69
TEXT FOR SLIDES SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern University ©AICHE NEW YORK, NEW YORK 1998, 2009 Abbreviations and Nomenclature: BP Boiling Point LC 50 Lethal Concentration at which 50% of the population dies LFL Lower flammable limit MIC Methyl isocyanate ppm parts per million (volume) PVH Process vent header RVVH Relief valve vent header

SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

  • Upload
    vuthuan

  • View
    237

  • Download
    5

Embed Size (px)

Citation preview

Page 1: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

TEXT FOR SLIDES

SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY

Prepared by

Ronald J. Willey

Department of Chemical Engineering

Northeastern University

©AICHE NEW YORK, NEW YORK 1998, 2009

Abbreviations and Nomenclature:

BP Boiling Point

LC50 Lethal Concentration at which 50% of the population dies

LFL Lower flammable limit

MIC Methyl isocyanate

ppm parts per million (volume)

PVH Process vent header

RVVH Relief valve vent header

Page 2: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

VGS Vent-gas scrubber system

Introduction:

This slide package is a summation of events that led to and concluded

with one of the most significant toxic releases that has occurred to date within the

chemical process industry. It is intended for the education of chemical engineering

students, and focuses on the chemical processes involved. It is best used in a course

devoted to process safety; however, it can be used in courses dealing with kinetics and

runaway reactions, or courses related to mass transfer and dispersion modeling.

The material sources were numerous and a partial listing of articles that

have appeared about the incident are given in the appendix. A major source for

technical descriptions of the accident came from a paper prepared by Ashok Kalelkar

of Arthur D. Little Inc. under contract to Union Carbide.

Chemical Engineering News has devoted numerous articles to the

incident including the February 11, 1985, December 2, 1985, and December 19, 1994

issues. These articles served as major references to the crisis and several portions of

this module are extracted from these sources. Also, Dr. Paul Shrivastava’s book

Bhopal: Anatomy of a Crisis 2nd Ed. 1992 published by Paul Chapman Publishing

Ltd, London, and Frank P. Lees’ case history section in Loss Prevention in the Process

Industries: Hazard Identification, 2nd Ed. 1993 published by Butterworth-Heinemann,

Oxford, England, served as other significant references for this case history.

Page 3: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 2. The incident

At about 12:45 A.M. on Monday, December 3, 1984, an event occurred in Bhopal, India, that changed the way chemical process safety is practiced throughout the world. On that morning, 41 metric tons of a toxic gas, mainly consisting of methyl isocyanate, entered into the atmosphere from a Union Carbide India Limited (UCIL) pesticide plant. The release traveled with the prevailing wind into heavily populated areas located near the plant. Although accurate figures reflecting deaths and injuries do not exist, it is known that more than 1,000 people were killed and thousands more were injured or affected. Panic prevailed in the city of 900,000 inhabitants.1 In terms of loss of life, this is the largest chemical plant disaster recorded to date.

A personal account is provided by Mr. Rajat Vaish age 10 at the time.2 “These people that were running through the streets were some of the people that actually made it past their sleep... many people never woke up.” Mr. Vaish was at his uncle’s house at the time of the accident. When the warning went off, Mr. Vaish fled. His uncle, for some reason, stayed. His uncle became sickened and later passed away.

Image drawn by Daniel Willey, Sept 2009, all rights reserved.

1. NY Times, 8 Dec 1984 p 7.

2. As presented by a friend, Ms. Kokila Katyal, in a Professional Practice course, Univ. Pitts, 1998.

Page 4: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 3. A trigger question for the information that follows.

Page 5: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 4. Location

The disaster begins in the center of India in a region called

Madhya Pradesh. Bhopal was established by a Hindu king named Raja Bhoj

around 1010 A.D. In the early 1700’s, Rani (Queen) Kamlavati ruled Bhopal

under the protection of the emperor of India. After the emperor’s death, she

invited Dost Mohammed Khan, a Muslim chieftain, to be the territory’s

protector. After her death Dost Mohammed Khan annexed Bhopal to be his

own kingdom -- thus a change in rulers from Hindus to Muslim occurred. In

1956, Bhopal was named the capital of the state of Madhya Pradesh. The city

grew rapidly during the 60's to 80’s (and it continues to grow today). The

population in 1961 was 102,000. By 1981 the population had increased to

895,815.3 Much of this growth came by migration from rural areas

surrounding Bhopal. City housing could not be afforded by many of these

migrants, and they thus became squatters, creating slums and shantytowns.

By 1984, Bhopal had 156 slum colonies.4

3. 1981 Census cited in the NY Times 22 Dec 1984 p 4.

4. P. Shrivastava, “Bhopal Anatomy of a Crisis,” 2nd Edition, Paul Chapman

Publishing Ltd. London, 1992, pp 48 to 49.

Page 6: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 5 Slum locations - Jaya Prakash Nagar and Kenchi Chola

Two slums were located very near the Union Carbide chemical

plant in an area not zoned for residential use. The huts were small, constructed

of mud walls and tin roofs. No water or sewage facilities existed. The

residents were illiterate. They consisted of very poor people who had migrated

from rural areas surrounding Bhopal several years earlier. Often these slums

were controlled by illegal landlords who exploited residents for protection.

The Union Carbide fence line is shown in the upper right hand quadrant of the

above figure.

