This Report is a translation of the original; “Chosa-Houkokusho” issued in March 2013 written

in Japanese, for convenience purpose only, and the original in Japanese shall prevail.

Nippon Shokubai Co., Ltd. Himeji Plant

Explosion and Fire at Acrylic Acid Production Facility

Investigation Report

March 2013

Accident Investigation Committee

Nippon Shokubai Co., Ltd.

Contents Pages

１．Preface 1

1.1. Introduction 1

1.2. Accident Investigation Committee 1

1.2.1. Mission and Members of the Committee 1

1.2.2. Progress of the Committee Meetings 2

２．Overview of the Accident 3

2.1. Location and Equipment Involved 3

2.2. Date and Time of Accident 3

2.3. Weather Conditions 3

2.4. Damages 3

2.4.1. Casualties 3

2.4.2. Property Damages 3

2.4.3. Environmental Impact 3

2.5. Situation after the Accident 3

３．Overview of the Himeji Plant and Equipment Involved 4

3.1. Overview of the Himeji Plant 4

3.2. Overview of Acrylic Acid Production Process 5

3.2.1. Properties of Acrylic Acid 5

3.2.2. Acrylic Acid Production Process 6

3.3. Overview of Equipment Involved 8

４．Development of the Accident 9

4.1. Events Leading to the Explosion and Fire 9

4.1.1. Events Timeline 9

4.1.2. Sequence of Events 11

4.2. Damages Resulted from the Explosion and Fire 16

4.2.1. Explosion Impact and its Damages 16

4.2.2. Fire Effect and Damages 27

５．Determining the Causes of the Accident 29

5.1. Contributing Factors of the Explosion and Fire 29

5.1.1. Direct Causes of the Accident 29

5.1.2. Contributing Factors of the Accident 29

5.1.3. Background of Accident Contributing Factors 31

5.2. Matters Related to Expanded Damage Resulted from the Explosion and Fire 42

６．Recommendations for Accident Recurrence Prevention Countermeasures 43

6.1. Recurrence Prevention Countermeasures for Equipment Involved 43

6.2. Countermeasures to Prevent Recurrence of Similar Accidents and its Lateral

Deployment 44

6.2.1. Countermeasures to Prevent Recurrence of Similar Accidents 44

6.2.2. Disaster Prevention Measures for Equipment Using Acrylic Acid 45

6.2.3. Lateral Deployment of the Accident Prevention Countermeasures 46

6.3. Fostering a Safety Culture of Safe Manufacturing Plant and Corporation 46

Afterword 62

Appendixes

1. Preface

1.1. Introduction

At about 14:35 on September 29, 2012, an explosion occurred at the Nippon Shokubai Co., Ltd.

Himeji Plant located in Himeji, Hyogo Prefecture, Japan. The explosion and subsequent fire in an acrylic

acid intermediate tank killed one person and injured 36.

As a result, on October 5, 2012, an Accident Investigation Committee consisting of four external and

three internal members was established with the objectives of determining the causes of the accident and

developing countermeasures to prevent the recurrence of accident.

The committee has thus far convened seven times. The committee has inspected the accident site,

verified records, various experimental data, analytical results, testimony etc. provided by Nippon

Shokubai’s Himeji Plant and discussed the causes of the accident as well as countermeasures. The

investigation has confirmed the causes of the accident that led to the explosion and fire and hence proposed

countermeasures to prevent the recurrence of accident. With these, the committee has prepared this final

report.

1.2. Accident Investigation Committee

1.2.1. Mission and Members of the Committee

From a fair standpoint, the committee aimed to find out facts and causes which led into the accident

and proposed countermeasures to prevent the recurrence of accident. The Accident Investigation

Committee consists of four external experts and three company employees who are responsible for safety

and production. The members of the committee are:

Chairman: Masamitsu Tamura, Emeritus Professor, The University of Tokyo

Members: Yasukazu Arai, Former Managing Director, The High Pressure Gas Safety Institute of

Japan

Mitsuo Koshi, Research Professor, Institute of Engineering Innovation, School of

Engineering, The University of Tokyo

Masayoshi Nakamura, Professor, The Graduate School of Technology Management,

Tokyo University of Agriculture and Technology

Yosuke Ogata, Representative Member of the Board, Senior Managing Executive

Officer, Nippon Shokubai Co., Ltd.

Masao Kitano, Director, Safe Production Technology Division, Nippon Shokubai Co.,

Ltd.

Hiroya Kobayashi, Director, Responsible Care Division, Nippon Shokubai Co., Ltd.

1

1.2.2. Progress of the Committee Meetings

1st Meeting October 23, 2012 (Tuesday) at the Himeji Plant, Nippon Shokubai Co., Ltd.

i. Finalized the organization of the Accident Investigation Committee

ii. Confirmed the overview of accident and inspected the accident site

iii. Deliberated on the investigation approach for the causes of accident

2nd Meeting November 15, 2012 (Thursday) at the Tokyo Office, Nippon Shokubai Co., Ltd.

i. Confirmed the accident timeline and V-3138 operation history.

ii. Sorted out the damages caused by explosion and fire

iii. Deliberated on the causes of accident

3rd Meeting December 4, 2012 (Tuesday) at the Tokyo Office, Nippon Shokubai Co., Ltd.

i. Deliberated on the causes of accident (cont.)

ii. Estimated the accident scenario and the direct causes

iii. Deliberated on direction of countermeasures for the direct causes

4th Meeting December 25, 2012 (Tuesday) at the Himeji Plant, Nippon Shokubai Co., Ltd.

i. Finalized the accident scenario and the direct causes

ii. Deliberated on countermeasures for the direct causes

iii. Deliberated on the investigation approach for the background causes

The committee has issued the “Accident Investigation Committee Interim Report” on January 18,

2013 (Friday) at the Osaka Office, Nippon Shokubai Co., Ltd.

i. Explosion and Fire at the Himeji Plant

5th Meeting January 25, 2013 (Friday) at the Himeji Plant, Nippon Shokubai Co., Ltd.

i. Summarized the accident scenario and the direct causes

ii. Deliberated on the background causes

iii. Deliberated on the directions for recurrence prevention countermeasures

6th Meeting February 14, 2013 (Thursday) at the Tokyo Office, Nippon Shokubai Co., Ltd.

i. Deliberated on the background causes

ii. Deliberated on the recurrence prevention measures

iii. Confirmed the ‘Table of Contents’ of the final report

7th Meeting February 22, 2013 (Friday) at the Tokyo Office, Nippon Shokubai Co., Ltd.

i. Finalized the background causes

ii. Finalized the recurrence prevention measures

2

2. Overview of the Accident

At about 14:35 on September 29, 2012, an explosion and fire occurred in an intermediate tank

(equipment item number: V-3138; nominal capacity: 70 m3) that temporarily stored bottom liquid from the

glacial acrylic acid rectifying column at the Nippon Shokubai’s Himeji Plant. The fire then spread to the

nearby equipment and buildings such as acrylic acid tanks, toluene tank and fire engines.

2.1. Location and Equipment Involved

(1) Location

Himeji Plant, Nippon Shokubai Co., Ltd.

992-1 Aza Nishioki, Okihama, Aboshi-ku, Himeji, Hyogo, Japan

(2) Equipment involved

An intermediate tank (V-3138) in the acrylic acid production facility

2.2. Date and Time of Accident

September 29, 2012 (Saturday) at about 14:35

2.3. Weather Conditions

Weather: Cloudy Temperature: 24~25°C Pressure: about 1010 hPa,

Humidity: about 60% Wind direction: WSW Wind speed: 2~3 m/sec

2.4. Damages

2.4.1. Casualties

Fatalities: 1 (firefighter)

Severely injured: 5 (2 firefighters and 3 employees)

Moderately injured: 13 (8 firefighters, 1 police officer and 4 employees)

Lightly injured: 18 (14 firefighters, 1 police officer and 3 employees) Total: 37 persons

2.4.2. Property Damage

The tank, V-3138 was destroyed and its surrounding equipment, racks, piping, cables, etc. were also

damaged.

2.4.3. Environmental Impact

As the outlet drain was closed, no hazardous materials were leaked out from the plant site.

2.5. Situation after the Accident

The authority temporarily issued a ‘Stop Work Order’ and hence all the facilities involved hazardous

materials were stopped. Since then, some of the facilities have resumed to operation after the authority’s

approval.

3

4

3. Overview of the Himeji Plant and Equipment Involved

3.1. Overview of the Himeji Plant

The site of the accident, Nippon Shokubai’s Himeji Plant, is a manufacturing plant with a site area of

about 900,000 m², situated in the southwest corner of the city of Himeji, Hyogo Prefecture. The Himeji

Plant produces basic chemicals (e.g. acrylic acid and acrylic esters), functional chemicals (e.g. super

absorbent polymers, electronic and information materials), catalysts and other products.

The acrylic acid intermediate tank (V-3138), where the explosion and fire occurred, is located in an

acrylic acid production facility in the basic chemicals production yard shown in Figure 3-1.

(Figure 3-1) Overview of the Himeji Plant, Nippon Shokubai

3.2. Overview of Acrylic Acid Production Process

3.2.1. Properties of Acrylic Acid

Acrylic acid is an organic compound with the chemical formula CH2=CHCOOH. It is a colorless,

transparent, flammable liquid having an irritating odor. Under the Japan’s Fire Service Act, acrylic acid is

classified as a Category IV, Class II Petroleum flammable liquid.

The main physical and chemical properties of the acrylic acid are as follows:

Density : 1.05 g/cm3 (20℃)

Melting Point : 13.5 ℃

Boiling Point : 141 ℃

Viscosity : 1.25 mPa･s

Ignition Point : 428 ℃

Flash Point : 51.4 ℃

Explosion Range : 2～17 vol%

Specific Heat : 2093 J/kg･℃

Heat of Dimerization : 145.3 kJ/kg

Heat of Polymerization : 1076 kJ/kg

Conductivity : >10-7 S/m

Reactivity : Due to its unstable double bond, acrylic acid is readily converted

into a dimer (DAA) via dimerization reaction as shown in Figure

3-2 and to polymerize through polymerization reaction shown in

Figure 3-3. In order to inhibit the polymerization, it is usually kept

in an atmosphere containing at least 5vol% of oxygen and added

with inhibitors.

(Figure 3-2) Chemical equation of diacrylic acid (DAA) formation

(Figure 3-3) Chemical equation of acrylic acid polymerization

5

3.2.2. Acrylic Acid Production Process

The acrylic acid production process has two production processes: crude acrylic acid production

process and glacial acrylic acid production process.

(1) Crude acrylic acid production process

The crude acrylic acid production process consists of oxidation process and purification process. In the

oxidation process, propylene is converted to acrylic acid vapor via vapor phase oxidation reaction. The

acrylic acid vapor is subsequently absorbed with water and formed the acrylic acid aqueous solution. In the

purification process, crude acrylic acid is obtained by separating water and other impurities from the acrylic

acid aqueous solution.

(2) Glacial acrylic acid production process

The glacial acrylic acid production facility consists of a rectifying column and a recovery column.

The crude acrylic acid is fed into the rectifying column where the impurities contained are removed

and obtained as the glacial acrylic acid.

The rectifying column bottom discharge liquid which contains the removed impurities, is fed into the

recovery column. In the recovery column, the impurities are discharged as waste oil and the recovered

acrylic acid is used as crude acrylic acid.

Inhibitors contained in the crude acrylic acid and inhibitors fed into the rectifying column are

concentrated in the rectifying column bottom liquid. This bottom liquid does not polymerize easily because

it contains more inhibitors than glacial acrylic acid.

Intermediate tank V-3138 was installed in the crude acrylic acid production facility and was used for

temporary storing the rectifying column bottom liquid.

(Figure 3-4) Schematic of the acrylic acid production facility

6

(3) Configuration of processes

The acrylic acid manufacturing plant has crude acrylic acid production facility and glacial acrylic acid

production facility as shown in Figure 3-5.

The crude acrylic acid production facility has four production lines (2AA to 5AA). Each production

line was constructed (completed) in the years stated in the following:

2AA: Y1973, 3AA: Y1985, 4AA: Y1990, 5AA: Y1995

The glacial acrylic acid production facility has five series of rectifying column (T-2108~T-6108) and

two series of recovery column (T-2701 & T-6701). The construction (completed) years of each series was

stated in the following:

<Rectifying column> T-2108: Y1985 T-3108: Y1985 T-4108: Y1990

T-5108: Y1994 T-6108: Y1998

<Recovery column> T-2701: Y1985 T-6701: Y1998

V-3138 is the only intermediate tank that is used to temporarily store rectifying column bottom liquid.

V-3138 was constructed (completed) in 1985.

(Figure 3-5) Configuration of acrylic acid production processes

7

3.3. Overview of Equipment Involved

The equipment involved was the acrylic acid intermediate tank V-3138, which had a nominal capacity

of 70 m3. It was an insulated (75mm thick) cone roof type storage tank. The tank was installed in

November 1985. It is used as an intermediate tank for temporarily storing the withdrawal liquid from the

rectifying column when, for example, the rectifying column stopped.

During normal operation, the bottom liquid is fed directly into the recovery column without passing

through V-3138 and therefore the intermediate tank is kept stagnant.

There was a cooling water coil inside V-3138 that serves to prevent the freezing of acrylic acid and

cooling the liquid that was fed into V-3138. The amount of liquid necessary to fully cover the top of the coil

is 25 m3.

Although acrylic acid is a flammable liquid, its vapor does not burn when it has an oxygen

concentration of 8% by volume or less. Therefore, a mixed gas (referred to as “M-Gas”) consisting of 7%

oxygen and 93% nitrogen by volume is fed into the tank for sealing purpose.

Liquid stored in V-3138 is circulated through pump P-3138C back into the same tank at two locations:

liquid level gauge nozzle set near the lower side of V-3138’s wall (referred to as “Recycle to Level Gauge”)

and nozzle set at V-3138’s top (referred to as “Recycle to Top”).

Initially when the V-3138 was installed in 1985, there was only the Recycle to Top line for pump

circulation. In order to prevent the precipitates accumulation and results in incorrect reading at the level

gauge, the Recycle to Level Gauge line was installed in 2009.

(Figure 3-6) Schematic of tank V-3138

8

4. Development of the Accident

4.1. Events leading to the Explosion and Fire

4.1.1. Events Timeline

The timelines for the explosion and fire accident is shown below:

September 18 to 20, 2012

Total power shutdown in the Himeji Plant for electrical and instrumentation

maintenance work (referred to as “Total Power Shutdown Work”)

September 20, 2012

Approx. 21:00 Intermediate tank V-3138 was re-commissioned after Total Power

Shutdown Work

(Cooling water charged into the coil, commissioned M-Gas seal, started

P-3138C)

September 21, 2012

Approx. 11:00 to Built up liquid level in recovery column (T-6701) from V-3138 and started

approx. 14:00 T-6701 operation. Subsequently, started rectifying column (T-6108) operation

and fed its bottom liquid directly into T-6701

(After building up liquid level in T-6701, V-3138 liquid volume reduced from

about 14 m³ to about 10 m³)

September 24, 2012

Approx. 10:00 Started rectifying column (T-5108) operation

Approx. 14:10 Started discharging T-5108 bottom liquid (fed to T-6701 via V-3138)

September 25, 2012

Approx. 9:30 Built up the V-3138 liquid level (stopped feeding T-6701 from V-3138)

September 28, 2012

Approx. 14:00 Liquid volume in V-3138 reached 60 m³ and then stopped feeding V-3138.

(T-5108 bottom liquid switched to feed T-6701 directly).

September 29, 2012

Approx. 13:17 V-3138 liquid level high alarm triggered

Approx. 13:20 Discovered white smoke emitting from V-3138’s vent

Approx. 13:25 Operators started spraying water onto V-3138

Approx. 13:40 Shift Leader made the plant-wide announcement and alerted self disaster

9

prevention team.

Reading of V-3138 liquid level gauge exceeded out-of-range limit (84.8 m3)

Approx. 13:48 to 13:49 Disaster prevention section notified public fire department through the

hotline.

Approx. 14:00 Self disaster prevention team started spraying water onto V-3138

Approx. 14:02 Public fire department arrived on site and cordoned off the accident site.

