviation Shutdown C2H4 Flow Alarm FT-624, FT-654 C2H4:HCl Ratio Alarm
^H^HC! RHo Shutdown
Reactor Pressure Alarm PT-732 Recycle Flow Alarm FT-734 Recycle Flow Alarm FT-736
Reactor Temperature Alarm TE-608 Reactor Temperature Alarm TE-618 Reactor Hot Spot Alarm Reactor Hot Spot Shutdown Reactor Bed Temperature Alarm Reactor Bed Temperature Shutdown
6 Months 6 Months 6 Months 6 Months 6 Months
6 Months 6 Months 6 Months
6 Months 6 Months 6 Months 6 Months 6 Months 6 Months
OHC REACTOR #4 OVERPRESSURE
Dowtherm Cooler Level Alarm LSL-609 Cooler Condensate Pump Shutdown Alarm UJR-555 Condensate Storage Tank Level Alarm LSL-551 Dowtherm Flow to Reactor Alarm FE-616 Reactor PCV to Scrubber PCV-729 Reactor Pressure Alarms PT-700 Reactor Pressure Alarms PT-702 Reactor Pressure Alarms PT-708 Reactor Pressure Alarms PT-720
I Year 1 Year 1 Year 1 Year 1 Year 1 Year 1 Year I Year 1 Year
DOWTHERM FURNACE
Dowtherm Pump Shutdown Alarm UJR-607 Dowtherm Head Tank Level Alarm LI-614 Dowtherm High Temperature Alarm TE-602 Natural Gas High Pressure Switch Dowtherm Low Flow Shutdown Dowtherm Low Flow Alarm FI-616 Dowtherm Furnace Shutdown System Dowtherm Temperature Alarm TE-618 Dowtherm Temperature Alarm TE-620
6 Months 6 Months 6 Months 6 Months
6Months 6Months 6Months 6Months 6Months
SL 011941
CONFIDENTIAL: Subject to Protective Order faf 14th Judicial District Court
No* 91-1145 TABLE VI (continued)
DH STILL
DH Still Bottoms Pump Alarm DH Still Bottoms Level Alarm LT-810 Steam Flow Alarm FV-804 Tempetature Alarm TE-815 Temperature Alarm TE--811 Cooling Tower Water Pump Shutdown Alarm DH Still Pressure Alarm PT-806 DH Still Pressure Alarm PT-808 DH Still Shutdown System
LIGHTS STILL
Lights Still Reflux Pump Shutdown Alarm Lights Still Reflux Flow Alarm FL-862 Lights Still Temperature Alarm TE-863 Lights Still Reflux Drum Level Alarm LT-864 Lights Still Bottoms Pump Shutdown Alarm Lights Still Bottoms Flow Alarm FT-851 Lights Still Steam Flow Alarm FT-858 Lights Still Temperature Alarm TE-880
LR-2215/13
1 Year 1 Year i Year 1 Year I Year 1 Year 1 Year 1 Year i Year
* * *
* * *
reasons. Suggested frequency would be on a one-year basis.
SL 011942
CONFIDENTIAL: Subject to Protective Order Of 14th Judicial District Court
No. 51-1145
LR-2215/14
appendix
SL 011943