Page 7: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 6. Photograph of Bhopal

Bhopal was a developing city. At the time of the accident, only

10,000 phone lines existed for a city of nearly 900,000 residents. Four major

hospitals existed with 1,800 beds and 300 doctors. Electricity supply was

sporadic. Even with these shortcomings, the city offered more than most

towns and cities in Madhya Pradesh. The Indian and local governments had a

strong desire to industrialize cities like Bhopal.

Photograph by R.J.Willey, Photograph taken Dec 2004, all rights reserved.

Page 8: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 7. Union Carbide Corporation Interest

The Union Carbide plant located in Bhopal was owned by

Union Carbide (India) Ltd. Union Carbide Corporation held 50.9% of the

shares of the UCIL.5 The remaining shares were held among 24,000

shareholders, including the government of India, which owned approximately

25% of the UCIL shares. The largest division of UCIL was the Battery

Product Division accounting for about 60% of the 1983 sales. The

Agricultural Products Division of UCIL controlled the Bhopal plant, which

manufactured agricultural products including fungicides, miticides, herbicides,

and insecticides. Just over 8% of the sales of UCIL came from this plant.

5. NY Times 8 Dec 1984 p 1.

Page 9: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 8. History of the Bhopal UCIL plant

UCIL entered the pesticide market in the early 1960’s. The

Agricultural Products Division began operating the Bhopal plant in 1969. The

manufacturing facility was located 2 miles north of the Bhopal railway station.

Initially, the plant was used only for formulation blending of pesticides. In

1974, UCIL was granted permission to manufacture pesticides and gradually

the plant was “backward integrated” so that by 19786 methyl isocyanate (MIC)

production began at the plant. In order to conserve foreign exchange and

promote the local pesticide market, the Indian government required the facility

to be “backward integrated,” which forced the manufacture of all intermediate

chemicals, such as MIC and phosgene, on site, rather than importing the raw

materials. The plant had a capacity of 5,250 metric tons per year of MIC.

[Photograph reprinted with permission from Chem. Eng. News, January 4,

1988, 66

(6), p 8. Copyright 1988 American Chemical Society]

Page 10: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

6. Lees, A 5/2.

Page 11: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 9. Plant siting

When the UCIL plant was originally built in the late 1960’s, it

was situated in an industrial area 2 miles north of the main railway station that

served Bhopal. At that time no residential areas were close by. As the plant

“backward integrated,” municipal authorities in Bhopal objected to the

continued use of the UCIL plant at its original location. 7 The city’s

development plan had designated the plant site for commercial or light

industrial use, but not for hazardous industries. The local authorities were

overruled by central and state authorities. As time continued, tens of

thousands of people migrated to Bhopal from rural areas. Housing in the city

was insufficient. As shantytowns developed near the UCIL plant property,

UCIL drew the government’s attention to them and requested that a

“greenbelt” be established around the plant, but to no avail. Instead, the

government granted ownership rights in the land to the occupants. Map

captured from Google Maps, searching on “India” August 31, 2009, © 2009

Google –Map data © 2009 Europa Technologies.

7. P. Shrivastava, p 35. Originally referenced from Town and Country

Planning Department, Bhopal Development Plan (Bhopal Municipal

Page 12: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Corporation, 1975)

Page 13: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 10. Chemical production at the plant

The UCIL plant manufactured Sevin®, a Union Carbide trade name for a

pesticide, whose active ingredient is 1-napthyl-N-methylcarbamate8 or the

generic name carbaryl. The reaction involved two reactants, methyl

isocyanate and alpha naphthol:

8. C&EN 11 Feb 1985 pp 30-31.

Page 14: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 11. Another trigger question for the slides that follow.

Page 15: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 12. Methyl isocyanate (MIC): an extreme toxin

“Methyl isocyanate (MIC) is reactive, toxic, volatile, and flammable.” So begins the first page of the Union Carbide pamphlet.9 Let us first focus on the toxic effects. The maximum exposure during an 8 hour period is 0.02 ppm (20 parts per billion). This level is called the TLV (threshold limit value)-TWA (time weighted average). By comparison, phosgene, another extremely toxic gas, is 0.1 ppm.10

Individuals begin to experience severe irritation of the nose and throat at exposures of MIC above 21 ppm. The LC50 for rats exposed for 4 hours is 5 ppm.11 LC50 is the lethal concentration in which 50% of the population dies when exposed at this level.

In humans, exposure to high concentrations can cause enough fluid accumulation in the lungs to cause drowning.12 At lower levels of exposure, the gas affects the eyes and lungs. It acts as a corrosive agent, eating away at moist vulnerable tissue, such as mucous membranes and eye surfaces. Because the MIC is soluble in water and degrades rapidly, acute effects are short term.13

Long term effects do exist. Although no accurate figures exist, it is generally believed that, of the many thousands of people exposed, a significant number suffered permanent injuries. 14

9. C&EN 11 Feb 1985, p 27 as quoted from a Union Carbide pamphlet on MIC.

10. C&EN 11 Feb 1985 p 38.

11. NY Times 4 Dec 1984 p A8.

12. Ibid

13. NY Times 15 Dec 1984 p 4.

14. C&EN 19 Dec 1994 p. 12.

Page 16: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 13. Other critical physical properties of MIC (BP & density as a vapor)

Methyl isocyanate has a boiling point of 39.1C and a vapor pressure of 348

mm Hg at 20C. As such, it is quite volatile and it will easily enter into the

surroundings at very high concentrations. Its molecular weight is 57. Thus, it

has a molecular weight about 2 times that of air and as a vapor, it will have a

higher density compared to air. It will tend to travel along the ground and

lower areas as a vapor when initially released into the atmosphere. However,

dilute mixtures in air will follow the predominant flow pattern of air. Dilute

mixtures still can be very toxic as this case proves.