Until approx. 14:35 V-3138’s contents leaked out from cracks

Approx. 14:35 Reading of V-3138 liquid level gauge dropped drastically

V-3138 liquid level low alarm triggered

V-3138 exploded and ruptured

Contents of V-3138 ignited and resulted in fire

22:36 Fire contained

September 30, 2012

15:30 Fire extinguished

10

4.1.2. Sequence of Events

The committee has divided the sequence of events leading up to the accident into four stages and

investigated the accident details and causes.

(1) Before started the storing operation in V-3138

(From September 21 to about 9:30 on September 25, 2012)

After the Total Power Shutdown Work was completed, cooling water was commissioned to pass

through V-3138’s coil and similarly the M-Gas sealing was also commissioned. Circulation from pump

P-3138C via Recycle to Level Gauge was commissioned too.

There was no abnormality observed in the operations of the crude acrylic acid production facility. The

correct amount of inhibitors was also fed into the crude acrylic acid obtained from the purification process

in accordance with the operation standards.

Operations of the rectifying column (T-6108) and the recovery column (T-6701) in the glacial acrylic

acid production facility were started on September 21. The bottom liquid of T-6108 was fed directly into

T-6701.

On September 24, another rectifying column (T-5108) was started. The bottom liquid of T-5108 was

fed to the T-6701 via V-3138. During this operation, the liquid volume in V-3138 was maintained

constantly at about 10m3.

The transfer piping of T-5108 bottom liquid was steam jacketed to prevent plugging due to

precipitation. The temperature of the bottom liquid as it entered V-3138 was estimated at about 100°C,

based on steam temperature and the length of the jacketed piping.

Inhibitors were fed into T-5108 in accordance with the operation standards.

(Figure 4-1) Status of tank V-3138 before liquid storing

11

(2) During storing operation in V-3138

(From about 9:30 on September 25 to about 14:00 on September 28, 2012)

At about 9:30 on September 25, the feed of bottom liquid from V-3138 to T-6701 was stopped and

commenced to build up the liquid volume in V-3138. This was to prepare a capacity load up test in the

recovery column which was scheduled at a later date.

There was no particular change in the crude acrylic acid fed to T-5108 or in the inhibitors. The

operation was running normally according to the operation standards.

The volume of liquid in V-3138 reached the planned 60m3 at about 14:00 on September 28,

approximately 77 hours after the commencement of storage. During this period, the circulation of Recycle

to Top was not commissioned. Figure 4-3 shows the V-3138 liquid level trend during storing operation.

The coil cooled the liquid in the bottom of V-3138 but was unable to cool the liquid above the top of the

coil effectively. This had created an uneven temperature distribution in the vertical direction of V-3138

liquid. It is presumed that this relatively high temperature section has led to the reaction of forming acrylic

acid dimer (diacrylic acid, referred to as “DAA”). With the heat of this reaction, V-3138 liquid temperature

has increased steadily. However, V-3138 was not equipped with any thermometer and it was not possible to

detect the temperature has increased.

(Figure 4-2) Status of tank V-3138 during liquid storing

(Figure 4-3) Trend of V-3138 liquid level gauge (until liquid storing completed)

12

(3) After storing operation in V-3138

(From about 14:00 to about 14:10 on September 28, 2012)

At about 14:00 on September 28, the liquid volume in V-3138 reached 60m3. Thereafter, T-5108

bottom liquid was switched back to feed T-6701 directly without passing through V-3138.

Even then, the circulation of Recycle to Top was still not commissioned. Consequently, the liquid

above the top of the coil was not cooled and remained at a relatively high temperature. In order to simulate

the temperature distribution in the V-3138 liquid, a 3-D model of V-3138 was developed. The ANSYS

Fluent Ver.13 was used to conduct the fluid analysis without taking the heat of reaction into consideration.

The analysis showed that if Recycle to Top was not commissioned, the liquid just below surface remained

at a relatively high temperature, average about 87°C immediately after reaching a volume of 60m³. The

analysis results are shown in Figure 4-5.

(Figure 4-4) Status of tank V-3138 after liquid storing

温度[℃]

温度[℃]

Actual condition model Reference

(Without Recycle to Top) (With Recycle to Top)

(Figure 4-5) Results of fluid analysis when V-3138 liquid has reached 60m³

13

(4) From liquid holding in intermediate tank (V-3138) to explosion and fire

(From about 14:10 on September 28 to about 14:35 on September 29, 2012)

It is presumed that the reaction of DAA formation accelerated in the liquid stored in V-3138 and the

liquid temperature continued to increase from the heat of DAA formation. In a laboratory experiment

conducted under adiabatic condition, the reaction was confirmed to proceed until the composition has

reached to approximately 40wt% of acrylic acid (AA), 60wt% of DAA and others. If the reaction is allowed

to proceed under adiabatic conditions up till these compositions, the amount of reaction heat is able to

increase the temperature of the V-3138 entire liquid volume by about 40°C.

It is also presumed that this continued rise in temperature has started the polymerization of acrylic acid

and the temperature of V-3138 liquid rose even more rapidly by the heat of polymerization.

At about 13:20 on September 29, an operator discovered white smoke (acrylic acid vapor) coming out

from V-3138 vent. It is estimated that the high temperature section of V-3138 liquid was about 160°C at

that time. This temperature was estimated by confirming water vapor is visible as white smoke in the

laboratory test and then by using Aspen Plus to confirm the vapor pressure with the composition of the

V-3138 liquid.

The V-3138 liquid level gauge reading (pressure differential type) has shown increasing trend. It

started to rise at about 13:20, reached the maximum liquid volume setting at about 13:30 and exceeded the

gauge’s out-of-range limit (84.8m3) at about 13:40. This trend has shown the V-3138 internal pressure has

increased. The increasing pressure is presumed to have occurred because the vent’s discharge capacity has

exceeded by the increased amount of evaporated vapor resulted from accelerated polymerization. Figure

4-6 shows the V-3138 liquid level trend.

The V-3138 pressure continued to rise thereafter and cracks appeared at the shell plate. It is estimated

the pressure had built up to 0.24~0.29MPaG based on tank structural analysis at the time the cracks

appeared. Independently, based on the adiabatic polymerization experiment data, the V-3138 internal

pressure was also estimated to reach about 0.27MPaG. The estimated trend of increased pressure is shown

in Figure 4-7 and the model used for estimation is explained in Appendix 2. Also, the V-3138 liquid was

estimated to be about 230~240°C in the high temperature zone. The reaction data of the adiabatic reaction

test is explained in Appendix 1 and the structural analysis results will be discussed in Section 4.2.1.

The V-3138 contents started to leak from the cracks and caused the V-3138 pressure dropped

drastically.

At about 14:35 on September 29, the V-3138 liquid started to boil violently due to the drastic drop in

pressure, while the liquid temperature still remained at high temperature. These resulted in Boiling Liquid

Expanding Vapor Explosion (BLEVE) inside V-3138 and subsequently ruptured the V-3138. The scattered

contents ignited and caused the fire. Based on the distance of scattered debris, it is estimated that the tank

internal pressure at the time of explosion was about 0.45~0.64MPaG. The possible ignition sources were

sparks generated from the impact of metals or from broken electrical wires during the explosion.

The explosion has also damaged the nearby tanks. Thus, the leaked acrylic acid and toluene from these

tanks caused the fire to spread further. Table 4-1 summarized the major events from V-3138 liquid reached

its boiling point till the explosion and fire occurred.

14

(Figure 4-6) Trend of V-3138 liquid level gauge (until the explosion)

(Figure 4-7) V-3138 liquid level gauge readings (converted to pressure) and pressure rise

(Table 4-1) Estimated sequence of events for V-3138 fire and explosion

# All internal pressure and temperature values are estimated

Sequence of events Time V-3138 Internal

Pressure [MPaG] V-3138 Internal

Temperature [℃] V-3138 internal pressure started to rise ~13:20 Atmospheric (160~170) Exceeded the liquid level gauge out-of-range limit

~13:40 (0.025) (175)

Self disaster prevention team started water spraying from fire engine (3,100 L/min)

~14:00 - -

V-3138 anchor bolts broken - (0.1) - Cracks initiated at outlet nozzle of cooling water coil (nozzle position: 270°)

Till ~14:35 (0.24~0.29) (234)

Cracks propagated and contents leaked out (at 270°, vertical directions)

Till ~14:35 (0.24~0.29) (234)

Cracks reached vapor phase parts and tank pressure dropped drastically

~14:35 (Atmospheric) (234)

Boiling Liquid ExpandingVapor Explosion ~14:35 (0.45~0.64) (234) Fire (acrylic acid: ~66 m³, toluene: ~28 m³) ~14:35 - - Fire contained ~22:36 - -

15

4.2. Damages Resulted from the Explosion and Fire

4.2.1. Explosion Impact and its Damages

(1) Equipment damage

<Condition of V-3138 tank and its surrounding equipment>

The Figure 4-8 shows the ruptured V-3138 resulted from explosion and the damage condition of its

surrounding equipment.

(1) Damaged Tank （view from southwest） (2) Zoom-in of Damaged Tank （view from southwest）

(3) Damaged Tank（top view from east） (4) Piperack at west of damaged tank （view from southeast）

(5) Piperack at west of damaged tank （view from south） (6) Pumps surrounding at west of damaged tank （view from south）

(Figure 4-8) Damages around V-3138

16

<Damage condition of V-3138>

The damaged conditions of the V-3138 main members are described below. Figure 4-9 also shows the

reconstruction figure of these members with other scattered fragments.

i. Roof Plate

The roof plate broke and separated near the roof-to-shell joint. The roof plate has split into two

and being thrown to a distance of about 50m.

ii. Shell Plate

The shell plate tore vertically in a near straight line at 270° position, deformed into casement-type

opening and left almost like a flat plate within the dike at the accident site.

iii. Bottom Plate

The bottom plate partly broke near the shell-to-bottom joint and remained inside the dike

Having observed the fracture surfaces of V-3138 members, most of them are slanted to the

through-thickness direction and have features of being formed through shear fracture caused by glide plane

decohesion. However, the fracture surface around the outlet nozzle of cooling water coil consisted of two

types of fracture surfaces: ductile fracture formed through tensile deformation and shear fracture. This part

was judged as a main initiation point of the cracks. The observation results of the fracture surfaces are

explained in Appendix 3.

The cracks formed in this part propagated and split the tank shell plate vertically in both directions at

270° position. The upwards crack broke into the roof plate and ran through it near the 135° position. The

roof and bottom plates generally cracked in the circumferential direction near their joints with the shell

plate. Figure 4-10 shows the estimated crack propagation paths of the roof and shell plate in development

view.

<Damage conditions of surrounding equipment>

The damage conditions of the surrounding equipment are shown in Figure 4-8. Six tanks which were

located within the same dike as V-3138 were severely damaged together with several pumps installed near

the west side of the dike. The six tanks were V-3116 (100m³: acrylic acid tank), V-3124 (50m³: toluene

tank), V-3145B, V-3147B, V-3147C, and V-3148C (5m³ each: inhibitor tanks). Furthermore, building

(electric room) at the south side of the dike, nearby piperacks, piping, cables and fire engines were severely

damaged also.

<Scattered conditions>

About 80 pieces of broken fragments from V-3138 and surrounding equipment scattered within 100m

radius. The tank contents were also scattered within 70m radius, particularly in the west direction of shell

plate’s opening (270° position).

17

(Figure 4-9) Reconstruction diagram (sketch) of V-3138 fragments

18

(Figure 4-10) V-3138 crack initiation and estimated propagation paths (top/ shell plates development view)

Crack diverted to shell plate

Propagated along shell-to-bottom joint

Vertical crack in the shell plate near the

270° position broke into the roof plate

Secondary damage caused by steel beam piercing

Independent cracks

Other initiation points

Crack broke from the roof-to-shell joint Crack broke into roof plate

Zoom-in of cooling water outlet nozzle One of the main initiation points

19

(2) Internal pressure at the time V-3138 ruptured by explosion

The internal pressure at the time tank ruptured by explosion was estimated by: i. the tank structural

analysis and ii. the distance of scattered fragments.

i. Estimation results of tank internal pressure by structural analysis

The V-3138 structural analysis used the finite element method to determine the tank internal pressure.

The Figure 4-11 shows the 3-D model developed for structural analysis. The model divided the tank into

100,000 elements for analysis. Main members of the modeled tank are shown in Figure 4-12.

Based on the structural analysis, the anchors were estimated to break at about 0.1MPaG and the

pressure which initiated the cracks at the tank was estimated to be 0.24~0.29MPaG. The generic nonlinear

structural analysis code, LS-DYNA was used for the analysis code.

The analysis of two predicted failure points was described as follows:

a) Anchor points

It is thought that as the tank internal pressure increased, the tank was uplifted but the tendency was

restrained by the anchors which created a large resistant force locally. This analysis was conducted by

using the 3-D model.

Figure 4-13 shows the equivalent plastic strain on the tank surface caused by the tank internal

pressure. Figure 4-14 shows the vertical reaction force acting on the anchors as a result of the tank

internal pressure. The lower limit for the tank pressure to create the anchor breaking load of 1005kN was

0.073MPaG. The tank pressure at which the vertical reaction force exceeds the breaking load was

estimated to be about 0.1MPaG. However, the strain on the shell plate at that time was about 2~3% which

was not a level rupture would occur at the shell plate.

b) Cooling water coil outlet nozzle

The thickness of the ruptured V-3138 shell plate measured at about 2.8mm which was the same

thickness as at the time of manufacture. Because of this, it was presumed that plate thickness reduction,

which begins when the elastic limit of the material is exceeded, had not occurred. The pressure at which

plate thickness reduction would become noticeable was estimated at about 0.25~0.30MPaG through

structural analysis using the 3-D model. Figure 4-15 shows the results of plate thickness reduction against

the tank internal pressure.

Presuming the cracks initiated before reaching this pressure, the structural analysis was conducted

focusing on the tank’s weak points. These included the nozzles of cooling water coil, roof-to-shell joints,

shell-to-bottom joint and the edges of the tank. The results have confirmed a larger strain appeared

locally at the outlet nozzle of cooling water coil than its surrounding. Figures 4-16 and 4-17 show the

analysis results.

In order to improve the analytical accuracy around this area, a local shell model was used to zoom-in

the nozzle vicinity for estimating the pressure at which the plastic deformation increase indefinitely

would occur. The pressure was estimated to be about 0.24~0.29MPaG. Figure 4-18 shows the results. The

reason why the cracks appeared in this area is presumed due to a structural discontinuity at which a

flange with a diameter of about 150mm and a plate thickness of about 8mm was attached to the shell

plate. It is presumed that excessive stress and strain were developed near the structural discontinuity.

20

剛体平面

(Figure 4-11) 3-D model used in structural analysis

(Figure 4-12) 3-D model of tank constituent components

Flange near 270 º position withdiameter of 150mm and 8mm platethickness

°

Anchor points. Some of the bottomplate thickness has been increased

Top Ring

Rafter

Anchor points (8 locations)

（１）Bird’s Eye View

（２）Shell Plate（Front: 270º position） （2）Roof

Rigid plane

21

0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.00 0.05 0.10 0.15 0.20 0.25

相当塑性歪

内圧（MPaG）

60℃

200℃

400℃

胴板のひずみ

P=0.073MPaG

84

2

b

dR

4

2

PD

F

アンカー破断荷重

圧力による荷重

d:アンカー直径(20mm)b: SS400破断強度(400MPa)

D: タンク円筒直径(4200mm)P: 内圧

算定式

(Figure 4-13) Equivalent plastic strain on the tank surface caused by the tank internal pressure

0

200

400

600

800

1000

1200

1400

1600

1800

2000

0.00 0.05 0.10 0.15 0.20 0.25

アンカー部の鉛直反力（kN）

内圧（MPaG）

60℃

200℃

400℃

1005kN

注意：アンカー8本分の合計値をプロット

(Figure 4-14) Vertical reaction force acting on the anchors as a result of the tank internal pressure

Tank internal pressure (MPaG)

Equ

ival

ent p

last

ic s

trai

n

Tank internal pressure (MPaG)

Note: total value of eight anchor points plotted

Anc

hors

’ ver

tical

rea

ctio

n fo

rce

(kN

)

22

1.0

1.2

1.4

1.6

1.8

2.0

2.2

2.4

2.6

2.8

3.0

3.2

0.0 0.1 0.2 0.3 0.4 0.5 0.6

胴板各部の板厚（m

m）

内圧（MPaG）

64254

51075

38946

26425

0.47MPaGで不安定化

(Figure 4-15) Plate thickness reduction against the tank internal pressure by 3-D model (at 200°C)

(Figure 4-16) Strain analysis by 3-D model (at 200°C)

0.2 MPaG 0.3 MPaG

Tank internal pressure (MPaG)

She

ll p

late

thic

knes

s in

eac

h se

ctio

n (m

m)

Destabilization at 0.47MPaG

23

(Figure 4-17) Strain analysis by 3-D model (at 200°C)

Zoom-in view of cooling water coil outlet nozzle

1.8

2.0

2.2

2.4

2.6

2.8

3.0

3.2

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.0 0.1 0.2 0.3 0.4 0.5

板厚（ m

m）

相当塑性歪

圧力換算値（MPaG）

60deg

200deg

400deg

60deg

200deg

400deg

相当塑性歪

板厚

板厚の参照要素歪の参照要素

(Figure 4-18) Analysis results of cooling water outlet nozzle vicinity

Zoom-in view of cooling water coil outlet nozzle at 0.2MPaG

Zoom-in view of cooling water coil outlet nozzle at 0.3MPaG

Tank internal pressure (MPaG)

Equ

ival

ent p

last

ic s

trai

n

Pla

te th

ickn

ess

(mm

)

Plate thickness

Equivalent plastic strain

Plate thickness reference element

Strain reference element

24

ii. Estimation results of tank internal pressure at ruptured by distance of V-3138 scattered fragments

From the scattered distance of V-3138 roof plates, deck and vent pipe, fragments’ minimum initial

velocity is determined and converted into Mach number. As a result the tank internal pressure at ruptured

was estimated to be 0.45~0.64MpaG.