Page 17: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 14. Reactivity of Methyl isocyanate -- the adjacent double bond system.

Methyl isocyanate, which is of the isocyanate family, is a very reactive

molecule because of the cumulative unsaturation system of R-N=C=O where

the effect of the adjacent double bonds adds to the instability of the

molecule.15.

15. C&EN 11 Feb 1985 p 27.

Page 18: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 15. The potential products of MIC reaction in the presence of water or a

catalyst

In the presence of small amounts of water, MIC reacts to form 1,3,5 trimethyl

biuret and CO2. In the presence of excess water, MIC reacts to form

methylamine that reacts with additional MIC to form 1,3 dimethylurea and

carbon dioxide.16. The overall reactions are exothermic and thus energy is

released. The initial amount of energy release is around 585 BTU per lb of

MIC or 3700 BTU per lb of water reacted. 17 A trace of acid or base will

catalyze the reactions. This energy, if not removed properly, will heat the

mixture, eventually bringing the mixture to its boiling point.

16. Ibid

17. C&EN 11 Feb 1985 p 28.

Page 19: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 16. MIC can react with itself exothermally to form a trimer.

In the presence of a basic catalyst, MIC can react with itself to form trimethyl

isocyanurate. Energy release for this reaction is about 540 BTU per lb of MIC.

Additional catalysts include iron, copper, tin and zinc.18.

18. C&EN 11 Feb 1985 p 29.

Page 20: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 17. Another trigger question for the slides that follow.

Page 21: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 18. A runaway reaction -- An exothermic reaction in a closed system.

As additional energy is released, in a closed system, both temperature and

pressure will increase approximately following the vapor-liquid equilibrium

line of the mixture, which includes CO2, a gaseous product. In closed systems,

the pressure continues to rise to the thermodynamically fixed maximum or it is

relieved.

Additional heat can be formed by secondary reactions involving

methylamine and MIC to give additional products.

Page 22: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 19. The production of MIC

The MIC required at the Bhopal facility was produced in batches every two to three months and stored in three cylindrical tanks. The last batch of MIC was made between October 7 and October 22, 1984. On October 22, the MIC unit was shut down for routine maintenance of the equipment, and the tanks were isolated. The process involves the following series of reactions done at the Bhopal site:

Formation of phosgene

2 C + O2 ---> 2 CO

CO + Cl2 ---> COCl2

phosgene

Next, methylamine was reacted with excess phosgene in the vapor phase to form methylcarbamoyl chloride and hydrogen chloride.19

heat

COCl2 + CH3NH2 ---> CH3NHCOCl + HCl

phosgene methylamine methylcarbamoyl

chloride

After the reaction, the reaction products were quenched in chloroform. The unreacted phosgene was separated by distillation from the quench liquid and recycled back to the

Page 23: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

phosgene/methylamine reactor. The hydrogen chloride was removed and the remaining material, methylcarbamoyl went to a pyrolysis unit where the following reaction occurred:

heat

CH3NHCOCl ---> CH3NCO + HCl

methylcarbamoyl methyl

chloride isocyanate

The HCl produced was sent to an absorber.20 Crude MIC from the pyrolysis unit went to a refining still where the top product was sent to the MIC storage tanks and the bottom product recycled back to the pyrolysis unit.

19. Adapted from Lees, A5/2.

20. C&EN 11 Feb 1985 p 32.

Page 24: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 20. The storage tanks

Three storage tanks existed for MIC produced at the site. Normally, two tanks

held production while a third was available for emergency use. The storage

tanks were constructed of 304 and 316 stainless steel. They were partially

submerged into the ground and the remaining exposed area was covered by an

earth mound upon which was placed a concrete barrier. Their dimensions

were 8 feet in diameter and 40 feet in length. Each tank’s volume was about

56,500 liters or 15,000 gals.21 Their pressure rating ranged from full vacuum

to 40 psig (2.72 bar) at 121 C. Each tank could hold about 45 tons of MIC,

which is sufficient for about 15 days of pesticide production. [Figure adapted

from Chemical Week/November 26, 1986 p 8. Original Source: Union

Carbide.]

21. Chemical Engineering 24 Dec 1984 p 17.

Page 25: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 21. Photograph of Storage Tanks

This is a photograph showing the pipe racks and storage tanks. The storage

tanks are located near the center of the photograph towards the right side.22

22. Courtesy, Mr. Wil Lepkowski, Senior Correspondent C & E News.

Page 26: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 22. Storage tanks continued, with process piping detailed

The process side of the storage tanks is shown in this slide. Material flowed

through a feed pipe opening located in the tank sump into one of two pumps.

One pump served as a circulation pump for a refrigeration unit. The other

pump served as the Derivatives Unit transfer pump that supplied MIC to the

Derivatives Unit. Two return lines existed -- one for the return from the

refrigeration unit and the other from the Derivatives Unit supply pump outlet.

Refrigeration was provided, in part, because surrounding temperatures in the

region could reach as high as 48 C (120 F) in the summer, is above the boiling

point of MIC at atmospheric pressure. Also, MIC is quite volatile even at

room temperature as Slide 11 pointed out. [Figure adapted from P. Shrivastava,

Bhopal: Anatomy of a Crisis, 2nd Ed. 1992 p 38. Original Source: Bhopal

Methyl Isocyanate Incident Investigation Team Report, Union Carbide

Corporation, March 1985.]

Page 27: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 23. Storage tanks continued, with nitrogen header added.

Nitrogen was supplied to the storage tanks for three reasons. First, bone dry,

high purity nitrogen was used at the Bhopal plant to reduce the risk of

contamination, especially from trace amounts of water. Secondly, MIC is a

fuel and thus can burn -- so contact with oxygen is avoided. MIC’s flash point

is -18C, and its LFL (lower flammability limit) is 6% in air. Thus nitrogen

“padded” the flammable contents of the tank. This is also called nitrogen

blanketing or nitrogen purging. Nitrogen or inert padding is commonly used

in tanks holding volatile fuels or reactive chemicals. Thirdly, the nitrogen pad

provided a positive head of 15 to 20 psig within the tank so that material could

be forced through the feed pipe to the transfer pumps. Nitrogen supply

pressure was controlled by control valve labeled N2 in this figure. [Slides 20

through 23 were adapted from Ashok S. Kalelkar (ASK), "Investigation of

Large-Magnitude Incidents: Bhopal as a Case Study," in I. Chem.E.

Symposium Series No. 110 The Institution of Chemical Engineers 1988, p 575,

Figure 7.]

Page 28: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 24. Storage tanks continued with process vent header

To the left of the nitrogen purge line existed a process vent header with a

control valve labeled “Valve 15” in this figure. Normally, Valve 15 was closed

except on vent line purging or control of moderate over pressure during filling.

The process vent header is used to vent normal process vents from various

parts of the unit to scrubbers or flares, so that they can be neutralized before

they reach the environment. Both the nitrogen pad line and the process vent

header line shared the same entry point to the storage tank through Valve 16.

Process vent lines are a common safety feature used for the processing of

hazardous chemicals. Finally, note the pressure gauge labeled “10” and the

valve directly below it. More details about this later.

Page 29: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 25. Storage tanks continued with relief valve vent header added

Relief systems are required on closed storage and reaction vessels. They serve

to protect the equipment and personnel from an explosion that may occur if the

vessel over pressurizes and to provide a mechanism to transfer and neutralize

material vented through the safety valves. If an over pressurization event

occurs, relief systems “trip” at some design pressure “relieving” the unit of the

excess pressure and preventing failure of the vessel. The relief system is

labeled as “11” and “2” in this diagram. In this case, the design involved a

rupture disk in series with a relief valve. This is common practice with highly

corrosive and toxic materials. The rupture disk isolates the stored material

from the relief valve, preventing damage to the relief valve during normal

service. Note the pressure gauge on the line. This gauge can indicate pin hole

leaks that may develop in the rupture disk over time.

Page 30: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 26. One last detail about the relief valve vent header

Note that the process vent header (PVH, #3) was connected to the relief valve

vent header (RVVH, #2) by a temporary jumper connection to permit routine

maintenance to be performed on the PVH, while the MIC unit was shut down.

One explanation of how the incident occurred is based on this connection

having these headers interconnected through open valves.

Page 31: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 27. Downstream of the relief valve -- the vent gas scrubber system

The relief valve vent header (RVVH) and process vent header (PVH) went

separately to a flare tower or a vent-gas scrubber system (VGS). This is good

engineering practice to prevent uncontrolled releases. In an unrelated incident

at a ICMESA plant in Seveso, Italy, an uncontrolled toxic release of dioxin

went straight into the surroundings. It resulted in the dispersion of about 2 kg

of dioxin throughout a 16 km2 area and forced the relocation of 700 inhabitants

in the most severely contaminated zones.23

At the Bhopal plant, the vent-gas scrubber system was a packed

column with three sections. The upper section was a 1.65 m diameter, 5.54 m

high section that held ceramic Berl saddles.24 A middle section, 1.65 m in

diameter and 2.1 m in height, separated the upper section from the bottom

section. The bottom section, 3.6 m in diameter and 6.9 m in height, was the

accumulator with a capacity of about 80,000 liters that held a 10% caustic soda

solution. Caustic soda neutralizes MIC before it releases to the atmosphere

through a 100-foot stack.

Piping and valving existed so that the RVVH and PVH lines

could also be routed to a flare tower. The primary function of the flare was to

burn vent gases from the carbon monoxide unit and the monomethylamine

Page 32: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

vaporizer safety valve. The flare also at times burned normal vents from the MIC

storage tanks (which could be routed to the VGS or directly to the flare), the VGS and

the MIC Refining Still. The vent gas scrubber system was part of an integrated

system designed to prevent, detect, and handle contamination, and was capable of

handling all reasonably foreseeable conditions.

23. See Seminar on Seveso Release Accident Case History, by Ronald J. Willey,

SACHE Slide Package 1994, AICHE.