(Table 4-2) Estimated tank internal pressure at ruptured by distance of V-3138 scattered fragments

Fragments Scattered Distance

Weight Exposed

Area Minimum Initial

Velocity Tank Internal

Pressure Roof plate – 1 52m 144kg 2m2 28.4m/s 0.49MPaG Roof plate – 2 49m 540kg 11.1m2 27.4m/s 0.45MPaG Deck 94m 171kg 2.3m2 35.7m/x 0.64MPaG Vent pipe 91m 14kg 0.02m2 31.5m/s 0.57MPaG

Also the Table 4-3 has listed down some of the V-3138 surrounding structures which were not affected

by the fire. In general, the structures impact by a load such as a blast wave, which has a very short action

time, should be examined by dynamic response analysis. However, it is supposed that the load received

duration was sufficiently long in relation to the natural frequency inherent to these structures. Therefore,

the blast pressure was estimated based on a static model.

(Table 4-3) Estimated blast pressure based on the damage sustained by V-3138 surrounding structures

Structures Distance from Tank Center

Damage Conditions

Specification Blast Pressure

[MPaG] Dike (south side) 3.3m Collapse 630mmH×150mmt More than 0.067

Dike (west side) 3.3m No damage 480mmH×150mmt Less than 0.115

H-steel piperacks 5.2m Severe bent 125mmW×4820mmH×7mmt 0.032~0.066

Lamp post 13.9m Bent 125A×940mmH×+

80A×2950mmt 0.014~0.018

Window glasses 74m Broken 430mmW×860mmH×3mmt 0.001~0.002

The explosive power was evaluated based on the blast pressure acted on the surrounding structures

and distance from the tank center. As a result, the mass of TNT equivalent was estimated at about 1~3kg.

Reference [4.2.1-1]

C.J.H. van den Bosch, R.A.P.M. Weterings, Methods for the calculation of physical effects - due to releases

of hazardous materials (liquids and gases) - 'Yellow Book'

25

(3) Types of explosion

Considering the following three types of explosion as the energy that brought the above mentioned

explosive power, the Boiling Liquid Expanding Vapor Explosion was thought to be the closest probable

type.

i. Vapor Cloud Explosion (Acrylic Acid Vapor)

This is a phenomenon in which a vapor cloud of flammable gas ignites and explodes. It is

accompanied by an extremely large explosive power.

Oxygen is necessary to form this type of explosion and it is unlikely the vapor cloud has

ignited inside the tank, since V-3138 was sealed with M-Gas. But given the flying directions of

the fragments and scattered contents, it is presumed that a huge pressure was generated in V-3138.

These were not coincided with the facts found after the explosion.

According to the estimated mass of TNT equivalent as the explosive power, the amount of

acrylic acid vapor that contributed to the explosion was calculated to be about 2~8kg. On the

other hand, the estimated amount of acrylic acid vapor released from V-3138’s vent to the

surrounding up till explosion was about 5,500kg (refer to Appendix 2). In view of explosive

power, it was hard to conclude that this type of explosion has occurred.

ii. Explosion (rupture) caused by V-3138 increased pressure due to runaway reaction (polymerization)

This was an explosion (rupture) associated with a rise in V-3138 internal pressure and it

coincided with the conditions of scattered fragments and contents.

The V-3138 internal pressure at the point of explosion was estimated to be about 0.27MPaG

from the data obtained in the adiabatic reaction test. Also the pressure that initiated the cracks

was estimated by the tank structural analysis as being 0.24~0.29MPaG. The estimated V-3138

internal pressures according to these two methods were nearly matched. However the tank

pressure estimated from the distance of scattered tank fragments at the time of V-3138 exploded

was about 0.45~0.64MPaG. This was nearly twice the estimated pressure that initiated the cracks

and these differences cannot be explained convincingly.

According to the process data, the recorded reading of V-3138 liquid level gauge has

exceeded the out-of-range limit at about 13:40 and dropped drastically before the explosion. This

dropped has triggered an abnormal low liquid level alarm at 58.6m3. These phenomena showed

that tank internal pressure had ever dropped just before the explosion occurred and this cannot be

explained convincingly if the explosion is caused by a runaway reaction.

iii. Boiling Liquid Expanding Vapor Explosion (BLEVE)

This is a phenomenon which phase equilibrium has destroyed by the rupture of a tank

containing a pressurized liquid above its boiling point. The dropped in liquid level gauge reading

(tank internal pressure) have coincided well to this phenomenon.

It has been reported in the literature that BLEVE will have an overpressure of about twice

the initial pressure. By comparing the estimated initial pressure which initiated the cracks

(0.24~0.29MPaG) and the estimated pressure at the time of explosion (0.45~0.64MPaG), the

pressure at the point of explosion was about twice the initial pressure.

26

The phenomenon of BLEVE occurred inside the tank has coincided with the scattered

conditions of V-3138 fragments and contents.

4.2.2. Fire Effect and Damages

The fire locations followed by the explosion have been confirmed based on the firefighting activities

and they are listed as below. Figure 4-19 shows the areas which caught fire.

i. Pool fire within the dike where V-3138 was installed

ii. Fire on the building (electric room) located south of V-3138

iii. Fire spread to surrounding cables and fire engines

iv. Fire spread due to the scattered contents that flew to the east direction

Table 4-4 shows the liquid volume before and after accident of each tank located within the same dike

as V-3138. Besides the V-3138, there are three tanks, i.e. V-3116, V-3124, and V-3148C, whose contents

varied before and after the accident. It is estimated that these tanks were damaged by the explosion and the

contents leaked out within the dike. These contents continued to fuel the pool fire (acrylic acid: approx.

67m3, and toluene: approx. 28m3).

(Table 4-4) The tanks installed within the dike and its liquid volume before and after accident.

Tank No. Main Contents Nominal

Capacity [m3]Liquid Volume [m3] Equipment Condition

After Accident Before After V-3138 Acrylic acid 70 60 - Ruptured V-3116 Acrylic acid 100 67 2 Damaged (leaked) V-3124 Toluene 50 30 1.5 Damaged (leaked) V-3145A Toluene 15 9.4 9.4 V-3145B Toluene 5.2 0.4 0.4 Damaged V-3146A Acrylic acid 15 8 8 V-3146B Acrylic acid 15 13 13 V-3147A Acrylic acid 10 7 7 V-3147B Acrylic acid 5.2 3 3 Damaged V-3147C Acrylic acid 5.2 4.2 4.2 Damaged V-3148C Acrylic acid 5.2 1.5 0.08 Damaged (leaked)

In addition, the highly possible causes of casualties are assumed due to the following factors:

i. Splashed by V-3138 high temperature contents which were released during the explosion.

ii. Burned by radiant heat generated from the pool fire at dike where V-3138 was installed.

iii. Burned by radiant heat generated from the possible fireball occurred after the BLEVE.

It is assumed that fire was possibly ignited by the following objects:

i. Sparks generated from the impact of metals at time of explosion

ii. Sparks generated from the broken electric cables

27

Location of Damaged Equipment

Location of Scattered Contents

Location of V-3138 Scattered Fragments

Cordon

①

②

③ ④

①

Main Fire Areas

Areas where Fire Spreaded

①

Foam spray from Chemical Fire Engine (Public). After extinguished fire at southwest of 3AB, fire engine moved towards south. After that, fire hoses extended to the east of accident tank yard for fire extinguishing.

②

③

④

Water Spray from Public Fire Engine

Water Spray from Fire Hydrant (NS)

Water Spray from Fire Hydrant (NS)

1

8

Fire Areas Around the Tank※ This information is prepared based on the status confirmed after the accident and Nippon Shokubai (NS) employees interviews.

※①～④ Measures taken after V-3138 Explosion and Fire

①

NS Fire Engine

Public Fire Engine

Location of Casualties

(Figure 4-19) Fire areas around the tank

28

5. Determining the Causes of the Accident

5.1. Contributing Factors of the Explosion and Fire

5.1.1. Direct Causes of the Accident

Based on the investigations and analysis, the committee has identified the direct causes of the accident

as following:

i. Even though high temperature T-5108 bottom liquid was building up in V-3138, the fact that the

circulation of Recycle to Top was not commissioned has caused acrylic acid to remain stagnant for

a significant long period of time at high temperature in the upper portion of the tank.

ii. DAA formation accelerated in the tank liquid with high temperature zones and the heat of

dimerization has caused the liquid temperature to increase. This has also caused the acrylic acid

start to polymerize and increased the liquid temperature further.

iii. Due to lack of thermometers and inadequate temperature monitoring, it was not possible to detect

the abnormal condition until polymerization had proceeded.

These have resulted in V-3138 exploded, followed by fire and caused enormous casualties and property

damage.

5.1.2. Contributing Factors of the Accident

(1) Fault-Tree diagram (FT diagram) of accident contributing factors

In light of the direct causes of the accident, an FT diagram was used to systematically analyze events

in order to clarify contributing factors that could have caused the accident and their connections. The

results are shown in Figure 5-1. The following eight items are determined as contributing factors that

caused the accident:

a) Excessive heating in tank feed liquid

The T-5108 bottom liquid temperature was about 65°C but it has increased to about 100°C by

the steam jacketed transfer piping before charging into V-3138.

b) Recycle to Top valve was closed

When tank liquid was to be maintained for several days above 25m3, Recycle to Top was

required to be commissioned. However this valve was not operated and remained closed.

c) Inadequate tank control temperature setting

d) Inadequate tank temperature detection

There was no setting for tank control temperature and there was no thermometer installed.

e) Unable to control the temperature within normal range

f) Unable to control below temperature upper limit

g) Unable to avoid abnormal situations

h) Unable to avoid crisis situations

29

No criteria were developed to judge the symptom and abnormality of tank temperature. Also the

response procedures to handle abnormal situations were not established.

Tank Explosion.Fire Started

Insufficient Inhibitors

Contamination

Unable to Avoid Abnormal Situation

AA Polymerization Proceeded. Temperature and

Pressure Increased

Heat Exchanger Capacity Reduced

Recycle to Top not commissioned

Higher Than Normal Feed Temperature

Unable to Control Within Normal Temperature

Inhibitor Solutions Mis-preparation

Inhibitors Inactivated due to Insufficient O2

Atmosphere

Polymers at Tank Vapor Part

Excessive Heat in Tank Feed Liquid

Upstream Temperature

Unable to Control

Leaked. Contaminated.

Unable to Avoid Crisis Situation

Poor Cooling Water Flow

Cooling Water Supply Temp.

Increased

Recycle to Top Closed

Circulation Pump Not Started

Valves Mis-operation

Inadequate Tank Control Temperature

Setting

Tank Temperature Exceed Normal Control Range ※

Inadequate Tank Temperature

Detection

AND

AND

OR

AND

AND

AND

OR

OR

OR OR

OR

Factor a)

Factor b)

Factor c) Factor d)

Factor e)

Factor f)

Factor g)

Factor h)

DAA formation Accelerated. Tank Temperature Increased

Stagnant in High Temperature（Exceed Temp. Upper Limit

※）

Unable to Control Temp. Below Upper Limit

Insufficient Cooling of Tank Liquid

OR

Inadequate Monitoring

Insufficient Tank Temperature

Control

OR

AND

※ With or without temperature control setting, the temperature has normal range. Determined as one of the contributing factors.

Inadequate Control of Inhibitor Injected

***** Excluded from contributing factors due to results of investigation and evaluation.

(Figure 5-1) Fault Tree Diagram (FT Diagram) of Contributing Factors

(2) Contributing factors of the accident and its management elements

In view of matters related to contributing factors of the accident, developing a safer system and

background of the surrounding circumstances etc., the management elements of each contributing factor

was investigated respectively. Table 5-1 shows the management elements related to each contributing

factors of the accident.

30

(Table 5-1) Management Elements of Respective Contributing Factors of the Accident

No. Contributing factors Management Elements

(1) a) Excessive heating in tank feed liquid Design review, validation and design

philosophies handover

b) Recycle to Top valve was closed

(2) -1) V-3138 Control Procedure Operation manuals development and

dissemination

(3) -2) T-6701 load up test Risk assessment and safe work management

for non-routine work

(4) c) Inadequate tank control temperature

setting

d) Inadequate tank temperature detection

Operating conditions setting and

management

(5) e) Unable to control the temperature

within normal range

f) Unable to cool down the temperature

below the upper limit

g) Unable to avoid abnormal situations

Criteria for abnormal situations and its

respective response procedures

(6) h) Unable to avoid crisis situations Crisis management and disaster prevention

activities

5.1.3. Background of Accident Contributing Factors

Investigate the management elements of contributing factors background

(1) Contributing factor a) Excessive heating in tank feed liquid

(Figure 5-2) History of glacial acrylic acid plants

31

i. Equipment history

In the past, hot water jacket was used for distillation columns and piping to prevent

polymerization and precipitation. As the polymerization prevention technology advanced, equipment

using hot water jacket has been reduced and hot water supply equipment also decreased.

When T-5108 was constructed, hot water jacket was not necessary for the T-5108 as a result of

technological improvement. On the other hand, the heating facility was still necessary for T-5108

bottom liquid transfer piping to prevent freezing and precipitation. However, since there was no hot

water equipment, steam jacket was chosen. Steam tracing was used for the T-6108 bottom liquid

transfer piping which was constructed at a later stage (See Figure 5-2).

ii. Design and construction stages

T-5108 was constructed as part of the 5AA construction project and commissioned in August

1994. The project was carried out by a mixed team of members from three departments: Production,

Technology and Engineering.

It was widely known that T-108 columns’ bottom liquid presented concerns of sludge

precipitation at low temperature and polymerization at high temperature. But, it had little recognition

on trouble may caused by insufficient cooling in V-3138 liquid, since the T-108 columns’ bottom

liquid contained a lot of inhibitors. A steam pressure regulating valve and a temperature control steam

trap were installed at the steam jacketed piping considered the T-5108 bottom liquid may polymerize

due to the excessive heat from steam jacket. However, the cooling capacity of the V-3138 coil was not

verified and the risk of causing inadequate cooling in the upper portion of the tank liquid was not

considered.

During that time, there was no design review system in place to verify, assess and review the

design of new equipment and hence the new installation has no appropriate evaluation from various

perspectives.

iii. Test run stage

The pressure setting of steam regulating valve at the steam jacketed piping that transfers T-5108

bottom liquid was set at 0.02MPaG based on setting stipulated in the operation manuals. The

temperature setting for the temperature control steam trap was set at 90°C based on the test run results.

But there was no monitoring device provided to monitor liquid temperature at the steam jacketed

piping outlet. It was assumed that at this point, the liquid has already been excessively heated to a

temperature higher than the T-5108 bottom temperature.

iv. After commercial operation

The T-5108 bottom liquid temperature at the steam jacketed piping outlet was not widely

recognized in the Production department. It was assumed to be about the same as the T-5108 bottom

temperature. This lack of accurate information (design concept, technical evaluation, adopted rational,

etc.) in the Production department was thought to be due to improper handover from the project team

to the Production department.