24. A.Ritchie, “The Bhopal Disaster -A Critical Study” 3rd Year Report, University

of Nottingham, Dept. Of Chem. Eng. March 1988, p 10.

Page 33: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 28. Proper safety systems were in place

This part of the plant, built in 1981, had acceptable safety systems in place at

the time to prevent a release of MIC from reaching the environment. A visit

by a Union Carbide Safety team was made in 1982 and 10 safety concerns

were brought forth at that time, one of which included the potential for release

of toxic materials due to equipment failure, operating problems, or

maintenance problems. Corrective actions were taken, including replacement

of corroded valves on the MIC unit. So what else happened?

Page 34: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 29. Operating conditions prior to the disaster:

1. The refrigeration unit was shut down in June of 1984. The refrigerant was

drained and used elsewhere within the plant. Thus, MIC was stored without

the benefit of cooling, and, if an overheating event should occur, no cooling

was available.

2. Several weeks before the incident, after the MIC unit was

shut down, a corroded portion of the PVH line leading to the flare was taken

out of service for maintenance. During this time, the process vents could no

longer be routed to the flare, and were rerouted directly to the VGS.25

3. The scrubber was turned off and put in stand-by mode

because the MIC production unit was not operating. At the time of the

incident, however, the MIC unit operators were able to turn the scrubber on.

25. Information provided by Union Carbide to author, Mar 1998.

Page 35: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 30. All were operations management decisions

It is likely that the decision to alter and by-pass established safety systems

were all management decisions. Typically these decisions are made by line

supervisors, and managers at the plant level. Although the reasons behind the

decisions are not fully known, the decision makers could not have foreseen the

catastrophe that ultimately occurred.

Page 36: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 31. Another trigger question for the slides that follow.

Page 37: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 32. Management of change: Guidelines now exist

Be cautious about decisions related to changing systems and procedures that

were originally installed for safety. Federal guidelines now govern major

changes in the chemical process industry (noted as “Management of Change”),

and a SACHE module is devoted to this issue.26 The potential consequences

of those decisions must be carefully considered.

26. Robert M. Bethea, Slide Package Chemical Process Safety Management

Flixborough and Pasadena (TX) Explosions Miscellaneous Case Studies 1994.

Page 38: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 33. Sequence of events that directly lead to the event

Given below is the sequence of events that led directly to the release of MIC

into the environment.

Page 39: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 34. Storage tank schematic

Two storage tanks, E610 and E611, contained MIC at the time of the incident.

Tank E610 had about 41 metric tons of MIC and Tank E611 had about 21

metric tons of MIC. During the month of November 1984, both tanks had

periods of low pressure. Tank E610 could not be used at the time of the

incident. Its pressure registered 2 psig instead of the normal 20 psig at

pressure gauge 10. Corrective maintenance work was performed on Tanks 610

and 611 on December 1 to correct the low pressure situation. Although the

plant personnel were able to correct the problem with 611, they were unable to

pressurize Tank 610. No work was done on either Tank 610 or 611 on

December 2 (two days before the event).

It is speculated that Tank E610 could not hold pressure because

normally closed Valve 15 was leaking.

Page 40: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 35. Reactivity

As previously mentioned, water and MIC react exothermally; therefore, efforts

were made continuously to keep water and MIC apart. However, as part of the

normal operating procedure, transfer lines had to be periodically flushed with

water because MIC can self react to form a polymeric species along pipe walls

that eventually lead to blockages.

Page 41: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 36. Two major explanations

Two major explanations have evolved to explain this event. Both explanations

agree that water entered into the MIC storage tank and runaway reactions

ensued. One explanation, referred to as the water washing theory, is based on

plant employee testimony and was issued by the authorities commissioned by

the Indian Government.27 The second, or sabotage theory, is based on

independent research conducted by Arthur D. Little, Inc., on behalf of Union

Carbide.28 We will begin with the first explanation. However, some details

about the second explanation will be included to assist in the understanding of

both explanations.

27. S. Varadarajan et al. “Report on Scientific Studies in the Factors Related to

Bhopal Toxic Gas Leakage (New Delhi: Council of Scientific and Industrial

Research, Dec 1985).

28. Ashok S. Kalelkar (ASK), “Investigation of Large-Magnitude Incidents:

Bhopal as a Case Study,” in I.Chem.E. Symposium Series No. 110 The

Institution of Chemical Engineers 1988, p 561.

Page 42: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 37. The order to clean the lines29

On December 3 at about 9:30 P.M. that evening, the second shift production superintendent ordered the MIC plant supervisor to flush out a 2-inch filter pressure safety valve line near the process filters in the production unit, over 600 feet away (by pipe) from the MIC storage tanks. The relief valve vent header (RVVH) is tied to the filter pressure safety valve line through Valve 19. Valve 19 is normally closed. Plant maintenance records dated 29 Nov 1984 support this contention, and Valve 19 was tagged as closed.30 Valve 19 was tested in July 1985 and was determined to be leak tight.31

Part of the normal operating procedure included the insertion of a blind on the line leading to the RVVH to prevent water from accidentally entering this line. This procedure was normally carried out by maintenance personnel. It is assumed that the blind was not inserted. Flushing began by operating personnel shortly after 9:30 P.M.32 The proponents of the water washing theory contend that the water leaked by Valve 19 and entered the RVVH. This means that water had to have at least a hydraulic head of 10.4 feet to reach the RVVH line, filling a 6-inch header pipe to reach the 8-inch RVVH pipe. Then, the proponents hypothesize that water flowed through 65 feet of the 8-inch pipe, followed through another 340 feet of 4-inch pipe around the jumper pipe into the process vent header (PVH). Before the water finally reached the Tank E610, water had to have filled an additional 340 feet

Page 43: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

of 2-inch PVH lines with several drops before entering the tank. Calculations showed that 540 gallons of water were required to fill all of the dead spaces in the pipe line before reaching the storage tanks.33 The lines could conceivably fill with water in 1.8 hours at 5 gallons per minute.

29. Diagram is from ASK Fig 4.

30. Ashok S. Kalelkar (ASK), “Investigation of Large-Magnitude Incidents: Bhopal as a Case Study,” in I.Chem.E. Symposium Series No. 110 The Institution of Chemical Engineers 1988, p 561.

31. ASK p 561.

32. P. Shrivastava, p 39.

33. ASK p 561.

Page 44: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 38. What went in, didn’t come out

As presented in Shrivastava’s book,34 the operator who was in charge of

flushing noticed that water wasn’t exiting all of the bleeder lines (the proper

exit for the flushing operation). He ceased the flushing operation until the

MIC plant supervisor ordered him to resume the process. As will be presented

below in the alternative explanation, the argument goes that 3 of the 4 bleeder

lines labeled 18 in Slide 37 were open; therefore, insufficient back pressure

existed to force water back through the closed Valve 19.

34. P. Shrivastava, p 39. From International Confederation of Free Trade

Unions, The Trade Union Report on Bhopal (Geneva: International

Confedieration of Free Trade Unions, July 1985).

Page 45: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 39. The path of the water flow

The proponents of the water washing theory speculate that water flowed from

the water washing area to the MIC storage tanks. It appears that the reason

that the tank E610 didn’t pressurize might have been that Valve 15 was faulty,

as it allowed water to enter Tank E610. About 500 kg of water entered the

tank and began to react exothermally with MIC. [Adapted from Ashok S.

Kalelkar (ASK), "Investigation of Large-Magnitude Incidents: Bhopal as a

Case Study," in I. Chem.E. Symposium Series No. 110 The Institution of

Chemical Engineers 1988, p 571, Figure 3.]

Page 46: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 40. Exothermic Reactions begin

Although widely contradictory accounts of the events have been given by the

operators, the proponents of the water wash theory believe that the following

occurred. At 10:20 P.M. the pressure within Tank E610 was reported at 2 psig.

At 11:00 P.M. the pressure was 10 psig as noted by another worker who had

come in on a shift change. He did not consider this abnormal at the time since

operating pressure was about 15 psig. By 12:15 A.M. this same worker noticed

that pressure had risen to between 25 and 30 psig. Within another 15 minutes

the pressure had exceeded 55 psig, and the worker ran to the tank. There he

found the tank with loud rumbling and screeching noises. Around 12:45 A.M.,

the relief system opened.35 As time continued, temperature within the tank

exceeded 200C. Average pressure within the tank exceeded 180 psig during

the release. The relief valve vent header system directed a stream of MIC to

the vent gas scrubber/flare system stack.

35. ASK p 559.

Page 47: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 41. Toxic Gas Release

The plant superintendent sounded the toxic gas alarm for the surrounding

community for a period of about 5 minutes and suspended operation of the

pesticide plant. Operators began futile efforts to spray water on the vent gas

stack (100 ft high); however, the water pressure was too low to reach the

emitting gases. The total release continued for two hours.

Page 48: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 42. The dispersion of the gas

Because of the slums built up around the plant, many thousands of people were exposed to MIC. Recall that the density of MIC as a vapor is two times that of air, thus it initially flowed towards ground level. As the MIC continued to be dispersed into the air the molecular weight of the mixture approached that of air and the MIC moved along with the prevalent air currents diffusing and dispersing into the surrounding environment. Even though the concentrations were very low (in the ppm levels), the mixture was still quite toxic as it moved across the UCIL property line. Secondly, this was a release during night time. At night the atmosphere is much calmer and sometimes inversions exist. Dispersion and mixing (dilution to concentrations below the threshold level) were hindered by the lack of atmospheric turbulence that normally occurs during daylight hours. The release blanketed an area of many square kilometers.36 Although accurate figures do not exist, more than 1,000 people were killed, and thousands more suffered injury. Hospitals and dispensaries were unable to cope with the flow of victims. The tragedy was profound. The toxic-gas alarm that was sounded for 5 minutes wasn’t heard by most. Further, many who heard the alarm didn’t understand its implication.

Instructors note the shape of some of estimates of the dispersion, circular in this case. Many dispersion models predict a long cigar shape. You may want to discuss dispersion modeling.

36. C&EN 10 Dec 1984 p 6.

Page 49: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 43. Findings after the accident

After the accident, the remaining contents of Tank E610 were determined. The

investigation concluded that 512 kg of water had entered the tank. This was

enough water to create enough temperature rise to promote additional catalytic

reactions including that of MIC to its trimer (a total of 6,938 kg remained).