Moreover, the temperature control steam trap has frequently experienced drain accumulation in

the jacket due to rust blockage at the drain outlet part. Ultimately it was decided to remove the

32

temperature control steam trap and allowed the steam to flow through the jacket continuously. As the

result, the T-5108 bottom liquid was heated to about 100°C. Also in December 2009, the steam

regulating valve was changed and the pressure setting was adjusted to 0.005MPaG. Therefore, it was

assumed that the temperature control steam trap was removed sometimes after this point. There was

no record on when the temperature control steam trap was removed or any investigation or procedure

relating to this removal.

(2) Contributing factor b) -1) Recycle to Top valve was closed (V-3138 Control Procedure)

M-Gas

V-3138

P-3138C

Manhole

Recycle to Level Gauge

Recycle to Top

Pressure Gauge

Manhole

Liquid Level Gauge

(ALI3138）

T-701 Columns

T-108 ColumnsBottom Liquid

（Since February, 2009）

（Since initial construction）

(Figure 5-3) V-3138 status before the accident

i. Chronology: Year 1985 to 2009

V-3138 was an intermediate tank between the T-108 columns and the T-701 columns. It had

a nominal capacity of 70m3 and was installed in 1985. At the time, it took two days to stop

T-2701 operation for maintenance and V-3138 was designed with a capacity that could store the

bottom liquid from the T-108 columns (T-2108/3108) during this two days period. Also, only

Recycle to Top circulation is provided in the tank. During normal operation of T-108 and T-701

columns, it was not necessary to store large volume of liquid in V-3138. Therefore, the liquid

volume was usually kept low not to exceed the top of cooling coil.

T-108 columns bottom liquid was supplied either via V-3138 or directly to the T-701

columns. But since Year 2000, it was supplied via V-3138 in order to stabilize the supply flow

rate to the T-701 columns. If V-3138 liquid temperature decreases, it would cause sludge

precipitation and accumulated inside the tank. This has caused problems such as wrong readings

in liquid level gauge and plugged the P-3138C strainer. It took a lot of works and efforts to

handle these problems. In February 2009, an additional circulation line, Recycle to Level Gauge,

was installed as a countermeasure to fix the wrong readings in the liquid level gauge.

33

ii. Chronology: Year 2009 to 2010

Even after the Recycle to Level Gauge was installed, there were still a lot of works to be

done such as cleaning and washing the pump strainer and its surrounding piping. After opening

up and cleaning the tank in August 2009, the procedures on how to use the tank was changed

temporarily starting from September 2009 in order to confirm the effectiveness of the new

procedure. The temporary procedure relating to tank usage was communicated and documented

in the Operation Knowledge Base system (a database which provides a centralized store of

operational related information). The content is presented as follows:

-1) Only withdrawing column bottom liquid to V-3138 when T-108 column is stopped.

-2) After feeding liquid into V-3138, supply the tank liquid to the T-701 column for

processing and reduce the tank liquid volume until 5m3, the minimum required liquid

volume which need to keep inside the tank.

-3) After processing the tank liquid, blow clean the surrounding piping and commission

only the Recycle to Level Gauge for tank circulation.

-4) Close the cooling water valve to the coil.

About five months later, at the end of January 2010, the above procedures were revised as

the “V-3138 Control Procedure”. Notification of this change was raised in the Operation

Knowledge Base and the content was listed in the following:

-1) When the T-108 columns are in operation, supply the T-108 columns bottom liquid

directly to the T-701 columns.

Only withdrawing column bottom liquid to V-3138 when T-108 column is stopped.

-2) After feeding liquid into V-3138, supply the tank liquid to the T-701 columns for

processing. Reduce the tank liquid volume until 5m3, the minimum required liquid

volume that covers the lower portion of the cooling coil.

-3) After processing the tank liquid, blow clean the surrounding piping and commission

only the Recycle to Level Gauge for tank circulation.

-4) In the situation where the liquid in the tank is unable to process and need to keep

inside the tank for several days, commission the Recycle to Top if the liquid volume has

more than 25m3 (liquid volume up to the top of the cooling coil).

-5) The cooling water valve to the coil should be opened normally.

This “V-3138 Control Procedure” was a permanent matter, however it has not reflected in

any documentation (ex. operation manuals etc.) or management system other than notification in

Operation Knowledge Base. However, in order to assist the site operators, signage was posted

above the Recycle to Top valve on site stating: “Normally Closed, Open Recycle if above 25m3.”

iii. Chronology: Year 2010 to 2012

After changing to V-3138 Control Procedure, the frequency of cleaning the pump strainer

has dropped sharply. This has significantly reduced the workload of operators. Moreover, the

tank internal has not been washed since the last cleaning in August 2009. Figure 5-4 shows the

frequency of pump strainer cleaning.

34

(Figure 5-4) Frequency of P-3138C strainer cleaning

Originally, the cleaning and washing works around V-3138 is due to the precipitate

accumulation inside the tank. By changing the procedures of operating the tank, the effectiveness

of this improvement has verified by the reduced frequency in strainer cleaning, etc. However, any

change in the amount of deposition inside the tank and the need for Recycle to Level Gauge was

not re-examined.

Thereafter, the chances of storing the liquid above 25m3 in V-3138 or opportunities to

operate the Recycle to Top valve were hardly experienced by operators. Hence the operators have

become unaware of the necessity to commission the Recycle to Top.

Furthermore, there were multiple rectifying columns connected to V-3138. Switching the

feed of T-108 columns bottom liquid to V-3138 or to the T-701 columns was carried out by

operating the header valve in different locations. These locations were about 15m away from the

P-3138C discharge header as indicated in Figure 5-3. For this reason, it might be difficult for

operators to notice the sign pertaining to Recycle to Top valve.

(3) Contributing factor b) -2) Recycle to Top valve was closed (T-6701 load up test)

i. Background relating to safe work management and operating instructions

In the Production department, “Permit-To-Work System” (PTW) is applied for any works

which required special attentions and safety measures. Also PTW is applied to “Standard for

Non-Regular Works” which are not specified in the Standard Operating Procedure. Approval is

necessary to carry out such works and safe work procedures are prepared in advance for each

different type of works. These safety measures and its contents are required to make known to all

members before any work is allowed to start.

However, in regards to operation and safe work management, there are no rules specified

35

things such as approval route, issuance of operating instructions. Practices on operating

instructions differ by each department. But basically if necessary, they will be issued with

Production Manager approval regardless whether there is any manual in place. Generally the

operating instructions are often issued to supplement the manuals.

ii. Risk management associated with changes

The rule of “Management of Change” (MOC) is applied for risk management associated

with permanent or temporary changes in equipment and processes, etc. This rule was established

in August 2004 and revamped completely in September 2009, after which this rule has been

implemented through database. Each department evaluated their practices to MOC and conducted

the necessary countermeasure.

According to this rule, any changes are evaluated using a MOC check sheet. The depth of

review and approval route are then determined based on the check sheet’s result. Changes in

which this rule is not applied are also described with examples. Examples which this rule is not

applicable included “changes involving same specification in facility, piping, equipment” or

“changes involving operating conditions and procedures but within the safe operating scope”.

However, the interpretations of whether this rule is applicable may differ between individuals. If

they think this rule is not applicable, the MOC check sheet will not be used.

iii. Load up test in Year 2009

The V-3138 liquid level built up in order to conduct a load up test in T-6701. This

intentionally liquid built up was not a normal but non-routine work. Moreover, the load up test

for T-6701 was an operation conducted under temporary conditions and was not covered in the

operation manuals.

The purpose of the T-6701 load up test was to confirm the operating conditions under which

the quality of distillate could be ensured. The testing conditions for T-6701 were thought to be

adjusted within the range of equipment capacity, given that the rectifying column load was within

the operating range of the past track records and the liquid volume to store in V-3138 was within

the tank’s nominal capacity. Accordingly, a testing plan indicating matters such as T-6701 test

method, conditions and schedule was prepared. However, the risks pertaining to storing liquid in

V-3138 were not investigated.

The testing plan was issued with the approval of Production Manager but no operating

instruction was issued. However at that time, the methods of provisional use of V-3138 were

verified and the Recycle to Top was commissioned after liquid built up in the tank.

iv. Load up test in Year 2012

The understanding regarding the T-6701 load up test was the same as test carried out in Year

2009. Hence the risks associated to build up V-3138 level were not investigated. Moreover, this

V-3138 storing operation was expected to carry out according to the “V-3138 Control Procedure”,

however its contents had not been reflected in manuals and no operating instructions were issued.

Therefore operators were not aware of the “V-3138 Control Procedure”.

36

(4) Contributing factor c) Inadequate tank control temperature setting

Contributing factor d) Inadequate tank temperature detection

i. V-3138 temperature control

An acrylic acid plant is made up of various unit operations such as absorption, distillation in

addition to the oxidation reaction. The temperatures in each of this main process unit need to be

precisely controlled.

Regarding the temperature control in V-3138, the tank temperature has to be maintained

within a range due to concerns of acrylic acid freezing, sludge precipitation at low temperature

and polymerization at high temperature. Normally only small volume of liquid with high contents

of inhibitors was stored in V-3138 and it could be cooled by the cooling coil inside the tank.

Hence the precise control of temperature was not necessary. Moreover, compared to the main

process equipment, there was relatively little awareness of the need for quantitative temperature

control in V-3138. This status has been the same since the tank constructed till now.

Furthermore, inhibitors and the atmosphere condition are not able to prevent the formation

of DAA, whereas the temperature has a big impact. But these risk information associated to DAA

formation were not widely shared. Hence it is assumed that awareness of temperature control to

prevent DAA formation was not established.

ii. Necessity of instrument installation

The need to install a thermometer and how to use it (on-site or remote monitoring) is

established based on matters such as the necessity of temperature control, the temperature

monitoring cycle and legal requirements. But, the plant has no standard on means of temperature

monitoring that reflects these points comprehensively. Meanwhile assessment for the need of

temperature control and monitoring was carried out by each plant individually. When the need for

temperature control is assessed as low and intermittent detection is sufficient, thermometers will

not be installed. This has created variation in the assessment and awareness of temperature

control and monitoring.

iii. V-3138 thermometer: design and construction stages

As described above, there was low awareness of the need for quantitative temperature

control in V-3138, and hence thermometer installation was not planned at the design stage. Later,

additional nozzle and other changes were made as the project moved into detailed design and

construction stages, the addition of thermometer nozzle was one of these changes. However, there

were no records available that can accurately identify the intention on how to use the nozzle or

the existence of a thermometer when the equipment commissioned.

iv. About the V-3138 thermometer: relation to lateral deployment following a trouble

In April 1994, polymerization trouble occurred in another acrylic acid intermediate tank.

Countermeasures proposed at that time included implementing continuous temperature

monitoring and installing external heat exchangers at similar tanks other than the tank where the

trouble occurred. Even though V-3138 was found of not having thermometer installed during that

37

time, it was not included in the lateral deployment scope. It is assumed the reason for this was

because the equipment subject to the lateral deployment was tanks that received column bottom

liquid at 80~100°C whereas the bottom temperature in the T-108 columns were only about

65~70°C. Also, T-5108 has not yet been constructed at that time.

During that time, the causes investigation and countermeasures implementation for this

trouble were concluded only within the involved department. There was no system in place to

involve participation of other departments likes Technology in investigation for preventing

similar trouble in other areas.

(5) Contributing factor e) Unable to control the temperature within normal range

Contributing factor f) Unable to cool down the temperature below the upper limit

Contributing factor g) Unable to avoid abnormal situations

i. Overall safety measures for acrylic acid

In acrylic acid plant, if any abnormality is detected and prevent the normal operation to

continue, the plant will be shutdown through interlock system. Conventionally, safety measures

for acrylic acid focused on measures to prevent trouble. However, emergency criteria and

response procedures were not established for the abnormal situations which cannot be controlled

even if the equipment is shut down.

ii. V-3138

Since there was insufficient temperature monitoring device in V-3138, it was not possible to

detect the temperature has exceeded until the equipment has been shut down or some sort of

actions have been taken. Also there were no criteria and means for handling abnormal situations

established and therefore, there was no equipment to handle this situation.

(6) Contributing factor h) Unable to avoid crisis situations

i. Past trouble

The Himeji Plant has not experienced a tank ruptured or explosion due to acrylic acid

polymerization. However, acrylic acid polymers have been discovered during tank internal

inspection and basic measures to prevent polymerization have been implemented for each tank,

such as the removal of unnecessary structures.

Measures taken in the past when acrylic acid polymerized in the tanks included two

instances of spraying water onto the tank and one instance of charging water and inhibitors from

the manhole. Other measures which have been taken included charging cooling water into the

tank jacket.

ii. Activities relating to responses to abnormal phenomena

With the aim of improving the disaster prevention capabilities of the safety team in the

Himeji Manufacturing Innovation (HMI) division, various activities have been conducted

together with the Responsible Care (RC) activities since Year 2007. Furthermore, other than the

complex-wide disaster prevention trainings, each of the various workplaces will also implement

38

its own planned disaster prevention exercises. One of these trainings is the fire fighting exercises

which include a yearly fire fighting competition. Others also include building up a good supply of

the disaster prevention equipment.

Immediate response procedures for abnormal phenomena such as hazardous materials or

high pressure gas leakage and fires, etc. have been developed in the Plant’s Emergency Response

Plan but none of these covered for runaway reaction. (In the training materials relating to

abnormal phenomena and alerting procedure (in the process for approval), there was illustration

to cool the tank by spraying water as a way of responding to runaway reaction).

iii. Status of rules development

A self disaster prevention manual has been developed as part of the Himeji Plant Rules and

Regulations. However, the role and responsibility on providing information to public fire

department was not clearly defined in self disaster prevention team. Also, there was no manual to

standardize each plant criteria for abnormality and its respective response procedure.

iv. Responses during this accident

Tank cooling was judged to be necessary based on the V-3138 conditions and hence the tank

was sprayed with water. At first, operators sprayed water from six fire hydrants (each at 400

liters/min.). Later, the self disaster prevention team sprayed additional water (3,100 liters/min)

from the fire engine. However V-3138 was insulated and it was assumed that water spraying did

not have much cooling effect to the tank liquid.

Since acrylic acid vapor was being released from V-3138, the intention of spraying water

onto the tank was to cool and to absorb the acrylic acid vapor. This action may have prevented

vapor cloud explosion. However under such circumstances additional measures such as charging

water and inhibitors inside the tank could not be taken. There was no means to control the

abnormal situations progress other than spraying water.

Based on the above, the backgrounds of each contributing factor are summarized in the Figure 5-5.

Also, each of the contributing factors, management elements of its background and consideration to prevent

recurrence are summarized in Table 5-2.

39

Risk

　Cr isis S itu at ion Occu rred

Risk Visualized 　Ac tual Risk

↑ ↑Risk Assessment

（Not Conducted）

　Assumed Risk

Less Many Work Frequency

Background of Low High Standardize

Accident Contributing Factors ↑ ↑ ↑ ↑ ↑ ↑ ↑

(1) Insufficient design review and design philosophy handover procedures

(2) Insufficient operation manuals development and dissemination

(3) Weak systems in safe work management and risk evaluating of changes, non-routine works

(4) Inadequate management of operating conditions settings

(5) No establishment of criteria for abnormal and crisis situations and its respective response procedures

(6) Poor coordination with public fire department during disaster prevention acitivities

↑ ↑ ↑ ↑ ↑ ↑ ↑ ↑

Safety culture and safety climate issues that derived from the above backgrounds

(Safety awareness and hazards sensibility, etc.)

(Lack of understanding of technical knowledge, expertise and hazardous information)

Invisble

Prepared for T-6701 load up

test, hence risk increasedVisible

Recognized risk while planning

T-6701 load up test

(No difference from normal

operation)

V-3138 Normal

ManagementV-3138 Storing

V-3138 Storing

V-3138 Disaster Prevention

Actitivities

(Figure 5-5) Summary of background of accident contributing factors

40

(Table 5-2) Consideration points for recurrence prevention

No. Contributing factors Management elements

Consideration Points for recurrence prevention

(1) a) Excessive heating in tank feed liquid Design review, validation and design philosophy handover

1) Conduct risk assessment and design review during the design stage.

2) Verify the operating conditions at the test run stage

3) Educate and disseminate any changes and its technical evaluation contents to operators.

4) Conduct risk assessment for any changes in equipment and verify the changes after the modification.

5) Manage the records and revision history thoroughly.

(2) b) Recycle to Top valve was closed

1) V-3138 Control Procedure

Operation manuals development and dissemination

1) Include the control procedure of intermediate tanks in the operation manuals

2) Conduct risk assessment for any changes in operating procedures and verify them after the changes

3) Disseminate the operating procedures to operators and confirm the procedures are executed.

4) Enhance the works’ reliability (easy-to-understand signage for field operators, etc.)

(3) b) Recycle to Top valve was closed

2) T-6701 load up test

Risk assessment and safe work management for

non-routine work

1) Confirm risk assessment is executed for non-routine works and changes.

2) Review the MOC system (risk evaluation scope, check sheets usage, etc.)

3) Establish system to ensure risk assessment is conducted before operation, issuance of operating instructions and its approval

routes, etc.

4) Enforce precaution checks before work implementations based on operating instructions.

5) Strengthen risk management system through education and personnel development

(4) c) Inadequate tank control temperature setting

d) Inadequate tank temperature detection

Operating conditions setting and management

1) Standardize temperature monitoring methods.

2) Standardize temperature control methods and improve the accuracy of temperature controlling.

3) Share the hazardous information of the handled materials.

4) Manage the equipment revision history thoroughly.

5) Strengthen participation of multiple departments including technology department in any trouble investigation.

(5) e) Unable to control the temperature within normal range

f) Unable to cool down the temperature below the upper limit

g) Unable to avoid abnormal situations

Criteria for abnormal situations and its respective response

procedures

1) Establish the criteria for abnormal situations and its respective emergency response procedures.

(6) h) Unable to avoid crisis situations Crisis management and disaster prevention activities

1) Establish the criteria for crisis situations and its respective emergency response procedures.

2) Improve cooperation with public fire department (information sharing development, review communication structure, etc.)

41

5.2. Matters Related to Expanded Damage Resulted from the Explosion and Fire

Based on Section 5.1 contributing factors of the accident, matters related to expanded damage are

discussed below:

(1) Property damage

i. The abnormal development that led to the tank explosion was due to the progressive

polymerization of acrylic acid. One of the reasons for damage to expand was due to the fact that

abnormal conditions could not be detected at early stage as a result of inadequate temperature

detection.

ii. As mentioned in Section 5.1.3, another factor that caused the property damaged to expand was no

other means were available to prevent the development from abnormal situations to crisis

situations besides water spraying.

(2) Casualties

i. Casualties in this accident were caused by tank explosion while one team was preparing and the

other team was spraying water onto the tank. They were exposed to fire radiant heat and splashed

by the high temperature contents from the tank. The casualties’ injuries were related to their

distance from the tank and personal protection equipment worn.

ii. For conducting disaster prevention activities from an appropriate distance with the right equipment,

factors such as criteria to determine abnormality, response plan, organization, etc. are important.

However these factors were not fully established for crisis situations and hence the casualties have

expanded.

42

6. Recommendations for Accident Recurrence Prevention Countermeasures

The following countermeasures are recommended to prevent the accident recurred, based on the

contributing factors and background that led to the accident as summarized in the preceding section. The

recurrence prevention countermeasures were derived using 6M-5E analysis based on the perspectives

described in Table 6-1. The details of the analysis are described in Table 6-2.

Table 6-1. Perspectives of 6M-5E analysis

6M: Perspectives on background types of accident contributing factors

5E: Perspectives of deriving the recurrence prevention countermeasures

Man Education

Machine Engineering

Method Enforcement

Material Environment

Measurement Example

Management

6.1. Recurrence Prevention Countermeasures for Equipment Involved

For restoring equipment involved, in order to ensure greater safety and stable production activities, the

following recurrence prevention countermeasures should be implemented.

(1) Revise the specification of the T-5108 bottom liquid transfer piping

In order to prevent overheating of tank feed liquid, the following items should be implemented:

i. Reconfirm the liquid properties and setting of the design conditions.

ii. Conduct risk assessment for the piping specification change and evaluate its propriety.

iii. Confirm the piping operating conditions and status through test run.

(2) Study the specifications of V-3138 and its associated equipment (new installation)

In order to control the V-3138 liquid temperature at the appropriate temperature and to ensure

evacuation criteria and to clarify its response procedures, the following items should be implemented:

i. Reconfirm the liquid properties and temperature control settings.

ii. Provide means to control the temperature (include thermometers installation, Recycle to Top line is

always commission).

iii. Establish the abnormal temperature criteria and its respective response procedures (study the

emergency inhibitors injection, etc.)

iv. Conduct risk assessment for the V-3138 and its associated equipment installation and evaluate its

propriety.

v. Confirm the operating conditions, status, etc. of V-3138 and its associated equipment through test

run.

(3) Development of manuals, etc.

To clarify safety of the restored equipment and necessary matters for stable operation, the following

43

items should be implemented:

i. Develop manuals and other documentations for T-5108 and V-3138

ii. Update the P&ID, etc. around T-5108 and V-3138.

iii. Prepare site signage around T-5108 and V-3138.

(4) Education and training

To operate the restored equipment safely and stably according to the manuals, the following items

should be implemented:

i. Provide training of T-5108 and V-3138 operation manuals, etc. to operators.

ii. Re-educate employees to fully understand the detailed behavior of DAA formation and the hazards

of acrylic acid.

6.2. Countermeasures to Prevent Recurrence of Similar Accidents and its Lateral Deployment

6.2.1. Countermeasures to Prevent Recurrence of Similar Accidents

To prevent similar accidents from occurring at other production facilities in the Himeji Plant, the

following items should be implemented:

(1) Operation and safe work management

To ensure risk assessment is conducted whenever there are changes in non-routine works, procedures,

methods, equipment, etc. and each work details are disseminated, the following items should be

implemented:

i. Establish basic rules for safe work management

Establish basic rules for safe work management and standardize the definitions and safe work

management system for each work. Also, provide training on the application of these rules, ensure

safety checks are enforced and consider all the potential hazards thoroughly before starting any works.

ii. Review rules of Management of Change

Prevent risk evaluation overlook by reviewing the target which covered under MOC and check

sheets exercise. Moreover, strengthen the risk management framework by providing training on MOC

application through case studies, by training the safety manager, etc.

iii. Review the practice of operating instructions

Specify the rules for issuing operating instructions and its approval routes with consideration of

risk evaluation, while revising safe work management and MOC systems. Also practices such as in

advance risk assessment, hazards evaluation before the works implementation and confirmation after

works completion should be enforced.

(2) Crisis Management

To strengthen the emergency response capabilities when the abnormal situations arise, the following

items should be implemented:

i. Develop the crisis management manual

Clarify the basic policy and the framework of crisis management, and set out the activities to be

carried out during normal circumstances for preparing any abnormal situations. In addition, when the

44

abnormal situations arise, establish measures to secure human safety, minimize the damages and

prevent secondary disaster. Clarify the communication routes between public fire department.

ii. Review the self disaster prevention manual

Review the individual roles and definition of self disaster prevention organization including

arrangement during evening and off-days, in views of the communication structure likes information

conveying to public fire department. Improvements of initial responses (includes reporting) to the

abnormal situations and disaster prevention equipment should be reviewed as well.

iii. Provide education and training on these manuals

Establish a self disaster management system which enables effective collaboration with public

fire department by implementation of trainings and drills.

(3) Other Countermeasures

i. Re-educate the hazards of the handled substances.

ii. Review the design standards for the tank associated equipment.

iii. Review the system of lateral deployment following a trouble.

Trouble investigation should be executed by multiple relevant parties including Technology

department for ensuring that no countermeasure is overlooked during lateral deployment process.

iv. Collect and make full use of the external incident case studies and technical information

Develop engineer and improve technical capabilities through the above countermeasures and

various informations collected externally.

6.2.2. Disaster Prevention Measures for Equipment Handling Acrylic Acid

To prevent the recurrence of similar disaster in the equipment handling acrylic acid, the following

items should be implemented:

i. Standardize tank temperature monitoring methods (thermometer installation).

ii. Review the tank control temperature range and temperature control methods

Review the control temperatures for tanks that storing acrylic acid. Establish methods to control

the tank temperature within the pre-defined temperature range and supplementary measures to control

the tank temperature.

iii. Establish the criteria for judging abnormal symptom and the basic concepts of the response

procedures

Establish the basic concept of temperature criteria for foreseeing abnormal situations and the

response procedures when the temperature criteria exceeded.

iv. Establish the temperature criteria of each equipment based on the criteria for judging abnormal

symptom

v. Establish the response procedure for each equipment considering the foreseeable abnormal events

and crisis situations.

The response procedures to the abnormal events include feed isolation, delaying the abnormality

progress, release, withdrawal and isolation. Depends on the characteristics of the equipment, the

45

suitable response procedures will be applied. In order to minimize the damage caused by abnormal

events, the existing response procedures should be further strengthened.

vi. Conduct risk assessment and evaluate its propriety for the above items i to v.

vii. Modify equipment handling acrylic acid based on the above items i to vi.

viii. Update the manuals and conduct the related trainings to all facilities handling acrylic acid.

6.2.3. Lateral Deployment of the Accident Prevention Countermeasures

Review the countermeasures described in Sections 6.2.1 & 6.2.2 and compare them against the current

practices in other plants.

Furthermore, share the knowledge and information gained through the accident investigation to other

companies and the industry. With these, contribute to the acrylic acid industry or even the entire chemicals

industry through the safer production activities.

6.3. Fostering a Safety Culture of Safe Manufacturing Plant and Corporation

Nippon Shokubai has until now implemented diverse safety activities under a corporate commitment

of “Safety takes precedence over production”. Under these circumstances, the company repeatedly

expanded the production capacity of the Himeji Plant as its backbone plant without any major accident.

Despite these, the explosion and fire which occurred in acrylic acid production facilities (the main facilities

in the Himeji Plant) has resulted in significant casualties and property damage.

Our investigation has revealed various diverse factors have complicatedly combined and caused the

accident in the Himeji Plant. The continuous stable production over the past years has softened the safety

awareness and became less sensitive towards dangers. As a result, these factors have led into the accident in

the Himeji Plant.

These factors should not be downplayed as being localized and restricted to specific organizations and

facilities. Rather, they should be recognized as the representation of Nippon Shokubai current safety

capabilities. The company therefore needs to fully accept this accident, each employee should contemplate

and realize the corporate commitment from now on.

To accomplish these, there is a need to renew employees understanding that safety is not something

others provide but rather they need to recognize and achieve safety by themselves. The company must

ensure that this attitude is reflected in the future behavior of its organization and individuals. Achieving

safety begins with “abiding” the safety rules and ”noticing” circumstances that could detract from safety,

these will result in “changing” to a safer corporation. In any event, it is unlikely to achieve safety without

the knowledge, realization, and expertise of organization and individuals. Therefore, the company needs to

treat these safety issues as the company-wide challenge including personnel development, alongside

implementation of the recurrence prevention countermeasures.

The lesson learned from the accident in the Himeji Plant must not be forgotten over time. In order to

guarantee the effectiveness of recurrence prevention countermeasures, evaluation must be continuously

carried out not only by Nippon Shokubai and the Himeji Plant but also third party. The third party

evaluation will also provide newfound awareness for ensuring safety.

46

A corporate safety culture is fostered over time by engaging in these kinds of diverse activities. The

company is therefore advised to formulate medium- and long-term plans for implementing the recurrence

prevention countermeasures derived through the accident investigation, rather than simply implementing a

short-term response. By ensuing these plans are implemented, it will help the entire Nippon Shokubai

organization to foster a safety culture.

47

(Tab

le 6

-2)

6M-5

E a

naly

sis

for

back

grou

nd o

f ac

cide

nt c

ontr

ibut

ing

fact

ors

and

recu

rren

ce p

reve

ntio

n co

unte

rmea

sure

s (1

)

Con

trib

utin

g fa

ctor

a)

Exc

essi

ve h

eatin

g in

tank

fee

d liq

uid

B

ackg

roun

d of

acc

iden

t co

ntri

butin

g fa

ctor

s M

an

Mac

hine

M

etho

d M

ater

ial

Mea

sure

men

t M

anag

emen

t

Des

ign

and

co

nstr

uctio

n st

ages

C

once

rns

T-10

8 co

lum

ns’ b

otto

m li

quid

at

low

and

hig

h te

mpe

ratu

res

wer

e w

idel

y kn

own.

But

, it

had

littl

e re

cogn

itio

n on

tr

oubl

e m

ay c

ause

d by

in

suff

icie

nt c

oolin

g in

V

-313

8 liq

uid

sinc

e it

co

ntai

ned

a lo

t of

inhi

bito

rs.

Hot

wat

er ja

cket

was

not

ne

cess

ary

for

the

T-51

08.

Stea

m ja

cket

was

inst

alle

d at

bot

tom

liqu

id tr

ansf

er

pipi

ng. C

once

rnin

g th

e bo

ttom

liqu

id w

ill b

e he

ated

up

due

to e

xces

sive

hea

t fr

om s

team

jack

et, a

ste

am

pres

sure

reg

ulat

ing

valv

e an

d a

tem

pera

ture

con

trol

st

eam

trap

wer

e in

stal

led.

T

he c

ooli

ng c

apac

ity

of th

e V

-313

8 co

il w

as n

ot

veri

fied

and

the

risk

of

caus

ing

inad

equa

te c

oolin

g in

the

uppe

r po

rtio

n of

the

tank

liqu

id w

as n

ot

cons

ider

ed.

T-

108

colu

mns

’ bo

ttom

li

quid

pre

sent

ed

conc

erns

of

slud

ge

prec

ipit

atio

n at

low

te

mpe

ratu

re a

nd

poly

mer

izat

ion

at h

igh

tem

pera

ture

.

T-

5108

was

con

stru

cted

as

part

of

the

5AA

co

nstr

uctio

n pr

ojec

t. T

he

desi

gn, c

onst

ruct

ion

and

proj

ect t

eam

con

sist

ed o

f m

embe

rs f

rom

de

part

men

ts:

M

anuf

actu

ring

, Te

chno

logy

and

E

ngin

eeri

ng.

The

re w

as n

o de

sign

re

view

sys

tem

in p

lace

to

veri

fy, a

sses

s an

d re

view

th

e de

sign

of

new

eq

uipm

ent a

nd h

ence

new

in

stal

lati

on h

ad n

o ap

prop

riat

e ev

alua

tion

fr

om v

ario

us p

ersp

ecti

ves.

Test

run

sta

ge

Stea

m p

ress

ure

regu

lati

ng v

alve

: 0.

02M

PaG

Te

mpe

ratu

re c

ontr

ol

stea

m tr

ap: 9

0°C

It w

as p

resu

med

that

at

this

poi

nt, t

he li

quid

ha

s al

read

y be

en

exce

ssiv

ely

heat

ed to

a

tem

pera

ture

hig

her

than

th

e T-

5108

bot

tom

te

mpe

ratu

re.

The

tem

pera

ture

set

ting

fo

r th

e te

mpe

ratu

re

cont

rol s

team

trap

was

se

t at 9

0°C

bas

ed o

n th

e re

sult

s of

the

test

run

re

sult

s. B

ut th

ere

was

no

mon

itor

ing

devi

ce

prov

ided

to m

onit

or th

e li

quid

tem

pera

ture

at t

he

stea

m ja

cket

ed p

ipin

g ou

tlet

.

The

pre

ssur

e se

ttin

g of

st

eam

reg

ulat

ing

valv

e at

th

e st

eam

jack

eted

pip

ing

that

tran

sfer

s T-

5108

bo

ttom

liqu

id w

as s

et a

t 0.

02M

PaG

bas

ed o

n se

ttin

g st

ipul

ated

in th

e op

erat

ion

man

uals

.

48

(Tab

le 6

-2)

6M-5

E a

naly

sis

for

back

grou

nd o

f ac

cide

nt c

ontr

ibut

ing

fact

ors

and

recu

rren

ce p

reve

ntio

n co

unte

rmea

sure

s (1

)

Con

trib

utin

g fa

ctor

a)

Exc

essi

ve h

eatin

g in

tank

fee

d liq

uid

Bac

kgro

und

of a

ccid

ent

cont

ribu

ting

fact

ors

Man

M

achi

ne

Met

hod

Mat

eria

l M

easu

rem

ent

Man

agem

ent

Aft

er c

omm

erci

al

oper

atio

n T

he T

-510

8 bo

ttom

liqu

id

tem

pera

ture

at t

he s

team

ja

cket

ed p

ipin

g ou

tlet

was

as

sum

ed to

be

abou

t the

sa

me

as th

e T-

5108

bo

ttom

tem

pera

ture

.