Additional materials found included 2,660 kg of dimethyl isocyanate, 390 kg

of cyclic dione, (Slide 40) 195 kg of trimethylurea, 423 kg of trimethylamine,

240 kg of dimethylamine and small percentages of trimethylbiuret,

monomethylamine, dimethylurea, and salts of sodium, iron, chromium, and

nickel.37 (Tables 1 & 2 provide more detail). The presence of the iron, most

likely coming from corrosion inside the tank (metal analysis showed the

composition matched the tank walls),38 or less likely from back flushing water,

promoted the trimerization reaction. This reaction, being exothermic, heated

the tank contents to 200C, which caused high pressures to build and finally

the opening of the relief valve system and the release of MIC into the relief

valve vent header. After the pressure had subsided, the relief valve did

reseat.39

37. C&EN 6 Jan 1986, p 6.

Page 50: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

38. Information provided by Union Carbide to author, Mar 1998.

39. C&EN 10 Dec 1984, p 6.

Page 51: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern
Page 52: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 45. An alternative explanation --“a deliberate act”

As discussed above, the point of water entry for the filter process was 600 feet

away and had elevation points 10 feet above the addition point. After an

extensive investigation, Union Carbide has rejected the water wash theory as

scientifically untenable,40 contending the addition of water to the MIC tank

was “a deliberate act”. 41

40. C&EN 6 Jan 1986, p 6.

41. Chemical Week 26 Nov p 8.

Page 53: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 46. Points made against the filter line cleaning explanation

Ashok S. Kalelkar of Arthur D. Little, Inc. Cambridge, Mass., presented a paper in London in May of 1988 regarding the cause of the Bhopal incident.42 Three strong arguments against the water-washing theory were presented.

1. Bleeder Valve Hydraulics:

With 3 of the 4 bleeder valves functioning, backpressure before the valves could not exceed 0.7 foot of hydraulic head. A minimum of 10.4 feet was required to reach the RVVH.

2. Closed Intermediate Valve:

As explained above, all evidence points to Valve 19 being closed and leak tight at the time of the incident.

3. Dry Header Piping:

If the header piping system, RVVH and PVH, had entirely filled with water (as would be required if water entered from the wash station), then water should have been observed in any low points located through this piping arrangement. When these lines were inspected no water whatsoever was found. On Feb. 8, 1985, the Superintendent of Police of the Indian Central Bureau of Investigation ordered a hole drilled at the lowest point in the PVH line (the line that supposedly filled from the water washing). This point was well away from the MIC tanks and far enough away from the RVVH (which would have emptied upon the release) so it would have been unaffected by heat. The line was found to be bone dry.

Page 54: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

42. ASK pp 553 to 575.

Page 55: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 47. Water entered by direct connection.

Union Carbide’s investigation team found compelling evidence that water was deliberately introduced into the tank. First, a witness, an instrument supervisor, who reported that the pressure gauge labeled 10 was missing on the morning after the incident and the line was unplugged. He also testified that a hose with running water was near.

Deliberate acts of mischief by workers in industrial plants are not uncommon and had, in fact, occurred previously in the Bhopal plant. The investigation team identified an employee of the plant, working on the third shift that night, who was keenly disgruntled. Immediately prior to the incident, his supervisors demoted him and he openly expressed resentment against management. Investigators found he confessed to being near the tank at about the time water would have been introduced and had the motive, the means, and the opportunity to commit sabotage. The investigators found that he introduced the water during the shift change by removing the pressure indicator and temporarily opening Valve 16. He likely believed he would only spoil a tank of MIC. Although the accounts given by plant personnel are wildly contradictory, the investigation team concluded that at 11:30 to 11:45 P.M., workers on the plant floor detected minor MIC leaks and the source was incorrectly identified near the scrubber flare tower. The incorrect source was wetted down with a fire hose spray. After tea time (around 12:00 midnight), several operators noted the high pressure on Tank E610. Workers ran to the tank and discovered the hose attached. Several solutions were discussed including the transfer of material from Tank E610 in hopes of getting the water out before anything else occurred. The release happened around this time. There is evidence that those involved in the discussion of the transfer decided to a cover-up.

Page 56: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 48. Another trigger question for the slides that follow.

Page 57: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 49. Lessons learned (Three Slides)

1) Have contingency plans available for dealing with major accidents.

Currently, in the United States, emergency response exercises

for training purposes are required on a frequent basis. It is not unusual in large

plants to have staged minor events monthly and larger events every six

months. Larger events can include fire drills that shut down a major chemical

line.

2) Public involvement in risk management and acceptance is now recognized

as a critical element of chemical manufacturing. Public involvement in risk

management programs are required under “Right to Know” laws and other

initiatives that exist across the U.S.A. New chemical industry programs were

implemented to better manage health, safety, and environmental risks and

encourage more public involvement, such as the Chemical Manufacturers

Association establishment of Community Action Emergency Response

(CAER), and Responsible Care programs.

Page 58: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 50. Lessons Learned: Reduce the inventory of the hazardous chemicals

3) Another lesson to be learned from this incident is consideration of inventory reduction for process intermediates. What you don’t have, can’t leak, catch on fire, or cause any other problems.43

Could the process be designed that kept the amount of MIC to a minimum? This would require integrating the MIC production unit with the carbaryl production unit, such that when production of carbaryl was required, the MIC unit would begin production and produce only a small amount of the intermediate. A present design exists in which the maximum inventory of MIC is 10 kg.