The

tem

pera

ture

con

trol

st

eam

trap

has

fre

quen

tly

expe

rien

ced

drai

n ac

cum

ulat

ion

in th

e ja

cket

due

to r

ust

bloc

kage

at t

he d

rain

ou

tlet

par

t. U

ltim

atel

y, th

e te

mpe

ratu

re c

ontr

ol s

team

tr

ap w

as r

emov

ed.

The

pre

ssur

e se

ttin

g of

st

eam

pre

ssur

e re

gula

ting

va

lve

was

adj

uste

d to

0.

005M

PaG

.

As

a re

sult

of

rem

ovin

g th

e te

mpe

ratu

re c

ontr

ol

stea

m tr

ap, t

he T

-510

8 bo

ttom

liqu

id w

as h

eate

d to

abo

ut 1

00°C

.

Im

prop

er h

ando

ver

of

accu

rate

info

rmat

ion

(des

ign

conc

ept,

tech

nica

l ev

alua

tion

, ado

pted

ra

tion

al, e

tc.)

fro

m th

e pr

ojec

t tea

m to

the

Pro

duct

ion

depa

rtm

ent.

The

re w

as n

o re

cord

on

whe

n th

e te

mpe

ratu

re

cont

rol s

team

trap

was

re

mov

ed.

P

ersp

ecti

ves

of

Cou

nter

mea

sure

s

Edu

cati

on

Edu

cate

ris

k m

anag

emen

t in

rel

atio

n to

MO

C

(Reg

ulat

ions

and

cas

e st

udie

s)

Edu

cate

ope

rato

rs th

e pi

ping

spe

cifi

cati

on

chan

ge

(Spe

cifi

cati

on, d

esig

n co

nditi

ons,

etc

.)

Edu

cate

ope

rato

rs th

e pi

ping

spe

cifi

cati

on

chan

ge

(Ope

rati

ng in

stru

ctio

ns,

cont

rol s

tand

ard,

etc

.)

Edu

cate

ope

rato

rs th

e pi

ping

spe

cifi

cati

on

chan

ge

(Liq

uid

prop

erti

es, e

tc.)

Edu

cate

ope

rato

rs th

e pi

ping

spe

cifi

cati

on

chan

ge

(Tes

t run

res

ults

)

Eng

inee

ring

Rev

ise

the

spec

ific

atio

n of

T-5

108

bott

om li

quid

tr

ansf

er p

ipin

g.

(Des

ign

cond

itio

ns, e

tc.)

.

Ris

ks e

valu

atio

n in

re

lati

on to

spe

cifi

cati

on

chan

ge.

Upd

ate

the

P&

ID, e

tc.

arou

nd T

-510

8.

Rev

ise

the

spec

ific

atio

n of

T-5

108

bott

om li

quid

tr

ansf

er p

ipin

g.

(Rec

onfi

rm li

quid

pr

oper

ties

, etc

.).

Con

firm

pip

ing

oper

atin

g co

nditi

ons,

sta

tus,

etc

. th

roug

h te

st r

un.

Eva

luat

e pr

opri

ety

of th

e sp

ecif

icat

ion

of T

-510

8 bo

ttom

liqu

id tr

ansf

er

pipi

ng.

Enf

orce

men

t E

nfor

ce o

f ri

sk e

valu

atio

n fo

r ch

ange

s an

d eq

uipm

ent d

esig

n.

U

pdat

e op

erat

ion

man

uals

of

gla

cial

acr

ylic

aci

d pr

oduc

tion

faci

liti

es.

Dev

elop

man

uals

and

ot

her

docu

men

tati

ons.

M

anag

e th

e re

visi

on

hist

ory

thor

ough

ly.

Env

iron

men

t

Upd

ate

site

sig

nage

in

rela

tion

to p

ipin

g sp

ecif

icat

ion

chan

ges

Exa

mpl

e

Pre

pare

MO

C f

or a

ctua

l ca

se s

tudi

es. (

Equ

ipm

ent

chan

ges)

.

49

(Tab

le 6

-2)

6M-5

E a

naly

sis

for

back

grou

nd o

f ac

cide

nt c

ontr

ibut

ing

fact

ors

and

recu

rren

ce p

reve

ntio

n co

unte

rmea

sure

s (2

)

Con

trib

utin

g fa

ctor

b)

-1)

R

ecyc

le to

the

Top

valv

e cl

osed

(V

-313

8 C

ontr

ol P

roce

dure

)

Bac

kgro

und

of a

ccid

ent

cont

ribu

ting

fact

ors

Man

M

achi

ne

Met

hod

Mat

eria

l M

easu

rem

ent

Man

agem

ent

Yea

r 19

85-2

009

Slu

dge

prec

ipit

atio

n

accu

mul

ated

insi

de th

e ta

nk a

lway

s ca

used

suc

h as

wro

ng r

eadi

ngs

in

liqu

id le

vel g

auge

, pl

ugge

d th

e P

-313

8C

stra

iner

, etc

. It t

ook

a lo

t of

wor

ks to

han

dle

thes

e pr

oble

ms.

The

V-3

138

had

nom

inal

ca

paci

ty o

f 70

m3 a

nd th

e li

quid

vol

ume

up to

the

top

of

cool

ing

coil

was

25m

3 . At t

he

tim

e of

con

stru

ctio

n, o

nly

Rec

ycle

to T

op w

as in

stal

led.

V

-313

8 liq

uid

volu

me

was

us

uall

y ke

pt a

t low

leve

l whe

n T-

108

and

T-70

1 co

lum

ns

wer

e ru

nnin

g no

rmal

ly.

V-3

138

liqui

d vo

lum

e w

as u

sual

ly k

ept l

ow n

ot

to e

xcee

d th

e to

p of

co

olin

g co

il.

T-10

8 co

lum

ns b

otto

m

liqu

id w

as s

uppl

ied

eith

er v

ia V

-313

8 or

di

rect

ly to

T-7

01

colu

mns

. But

sin

ce Y

ear

2000

, the

T-1

08 c

olum

ns

botto

m li

quid

was

su

pplie

d to

T-7

01

colu

mns

via

V-3

138

only

.

Slu

dge

prec

ipit

atio

n w

ill

accu

mul

ate

insi

de th

e ta

nk if

V-3

138

liqui

d te

mpe

ratu

re is

low

.

Yea

r 20

09-2

010

Eve

n af

ter

the

Rec

ycle

to

Lev

el G

auge

was

in

stal

led,

ther

e w

ere

stil

l a

lot o

f cl

eani

ng a

nd

was

hing

wor

ks a

t pum

p st

rain

er, t

he s

urro

undi

ng

pipi

ng, e

tc.

In F

ebru

ary

2009

, the

Rec

ycle

to

Lev

el G

auge

was

inst

alle

d as

a c

ount

erm

easu

re to

fix

the

wro

ng r

eadi

ngs

in li

quid

leve

l ga

uge

due

to s

ludg

e pr

ecip

itat

ion

accu

mul

ated

in

side

the

tank

. A

t the

tim

e w

hen

V-3

138

Con

trol

Pro

cedu

re w

as

revi

sed,

a s

igna

ge w

as p

oste

d ab

ove

the

Rec

ycle

to T

op

valv

e on

sit

e st

atin

g:

“Nor

mal

ly C

lose

d, O

pen

Rec

ycle

if a

bove

25m

3 ”

Fro

m S

epte

mbe

r 20

09,

the

proc

edur

e on

how

to

use

V-3

138

was

te

mpo

rari

ly c

hang

ed to

w

ithd

raw

col

umn

botto

m li

quid

to V

-313

8 on

ly w

hen

T-10

8 co

lum

n is

sto

pped

. W

ithi

n th

e ne

xt f

ive

mon

ths,

the

effe

ctiv

enes

s of

this

pr

oced

ure

is c

onfi

rmed

. In

Jan

uary

201

0,

incl

udin

g ch

ange

s in

ta

nk r

ecyc

le (

depe

nds

on li

quid

vol

ume)

, the

V

-313

8 C

ontr

ol

Pro

cedu

res

was

rev

ised

.

The

tem

pora

ry p

roce

dure

w

as c

omm

unic

ated

in th

e O

pera

tion

Kno

wle

dge

Bas

e sy

stem

. The

no

tifi

cati

on o

f th

e pe

rman

ent V

-313

8 C

ontr

ol

Pro

cedu

re w

as o

nly

stop

ped

at O

pera

tion

Kno

wle

dge

Bas

e sy

stem

. It

has

not

ref

lect

ed in

op

erat

ion

man

uals

etc

.

50

(Tab

le 6

-2)

6M-5

E a

naly

sis

for

back

grou

nd o

f ac

cide

nt c

ontr

ibut

ing

fact

ors

and

recu

rren

ce p

reve

ntio

n co

unte

rmea

sure

s (2

)

Con

trib

utin

g fa

ctor

b)

-1)

R

ecyc

le to

the

Top

valv

e cl

osed

(V

-313

8 C

ontr

ol P

roce

dure

) B

ackg

roun

d of

acc

iden

t co

ntri

butin

g fa

ctor

s M

an

Mac

hine

M

etho

d M

ater

ial

Mea

sure

men

t M

anag

emen

t

Yea

r 20

10-2

012

The

pum

p st

rain

er c

lean

ing

freq

uenc

y ha

s dr

oppe

d sh

arpl

y w

hich

has

sig

nifi

cant

ly r

educ

ed

the

wor

kloa

d of

ope

rato

rs.

How

ever

aft

er th

ese,

op

port

uniti

es to

ope

rate

the

Rec

ycle

to T

op v

alve

wer

e ha

rdly

ex

peri

ence

d. O

pera

tors

hav

e be

com

e un

awar

e of

the

nece

ssit

y to

com

mis

sion

the

Rec

ycle

to T

op

valv

e.

Sw

itch

ing

the

feed

of

T-10

8 co

lum

ns b

otto

m li

quid

to

V-3

138

or to

the

T-70

1 co

lum

ns w

as c

arri

ed o

ut b

y op

erat

ing

the

head

er v

alve

in

diff

eren

t loc

atio

ns w

hich

wer

e aw

ay f

rom

the

P-3

138C

di

scha

rge

head

er. F

or th

is

reas

on, i

t mig

ht b

e di

ffic

ult

for

oper

ator

s to

not

ice

the

sign

age

pert

aini

ng to

Rec

ycle

to

the

Top

valv

e.

T

he e

ffec

tive

ness

was

ve

rifi

ed b

y th

e re

duce

d fr

eque

ncy

in s

trai

ner

clea

ning

. How

ever

, any

ch

ange

in th

e am

ount

of

depo

sitio

n in

side

the

tank

w

as n

ot c

heck

ed a

nd th

e ne

ed f

or R

ecyc

le to

the

Lev

el G

auge

was

not

re

-exa

min

ed.

Per

spec

tive

s of

C

ount

erm

easu

res

Edu

cati

on

E

duca

te o

pera

tors

the

new

de

velo

pmen

t in

V-3

138

(Spe

cifi

cati

on, d

esig

n co

nditi

ons)

.

Edu

cate

ope

rato

rs th

e ne

w d

evel

opm

ent i

n V

-313

8 (O

pera

ting

in

stru

ctio

ns, c

ontr

ol

stan

dard

, etc

.).

Edu

cate

ope

rato

rs

the

new

de

velo

pmen

t in

V-3

138

(Rec

onfi

rm li

quid

pr

oper

ties

, etc

.).

Edu

cate

ope

rato

rs th

e ne

w d

evel

opm

ent i

n V

-313

8 (T

est r

un r

esul

ts).

Eng

inee

ring

Stud

y th

e sp

ecif

icat

ions

of

V-3

138

and

its

asso

ciat

ed

equi

pmen

t (ne

w in

stal

lati

on)

(Con

trol

tem

pera

ture

, mea

ns

to c

ontr

ol, e

tc.)

R

isk

asse

ssm

ent f

or n

ew

inst

alla

tion

.

Upd

ate

the

P&

ID, e

tc.

arou

nd V

-313

8.

Stud

y th

e sp

ecif

icat

ions

of

V-3

138

and

its

asso

ciat

ed

equi

pmen

t (ne

w

inst

alla

tion

) (r

econ

firm

atio

n of

li

quid

pro

pert

ies)

.

Con

firm

the

oper

atin

g co

nditi

ons,

equ

ipm

ent

stat

us, e

tc. t

hrou

gh te

st

run.

Eva

luat

e sp

ecif

icat

ion

prop

riet

y of

V-3

138

and

its

asso

ciat

ed e

quip

men

t

Enf

orce

men

t C

onfi

rm th

e pr

oced

ures

wer

e ex

ecut

ed.

U

pdat

e op

erat

ion

man

uals

of

glac

ial

acry

lic

acid

pr

oduc

tion

faci

liti

es.

Ref

lect

met

hods

to

cont

rol i

nter

med

iate

tank

in

ope

rati

on m

anua

ls.

Env

iron

men

t E

nhan

ce th

e w

orks

’ rel

iabi

lity

(e

asy-

to-u

nder

stan

d si

gnag

e et

c.).

Pre

pare

sit

e si

gnag

e fo

r ne

w

inst

alla

tion

.

Exa

mpl

e

Est

abli

sh m

etho

ds a

nd

supp

lem

enta

ry m

easu

res

to

cont

rol t

he ta

nk te

mpe

ratu

re.

Rev

iew

the

equi

pmen

t bas

ed

on a

bove

est

abli

shm

ent.

Rev

iew

des

ign

stan

dard

s fo

r ta

nk

asso

ciat

ed e

quip

men

t

51

(Tab

le 6

-2)

6M-5

E a

naly

sis

for

back

grou

nd o

f ac

cide

nt c

ontr

ibut

ing

fact

ors

and

recu

rren

ce p

reve

ntio

n co

unte

rmea

sure

s (3

)

Con

trib

utin

g fa

ctor

b)

-2)

Rec

ycle

to th

e To

p va

lve

was

clo

sed

(T-6

701

load

up

test

)

Bac

kgro

und

of a

ccid

ent

cont

ribu

ting

fact

ors

Man

M

achi

ne

Met

hod

Mat

eria

l M

easu

rem

ent

Man

agem

ent

Bac

kgro

und

rela

ting

to

safe

wor

k m

anag

emen

t an

d op

erat

ing

inst

ruct

ions

Saf

ety

mea

sure

s (P

TW

, saf

e w

ork

proc

edur

e, e

tc.)

and

its

cont

ents

are

req

uire

d to

mak

e kn

own

to a

ll m

embe

rs b

efor

e an

y w

ork

is a

llow

ed to

sta

rt.

Bas

ical

ly, o

pera

ting

in

stru

ctio

ns w

ill b

e is

sued

wit

h P

rodu

ctio

n M

anag

er a

ppro

val

rega

rdle

ss w

heth

er th

ere

is a

ny

man

ual i

n pl

ace.

In

the

Pro

duct

ion

Dep

artm

ent,

PT

W is

ap

plie

d fo

r an

y w

orks

whi

ch

requ

ired

spe

cial

at

tent

ions

and

saf

ety

mea

sure

s. A

lso

PT

W

is a

ppli

ed to

“S

tand

ard

for

Non

-Reg

ular

Wor

ks”

whi

ch a

re n

ot

spec

ifie

d in

the

Stan

dard

Ope

ratin

g P

roce

dure

. In

reg

ards

to

oper

atio

n an

d sa

fe

wor

k m

anag

emen

t, th

ere

are

no r

ules

sp

ecif

ied

thin

gs s

uch

as a

ppro

val r

oute

, is

suan

ce o

f op

erat

ing

inst

ruct

ions

.

App

rova

l is

nece

ssar

y to

ca

rry

out w

orks

and

saf

e w

ork

proc

edur

es a

re

prep

ared

in a

dvan

ce f

or

each

dif

fere

nt ty

pe o

f w

orks

.

Pra

ctic

es o

n op

erat

ing

inst

ruct

ions

dif

fer

by e

ach

depa

rtm

ent b

ut g

ener

ally

th

ey a

re o

ften

issu

ed to

su

pple

men

t the

man

uals

.