Process safety analysis should be done early in the design stages. You need to ask the question: if hazardous intermediates are involved, can they be eliminated by alternative methods (see in the next slide) or, if possible, can these be minimized? Inherently safer plants would be toward the total elimination of the hazardous intermediate like MIC. For a SACHE lecture on inherently safer plants, see Kubias.44

43. T. Kletz, “Lesson from Disaster” Institution of Chemical Engineers, Rugby, Warwickshire CV21 3HQ, UK, 1993, p 83.

44. O. Kubias, “Inherently Safer Plants,” SACHE-AIChE 1996.

Page 59: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 51. Alternatives

Another possibility is to look at an alternative route to carbaryl that involves

less hazardous intermediates. One such route is the direct reaction of alpha

naphthol with phosgene to form alpha naphthol chloroformate. Alpha

naphthol chloroformate can then be reacted with methyl amine to give the

desired product carbaryl.45

45. C&EN 11 Feb 1985 p 30.

Page 60: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 52. Siting of the plant

The plant was originally designed to formulate pesticides. As it “backward

integrated,” it became more hazardous. Hazardous plants should be sited away

from populated areas and “green belts” established around them.

Page 61: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 53. Release mitigation

Release mitigation involves methods to lessen the effects of any release event

that may impact the surroundings.

1. Conduct toxic release modeling based on a potential release

scenarios.

2. Evaluate measures to mitigate, such as reduction of

inventory, spill containment, proper maintenance, detection by sensors, and

water sprays as a few examples.46

3. Emergency and safety training of operators.

4. Emergency response exercises with local community

services.

46. R.W. Prugh and R.W.Johnson, Guidelines for Vapor Release Mitigation

Page 62: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

AICHE, New York: 1988.

Page 63: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 54. Maintaining the integrity of safety systems

Safety systems, such as flares, scrubbers, refrigeration systems, relief devices,

or emergency shutdown devices, are critical to the safe operation of the

processes in which they are employed. Any changes affecting the integrity of

these systems must be managed carefully. Management of change procedures

are now required through regulation in the US.

Page 64: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 55. Risk management overseas

Another important concept is that risks must be managed in a consistent way

no matter where the plant is located around the world. U.S.-based

corporations must hold their U.S. locations to the same standard as those

outside the country. This applies both to the design and the operation of these

facilities.

Page 65: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 56. Employee threats of violence programs

Acts of violence and mischief in the workplace have increased in prevalence in

recent years. As a result, many companies have instituted programs to detect

disgruntled or potentially violent employees. Companies should adapt threat

of violence programs that include training and education, procedures to

recognize early indicators and procedures to manage threats once they have

occur. In recognition of the problem, OSHA has begun work on a workplace

violence standard.47

47. For more details see the following web sites that were active in 1998:

“http://www.osha-slc.gov/SLTC/WorkplaceViolence/index.html” and

“http://www.osha-slc.gov/workplace_violence/wrkplaceViolence.Table.html”

Page 66: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 57. Other outcomes after the incident

As news of the incident spread, several events occurred. The chairman of Union Carbide, Warren M. Anderson, traveled to Bhopal. He was arrested by the Madhya Pradesh authorities; then released at the end of the day. He then returned to New Delhi.48 Within in a week of the incident, Union Carbide set up a $1 million fund to assist in direct aid at Bhopal. The remaining MIC located in Tank E611 was converted to pesticide two weeks later under very careful precautionary means that included a helicopter spraying water on the plant.49 The Indian Government termed the processing “Operation Faith”.50 Before the disposal, tens of thousands of remaining Bhopal residents fled the city in fear of another release from the plant. In the ensuing months, Union Carbide gave another $5 million to the Indian Red Cross and donated additional millions of dollars to other relief organizations for humanitarian work in Bhopal.

Union Carbide settled with the Indian government for $470 million in 1989 and contributed an additional $20 million for the construction and operation of a new hospital. In the fall of 1994, Union Carbide sold its share in UCIL, and about $74 million from the sale went into the hospital’s trust fund.

48. NY Times 8 Dec 1984 p 1.

49. NY Times 17 Dec 1984, p A 8.

50. ASK p 556.

Page 67: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 58. U.S. legislative initiatives.

A number of U.S. legislative initiatives were instituted and aim at prevention

of such an incident in the future. These include Process Safety Management

Standard, OSHA 29CFR1910.119, and EPA’s Risk Management Program Rule,

40CFR68.

Page 68: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 59. Additional Initiatives

Page 69: SEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY - AIChE · PDF fileSEMINAR ON THE BHOPAL DISASTER: A CASE HISTORY Prepared by Ronald J. Willey Department of Chemical Engineering Northeastern

Slide 60. Acknowledgments

This slide module was prepared from materials originally referenced by Dr.

Walt Howard. The author is thankful to have had this list and this listing is

presented in a portion of the index for those who wish to read further. The

author acknowledges: Ms. Cristy Godoy de Urruela for collecting the many

citations related to the incident and creating the tables that are attached as an

appendix to this module; Mr. Wilbert C. Lepkowski, Senior Correspondent,

Chemical & Engineering News, for several original photographs used in this

work; and Mr. Bob G. Perry for a copy of the Arthur D. Little paper presented

in 1988 that assisted in understanding the plant process properly. Finally, the

author acknowledges the SACHE committee whose reviews of text and slides

are gratefully accepted.