Ris

k m

anag

emen

t as

soci

ated

wit

h ch

ange

s In

terp

reta

tion

s of

whe

ther

M

OC

is a

pplic

able

may

dif

fer

betw

een

indi

vidu

als.

If

they

th

ink

it is

not

app

lica

ble,

the

MO

C c

heck

she

et w

ill n

ot b

e us

ed.

T

he r

ule

of M

OC

is

appl

ied

for

risk

m

anag

emen

t as

soci

ated

wit

h pe

rman

ent o

r te

mpo

rary

cha

nges

in

equi

pmen

t and

pr

oces

ses,

etc

. C

hang

es w

hich

MO

C

is n

ot a

pplie

d ar

e al

so il

lust

rate

d w

ith

exam

ples

.

MO

C h

as b

een

impl

emen

ted

thro

ugh

data

base

. Any

cha

nges

are

ev

alua

ted

usin

g a

MO

C

chec

k sh

eet.

The

dep

th o

f re

view

and

app

rova

l rou

te

are

then

det

erm

ined

bas

ed

on th

e ch

eck

shee

t’s

resu

lt.

52

(Tab

le 6

-2)

6M-5

E a

naly

sis

for

back

grou

nd o

f ac

cide

nt c

ontr

ibut

ing

fact

ors

and

recu

rren

ce p

reve

ntio

n co

unte

rmea

sure

s (3

)

Con

trib

utin

g fa

ctor

b)

-2)

Rec

ycle

to th

e To

p va

lve

was

clo

sed

(T-6

701

load

up

test

) B

ackg

roun

d of

acc

iden

t co

ntri

butin

g fa

ctor

s M

an

Mac

hine

M

etho

d M

ater

ial

Mea

sure

men

t M

anag

emen

t

Loa

d up

test

in Y

ear

2009

T

he p

urpo

se o

f th

e T-

6701

load

up

test

was

to

con

firm

the

oper

atin

g co

nditi

ons

unde

r w

hich

th

e qu

alit

y of

dis

till

ate

coul

d be

ens

ured

. The

te

stin

g co

ndit

ions

wer

e th

ough

t to

be a

djus

ted

wit

hin

the

rang

e of

eq

uipm

ent c

apac

ity

base

d on

pas

t tra

ck r

ecor

ds.

The

pas

t tra

ck r

ecor

ds o

f co

lum

n lo

ad h

ave

the

T-67

01 te

stin

g co

ndit

ions

. T

he li

quid

vol

ume

to

stor

e in

V-3

138

was

w

ithi

n th

e ta

nk’s

nom

inal

ca

paci

ty.

The

ris

ks p

erta

inin

g to

st

orin

g liq

uid

in V

-313

8 w

ere

not i

nves

tiga

ted.

Test

ing

plan

indi

cati

ng

mat

ters

suc

h as

T-6

701

test

met

hod,

con

ditio

ns

and

sche

dule

was

pr

epar

ed.

A

t tha

t tim

e, th

e m

etho

ds

of p

rovi

sion

al u

se o

f V

-313

8 w

ere

veri

fied

and

th

e R

ecyc

le to

Top

was

co

mm

issi

oned

aft

er li

quid

bu

ilt u

p in

the

tank

.

The

inte

ntio

nall

y li

quid

bu

ilt u

p w

as n

ot a

nor

mal

bu

t non

-rou

tine

wor

k.

The

T-6

701

load

up

test

w

as a

n op

erat

ion

not

cove

red

in th

e op

erat

ion

man

uals

. Tes

ting

pla

n w

as

issu

ed w

ith

the

appr

oval

of

Pro

duct

ion

Man

ager

bu

t no

oper

atin

g in

stru

ctio

n w

as is

sued

.

Test

impl

emen

ted

in

fisc

al 2

012

The

test

was

und

erst

ood

sam

e as

test

car

ried

out

in

Yea

r 20

09.

Sam

e as

Yea

r 20

09, t

he

risk

s as

soci

ated

to b

uild

up

V-3

138

leve

l wer

e no

t in

vest

igat

ed.

The

test

ing

plan

was

pr

epar

ed in

the

sam

e m

anne

r as

Yea

r 20

09.

Sam

e as

Yea

r 20

09, n

o op

erat

ing

inst

ruct

ion

was

is

sued

.

Ope

rato

rs w

ere

not a

war

e of

the

“V-3

138

Con

trol

P

roce

dure

”.

Per

spec

tive

s of

C

ount

erm

easu

res

Edu

cati

on

Edu

cate

ris

k m

anag

emen

t fo

r no

n-ro

utin

e w

orks

and

ch

ange

s.

(Ris

k as

sess

men

t, K

Y,

etc.

)

E

duca

te a

bout

saf

e w

ork

man

agem

ent a

nd M

OC

.

(Rul

es, c

ase

stud

ies,

etc

.)

Eng

inee

ring

E

nfor

cem

ent

Con

firm

ris

k as

sess

men

t is

exe

cute

d fo

r no

n-ro

utin

e w

orks

and

ch

ange

s.

Stre

ngth

en r

isk

man

agem

ent s

yste

m

thro

ugh

educ

atio

n an

d pe

rson

nel d

evel

opm

ent.

E

stab

lish

sys

tem

to

ensu

re r

isk

asse

ssm

ent i

s co

nduc

ted

befo

re

oper

atio

n, is

suan

ce o

f op

erat

ing

inst

ruct

ions

and

it

s ap

prov

al r

oute

s, e

tc.

Rev

iew

MO

C s

yste

m

Rev

iew

ris

k ev

alua

tion

sc

ope

and

chec

k sh

eet

usag

e fo

r M

OC

.

Env

iron

men

t

E

xam

ple

Pre

pare

gen

eral

exa

mpl

es

for

MO

C

53

(Tab

le 6

-2)

6M-5

E a

naly

sis

for

back

grou

nd o

f ac

cide

nt c

ontr

ibut

ing

fact

ors

and

recu

rren

ce p

reve

ntio

n co

unte

rmea

sure

s (4

)

Con

trib

utin

g fa

ctor

c)

Inad

equa

te ta

nk c

ontr

ol te

mpe

ratu

re s

ettin

g C

ontr

ibut

ing

fact

or d

)

Inad

equa

te ta

nk te

mpe

ratu

re d

etec

tion

B

ackg

roun

d of

acc

iden

t co

ntri

butin

g fa

ctor

s M

an

Mac

hine

M

etho

d M

ater

ial

Mea

sure

men

t M

anag

emen

t

V-3

138

tem

pera

ture

co

ntro

l C

ompa

re to

the

mai

n pr

oces

s eq

uipm

ent,

ther

e w

as r

elat

ivel

y li

ttle

aw

aren

ess

of th

e ne

ed f

or

quan

titat

ive

tem

pera

ture

co

ntro

l in

V-3

138.

R

isk

info

rmat

ion

asso

ciat

ed to

D

AA

for

mat

ion

was

not

wid

ely

shar

ed.

The

tem

pera

ture

of

mai

n pr

oces

s un

its

in

acry

lic

acid

pla

nt n

eed

to b

e pr

ecis

ely

cont

roll

ed.

T

he V

-313

8 liq

uid

coul

d be

coo

led

by th

e co

olin

g co

il. H

ence

th

e pr

ecis

e co

ntro

l of

tem

pera

ture

was

not

ne

cess

ary.

W

ith

rega

rd to

V-3

138

liqu

id, t

here

are

co

ncer

ns o

f ac

ryli

c ac

id f

reez

ing

and

slud

ge p

reci

pita

tion

at

low

tem

pera

ture

s an

d po

lym

eriz

atio

n at

hig

h te

mpe

ratu

res

but t

he

liqu

id h

as h

igh

cont

ents

of

inhi

bito

rs.

Inhi

bito

rs a

nd th

e at

mos

pher

e co

ndit

ion

are

not a

ble

to p

reve

nt

form

atio

n of

DA

A,

whe

reas

the

tem

pera

ture

has

a b

ig

impa

ct.

The

tank

tem

pera

ture

has

to

be m

aint

aine

d w

ithi

n a

rang

e.

Nec

essi

ty o

f in

stru

men

t in

stal

lati

on

Whe

n th

e ne

ed f

or te

mpe

ratu

re

cont

rol i

s as

sess

ed a

s lo

w,

tem

pera

ture

wil

l not

be

mon

itor

ed c

onti

nuou

sly.

Whe

n in

term

itte

nt

dete

ctio

n is

suf

fici

ent,

ther

mom

eter

s w

ill n

ot

be in

stal

led.

The

nee

d to

inst

all a

th

erm

omet

er a

nd

how

to u

se it

is

esta

blis

hed

base

d on

m

atte

rs s

uch

as th

e ne

cess

ity

of

tem

pera

ture

con

trol

, th

e te

mpe

ratu

re

mon

itor

ing

cycl

e an

d le

gal r

equi

rem

ents

.

The

pla

nt h

as n

o st

anda

rd o

n m

eans

of

tem

pera

ture

m

onit

orin

g.

Ass

essm

ent f

or th

e ne

ed o

f te

mpe

ratu

re c

ontr

ol a

nd

mon

itor

ing

was

car

ried

out

by

eac

h pl

ant i

ndiv

idua

lly.

T

his

has

crea

ted

vari

atio

n in

the

asse

ssm

ent a

nd

awar

enes

s of

tem

pera

ture

co

ntro

l and

mon

itor

ing.

54

(Tab

le 6

-2)

6M-5

E a

naly

sis

for

back

grou

nd o

f ac

cide

nt c

ontr

ibut

ing

fact

ors

and

recu

rren

ce p

reve

ntio

n co

unte

rmea

sure

s (4

)

Con

trib

utin

g fa

ctor

c)

Inad

equa

te ta

nk c

ontr

ol te

mpe

ratu

re s

ettin

g C

ontr

ibut

ing

fact

or d

)

Inad

equa

te ta

nk te

mpe

ratu

re d

etec

tion

Bac

kgro

und

of a

ccid

ent

cont

ribu

ting

fact

ors

Man

M

achi

ne

Met

hod

Mat

eria

l M

easu

rem

ent

Man

agem

ent

V-3

138

ther

mom

eter

: de

sign

and

con

stru

ctio

n st

ages

T

herm

omet

er in

stal

lati

on

was

not

pla

nned

at t

he

V-3

138

desi

gn s

tage

. T

herm

omet

er n

ozzl

e w

as

adde

d as

the

proj

ect

mov

ed in

to d

etai

led

desi

gn a

nd c

onst

ruct

ion

stag

es.

T

here

wer

e no

rec

ords

av

aila

ble

that

can

ac

cura

tely

iden

tify

the

inte

ntio

n on

how

to u

se

the

nozz

le o

r th

e ex

iste

nce

of a

th

erm

omet

er w

hen

the

equi

pmen

t com

mis

sion

ed.

A

bout

the

V-3

138

ther

mom

eter

: rel

atio

n to

la

tera

l dep

loym

ent

foll

owin

g a

trou

ble

Eve

n th

ough

V-3

138

was

fo

und

of n

ot h

avin

g th

erm

omet

er in

stal

led,

it

was

not

incl

uded

in th

e la

tera

l dep

loym

ent s

cope

.

Aft

er p

olym

eriz

atio

n tr

oubl

e oc

curr

ed in

an

othe

r ac

ryli

c ac

id

inte

rmed

iate

tank

(in

A

pril

199

4), t

he p

ropo

sed

coun

term

easu

res

incl

uded

im

plem

enti

ng c

onti

nuou

s te

mpe

ratu

re m

onit

orin

g an

d in

stal

ling

ext

erna

l he

at e

xcha

nger

s si

mil

ar

tank

s ot

her

than

the

tank

w

here

the

trou

ble

occu

rred

.

The

equ

ipm

ent s

ubje

ct to

th

e la

tera

l dep

loym

ent

was

tank

s th

at r

ecei

ve

colu

mn

botto

m li

quid

at

80～

100°

C.

T

he b

otto

m te

mpe

ratu

res

of T

-108

col

umns

wer

e on

ly a

bout

65～

70°C

.

(T-5

108

had

not y

et b

een

cons

truc

ted

at th

at ti

me)

.

The

cau

ses

inve

stig

atio

n an

d co

unte

rmea

sure

s im

plem

enta

tion

for

this

tr

oubl

e w

ere

conc

lude

d on

ly w

ithi

n th

e in

volv

ed

depa

rtm

ent.

The

re w

as n

o sy

stem

in p

lace

to in

volv

e pa

rtic

ipat

ion

of o

ther

de

part

men

ts li

kes

Tech

nolo

gy in

in

vest

igat

ion

for

prev

enti

ng s

imil

ar tr

oubl

e in

oth

er a

reas

.

Per

spec

tive

s of

C

ount

erm

easu

res

Edu

cati

on

E

duca

te o

pera

tors

the

new

dev

elop

men

t in

V-3

138

(V-3

138

and

its

asso

ciat

ed e

quip

men

t)

Edu

cate

ope

rato

rs th

e ne

w d

evel

opm

ent i

n V

-313

8 (O

pera

ting

inst

ruct

ions

, co

ntro

l sta

ndar

d, e

tc.)

.

Edu

cate

ope

rato

rs th

e ne

w d

evel

opm

ent i

n V

-313

8 (L

iqui

d pr

oper

ties

, DA

A

beha

vior

, etc

.)

Edu

cate

ope

rato

rs th

e ne

w d

evel

opm

ent i

n V

-313

8 (T

est r

un r

esul

ts)

Eng

inee

ring

D

evel

op e

ngin

eer

and

impr

ove

tech

nica

l ca

pabi

liti

es th

roug

h tr

oubl

e in

vest

igat

ion

invo

lvem

ent.

Stud

y th

e sp

ecif

icat

ions

of

V-3

138

and

its

asso

ciat

ed e

quip

men

t (n

ew in

stal

lati

on)

(Con

trol

tem

pera

ture

, m

eans

to c

ontr

ol, e

tc.)

R

isk

asse

ssm

ent f

or n

ew

inst

alla

tion

.

Upd

ate

the

P&

ID, e

tc.

arou

nd V

-313

8.

Stud

y th

e sp

ecif

icat

ions

of

V-3

138

and

its

asso

ciat

ed e

quip

men

t (n

ew in

stal

lati

on)

(rec

onfi

rmat

ion

of li

quid

pr

oper

ties

).

Exp

lain

the

deta

ils

of

DA

A f

orm

atio

n.

Con

firm

the

oper

atin

g co

nditi

ons,

sta

tus,

etc

. th

roug

h te

st r

un.

Eva

luat

e sp

ecif

icat

ion

prop

riet

y of

V-3

138

and

its

asso

ciat

ed e

quip

men

t

55

(Tab

le 6

-2)

6M-5

E a

naly

sis

for

back

grou

nd o

f ac

cide

nt c

ontr

ibut

ing

fact

ors

and

recu

rren

ce p

reve

ntio

n co

unte

rmea

sure

s (4

)

Con

trib

utin

g fa

ctor

c)

Inad

equa

te ta

nk c

ontr

ol te

mpe

ratu

re s

ettin

g C

ontr

ibut

ing

fact

or d

)

Inad

equa

te ta

nk te

mpe

ratu

re d

etec

tion

Bac

kgro

und

of a

ccid

ent

cont

ribu

ting

fact

ors

Man

M

achi

ne

Met

hod

Mat

eria

l M

easu

rem

ent

Man

agem

ent

Per

spec

tive

s of

C

ount

erm

easu

res

Enf

orce

men

t

Stan

dard

ize

tank

te

mpe

ratu

re m

onit

orin

g (t

herm

omet

er

inst

alla

tion

).

Dev

elop

the

oper

atio

n m

anua

ls f

or g

laci

al

acry

lic

acid

pro

duct

ion

faci

liti

es.

Ref

lect

met

hods

to

cont

rol i

nter

med

iate

tank

in

ope

rati

on m

anua

ls.

Rev

iew

the

syst

em o

f la

tera

l dep

loym

ent

foll

owin

g a

trou

ble

Env

iron

men

t

Pre

pare

sit

e si

gnag

e fo

r ne

w in

stal

lati

on.

Exa

mpl

e

Est

abli

sh m

etho

ds a

nd

supp

lem

enta

ry m

easu

res

to c

ontr

ol th

e ta

nk

tem

pera

ture

. R

evie

w th

e eq

uipm

ent

base

d on

abo

ve

esta

blis

hmen

t.

Rev

iew

des

ign

stan

dard

s fo

r ta

nk a

ssoc

iate

d eq

uipm

ent

56

(Tab

le 6

-2)

6M-5

E a

naly

sis

for

back

grou

nd o

f ac

cide

nt c

ontr

ibut

ing

fact

ors

and

recu

rren

ce p

reve

ntio

n co

unte

rmea

sure

s (5

)

Con

trib

utin

g fa

ctor

e)

Una

ble

to c

ontr

ol th

e te

mpe

ratu

re w

ithin

nor

mal

ran

ge

Con

trib

utin

g fa

ctor

f)

Una

ble

to c

ool d

own

the

tem

pera

ture

bel

ow th

e up

per

limit

C

ontr

ibut

ing

fact

or g

) U

nabl

e to

avo

id a

bnor

mal

sit

uatio

ns

B

ackg

roun

d of

ac

cide

nt c

ontr

ibut

ing

fact

ors

Man

M

achi

ne

Met

hod

Mat

eria

l M

easu

rem

ent

Man

agem

ent

Acr

ylic

aci

d sa

fety

m

easu

res

over

all

In

acr

ylic

aci

d pl

ant,

if

any

abno

rmal

ity

is

dete

cted

and

pre

vent

th

e no

rmal

ope

rati

on

to c

onti

nue,

the

plan

t w

ill b

e sh

utdo

wn

thro

ugh

inte

rloc

k sy

stem

.

Em

erge

ncy

crit

eria

an

d re

spon

se

proc

edur

es w

ere

not

esta

blis

hed

for

the

abno

rmal

sit

uati

ons

whi

ch c

anno

t be

cont

roll

ed e

ven

if th

e eq

uipm

ent i

s sh

utdo

wn.

Con

vent

iona

lly,

the

safe

ty

mea

sure

s fo

cuse

d on

m

easu

res

to p

reve

nt

trou

ble.

V-3

138

It w

as n

ot p

ossi

ble

to d

etec

t the

te

mpe

ratu

re h

as e

xcee

ded

unti

l th

e eq

uipm

ent h

as b

een

shut

do

wn

or s

ome

sort

of

acti

ons

had

been

take

n.

(Ina

dequ

ate

tem

pera

ture

de

tect

ion)

. T

here

was

no

equi

pmen

t to

hand

le

the

abno

rmal

si

tuat

ions

.

(Ina

dequ

ate

cont

rol

tem

pera

ture

set

ting

).

The

re w

ere

no

crit

eria

and

mea

ns

for

hand

ling

abno

rmal

sit

uati

ons

esta

blis

hed.

Per

spec

tive

s of

C

ount

erm

easu

res

Edu

cati

on

Mai

ntai

n an

d im

prov

e th

e em

erge

ncy

resp

onse

cap

abil

itie

s th

roug

h tr

aini

ng.

Edu

cate

and

trai

n op

erat

ors

on r

espo

nse

proc

edur

es f

or V

-313

8 ab

norm

al s

itua

tion

s.

(Spe

cifi

cati

ons,

des

ign

cond

ition

s, e

tc.)

.

Edu

cate

and

trai

n op

erat

ors

on

resp

onse

pro

cedu

res

for

V-3

138

abno

rmal

si

tuat

ions

.

(Ope

rati

ng

inst

ruct

ions

, con

trol

st

anda

rd, e

tc.)

.

Edu

cate

and

trai

n op

erat

ors

on r

espo

nse

proc

edur

es f

or V

-313

8 ab

norm

al s

itua

tion

s.

(Han

dlin

g of

em

erge

ncy

inhi

bito

rs).

Con

firm

the

requ

ired

tim

e of

res

pons

e pr

oced

ure

for

abno

rmal

sit

uati

ons

by

trai

ning

.

Impl

emen

t the

pla

nned

em

erge

ncy

resp

onse

dri

lls

57

(Tab

le 6

-2)

6M-5

E a

naly

sis

for

back

grou

nd o

f ac

cide

nt c

ontr

ibut

ing

fact

ors

and

recu

rren

ce p

reve

ntio

n co

unte

rmea

sure

s (5

)

Con

trib

utin

g fa

ctor

e)

Una

ble

to c

ontr

ol th

e te

mpe

ratu

re w

ithin

nor

mal

ran

ge

Con

trib

utin

g fa

ctor

f)

Una

ble

to c

ool d

own

the

tem

pera

ture

bel

ow th

e up

per

limit

C

ontr

ibut

ing

fact

or g

) U

nabl

e to

avo

id a

bnor

mal

sit

uatio

ns

Bac

kgro

und

of a

ccid

ent

cont

ribu

ting

fact

ors

Man

M

achi

ne

Met

hod

Mat

eria

l M

easu

rem

ent

Man

agem

ent

Per

spec

tive

s of

C

ount

erm

easu

res

Eng

inee

ring

D

evel

op e

ngin

eer

and

impr

ove

tech

nica

l ca

pabi

liti

es th

roug

h an

alyz

ing

exte

rnal

in

cide

nts.

Eva

luat

e re

spon

se

proc

edur

es f

or V

-313

8 ab

norm

al s

itua

tion

s

(Tem

pera

ture

cri

teri

a,

emer

genc

y in

hibi

tor

equi

pmen

t, et

c.)

Eva

luat

e re

spon

se

proc

edur

e fo

r ab

norm

al

situ

atio

ns f

or e

quip

men

t ha

ndlin

g ac

ryli

c ac

id

(Tem

pera

ture

cri

teri

a,

equi

pmen

t spe

cifi

cati

ons,

et

c.)

Con

duct

ris

k as

sess

men

t on

res

pons

e pr

oced

ures

fo

r ab

norm

al s

itua

tion

s.

Eva

luat

e re

spon

se

proc

edur

es f

or V

-313

8 ab

norm

al s

itua

tion

s

(Ope

rati

ng in

stru

ctio

ns,

cont

rol s

tand

ard,

etc

.)

Eva

luat

e re

spon

se

proc

edur

e fo

r ab

norm

al

situ

atio

ns f

or e

quip

men

t ha

ndlin

g ac

ryli

c ac

id

(Ope

rati

ng in

stru

ctio

ns,

cont

rol s

tand

ard,

etc

.)

Dev

elop

the

crit

eria

for

ju

dgin

g ab

norm

al

sym

ptom

and

the

basi

c co

ncep

ts o

f th

e re

spon

se

proc

edur

es.

Eva

luat

e re

spon

se

proc

edur

es f

or V

-313

8 ab

norm

al s

itua

tion

s

(Sel

ecti

on o

f em

erge

ncy

inhi

bito

rs a

nd u

nder

stan

d it

s pr

oper

ties

, etc

.)

Eva

luat

e re

spon

se

proc

edur

e fo

r ab

norm

al

situ

atio

ns f

or e

quip

men

t ha

ndlin

g ac

ryli

c ac

id

(Sel

ecti

on o

f em

erge

ncy

inhi

bito

rs a

nd u

nder

stan

d it

s pr

oper

ties

, etc

.)

E

valu

ate

the

prop

riet

y of

V

-313

8 ab

norm

al

situ

atio

ns c

rite

ria,

re

spon

se p

roce

dure

s.

Eva

luat

e th

e pr

opri

ety

of

equi

pmen

t han

dlin

g ac

ryli

c ac

id a

bnor

mal

si

tuat

ions

cri

teri

a,

resp

onse

pro

cedu

res.

R

evie

w e

quip

men

t bas

ed

on th

e de

sign

con

cept

s.

Enf

orce

men

t St

reng

then

the

impl

emen

tati

on o

f da

ily

haza

rd e

valu

atio

n

D

evel

op V

-313

8 m

anua

ls

and

othe

r do

cum

enta

tion

s.

Rev

iew

ope

rati

on

man

uals

of

equi

pmen

t ha

ndlin

g ac

ryli

c ac

id.

Env

iron

men

t

Pre

pare

sit

e si

gnag

e fo

r eq

uipm

ent i

nsta

llat

ion.

Exa

mpl

e

Col

lect

and

mak

e fu

ll u

se

of th

e ex

tern

al in

cide

nt

case

stu

dies

and

tech

nica

l in

form

atio

n.

58

(Tab

le 6

-2)

6M-5

E a

naly

sis

for

back

grou

nd o

f ac

cide

nt c

ontr

ibut

ing

fact

ors

and

recu

rren

ce p

reve

ntio

n co

unte

rmea

sure

s (6

)

Con

trib

utin

g fa

ctor

h)

Una

ble

to a

void

cri

sis

situ

atio

ns

B

ackg

roun

d of

acc

iden

t co

ntri

butin

g fa

ctor

s M

an

Mac

hine

M

etho

d M

ater

ial

Mea

sure

men

t M

anag

emen

t

Pas

t tro

uble

Him

eji P

lant

has

not

ex

peri

ence

d a

tank

ru

ptur

ed o

r ex

plos

ion

due

to a

cryl

ic a

cid

poly

mer

izat

ion.

A

cryl

ic a

cid

poly

mer

s ha

ve b

een

disc

over

ed

duri

ng ta

nk in

tern

al

insp

ecti

on a

nd b

asic

m

easu

res

to p

reve

nt

poly

mer

izat

ion

have

be

en im

plem

ente

d fo

r ea

ch ta

nk.

Mea

sure

s ta

ken

whe

n ac

ryli

c ac

id

poly

mer

ized

in th

e ta

nks

incl

uded

two

inst

ance

s of

spr

ayin

g w

ater

ont

o th

e ta

nk

and

one

inst

ance

of

char

ging

wat

er a

nd

inhi

bito

rs f

rom

the

man

hole

. Oth

er

mea

sure

s w

hich

hav

e ta

ken

incl

uded

ch

argi

ng c

oolin

g w

ater

in

to th

e ta

nk ja

cket

.

Act

ivit

ies

rela

ting

to

resp

onse

s to

abn

orm

al

phen

omen

a

Oth

er th

an th

e co

mpl

ex-w

ide

disa

ster

pr

even

tion

trai

ning

s, e

ach

of th

e va

riou

s w

orkp

lace

s w

ill a

lso

impl

emen

t its

ow

n pl

anne

d di

sast

er

prev

enti

on e

xerc

ises

. One

of

thes

e tr

aini

ngs

is th

e fi

re

figh

ting

exe

rcis

es w

hich

in

clud

e a

year

ly f

ire

figh

ting

com

peti

tion

.

To im

prov

e th

e di

sast

er

prev

enti

on c

apab

ilit

ies,

th

e pl

ant h

as a

lso

build

ing

up a

goo

d su

pply

of

the

disa

ster

pr

even

tion

equ

ipm

ent.

Imm

edia

te r

espo

nse

proc

edur

es f

or

abno

rmal

phe

nom

ena

have

bee

n de

velo

ped

in th

e P

lant

’s

Em

erge

ncy

Res

pons

e P

lan

but n

one

of th

ese

cove

red

for

runa

way

re

acti

on. (

In th

e tr

aini

ng m

ater

ials

re

lati

ng to

abn

orm

al

phen

omen

a an

d al

erti

ng p

roce

dure

(in

th

e pr

oces

s fo

r ap

prov

al),

ther

e w

as

illu

stra

tion

to c

ool t

he

tank

by

spra

ying

wat

er

as a

way

of

resp

ondi

ng

to r

unaw

ay r

eact

ion)

.

Wit

h th

e ai

m o

f im

prov

ing

the

disa

ster

pre

vent

ion

capa

bili

ties

in th

e H

MI

divi

sion

(sa

fety

team

),

vari

ous

acti

viti

es h

ave

been

con

duct

ed to

geth

er

wit

h th

e R

C a

ctiv

itie

s si

nce

Yea

r 20

07.

59

(Tab

le 6

-2)

6M-5

E a

naly

sis

for

back

grou

nd o

f ac

cide

nt c

ontr

ibut

ing

fact

ors

and

recu

rren

ce p

reve

ntio

n co

unte

rmea

sure

s (6

)

Con

trib

utin

g fa

ctor

h)

Una

ble

to a

void

cri

sis

situ

atio

ns

B

ackg

roun

d of

acc

iden

t co

ntri

butin

g fa

ctor

s M

an

Mac

hine

M

etho

d M

ater

ial

Mea

sure

men

t M

anag

emen

t

Stat

us o

f ru

les

deve

lopm

ent

The

re w

as n

o m

anua

l to

sta

ndar

dize

eac

h pl

ant c

rite

ria

for

abno

rmal

ity

and

its

resp

ecti

ve r

espo

nse

proc

edur

es.

The

rol

e an

d re

spon

sibi

lity

on

prov

idin

g in

form

atio

n to

pu

blic

fir

e de

part

men

t w

as n

ot c

lear

ly d

efin

ed in

se

lf d

isas

ter

prev

enti

on

team

. R

espo

nses

dur

ing

this

ac

cide

nt

Tank

coo

ling

was

judg

ed to

be

nec

essa

ry b

ased

on

the

V-3

138

cond

ition

s an

d he

nce

the

tank

was

spr

ayed

w

ith

wat

er. T

he in

tent

ion

was

to c

ool a

nd to

abs

orb

the

acry

lic

acid

vap

or.

Ope

rato

rs s

pray

ed

wat

er f

rom

six

fir

e hy

dran

ts (

each

at 4

00

lite

rs/m

in.)

. Lat

er, t

he

self

dis

aste

r pr

even

tion

team

sp

raye

d ad

ditio

nal

wat

er (

3,10

0 li

ters

/min

) fr

om th

e fi

re e

ngin

e.

Sin

ce a

cryl

ic a

cid

vapo

r w

as b

eing

re

leas

ed, a

ddit

iona

l m

easu

res

such

as

char

ging

wat

er a

nd

inhi

bito

rs in

side

V

-313

8 co

uld

not b

e ta

ken.

The

re w

as n

o m

eans

to c

ontr

ol th

e ab

norm

al s

itua

tion

s pr

ogre

ss o

ther

than

sp

rayi

ng w

ater

.

V-3

138

was

insu

late

d an

d it

was

ass

umed

th

at w

ater

spr

ayin

g di

d no

t hav

e m

uch

cool

ing

effe

ct to

the

tank

liqu

id b

ut m

ay

have

pre

vent

ed v

apor

cl

oud

expl

osio

n.

Per

spec

tive

s of

C

ount

erm

easu

res

Edu

cati

on

Mai

ntai

n an

d im

prov

e th

e em

erge

ncy

resp

onse

ca

pabi

liti

es th

roug

h tr

aini

ng.

Pro

vide

edu

cati

on

and

trai

ning

bas

ed o

n cr

isis

man

agem

ent/

se

lf d

isas

ter

prev

enti

on m

anua

l

(Spe

cifi

cati

ons,

de

sign

con

ditio

ns,

etc.

)

Pro

vide

edu

cati

on a

nd

trai

ning

bas

ed o

n cr

isis

man

agem

ent/

se

lf d

isas

ter

prev

enti

on m

anua

l.

(Im

prov

e im

med

iate

ac

tion

s, e

tc.)

Re-

educ

ate

the

haza

rds

of h

andl

e m

ater

ials

.

Con

firm

the

requ

ired

tim

e of

res

pons

e pr

oced

ure

for

abno

rmal

sit

uati

ons

by

trai

ning

.

Impl

emen

t pla

ns f

or

educ

atio

n an

d tr

aini

ng.

Eng

inee

ring

D

evel

op e

ngin

eer

and

impr

ove

tech

nica

l ca

pabi

liti

es th

roug

h an

alyz

ing

exte

rnal

in

cide

nts.

Eva

luat

e th

e re

spon

se

proc

edur

es f

or c

risi

s si

tuat

ions

(T

empe

ratu

re c

rite

ria,

sp

ecif

icat

ions

, etc

.)

Con

duct

ris

k as

sess

men

t on

resp

onse

pro

cedu

res.

Eva

luat

e th

e re

spon

se

proc

edur

es f

or c

risi

s si

tuat

ions

(O

pera

ting

in

stru

ctio

ns, c

ontr

ol

stan

dard

, etc

.)

Eva

luat

e th

e pr

opri

ety

of

resp

onse

pro

cedu

res.

60

(Tab

le 6

-2)

6M-5

E a

naly

sis

for

back

grou

nd o

f ac

cide

nt c

ontr

ibut

ing

fact

ors

and

recu

rren

ce p

reve

ntio

n co

unte

rmea

sure

s (6

)

Con

trib

utin

g fa

ctor

h)

Una

ble

to a

void

cri

sis

situ

atio

ns

Bac

kgro

und

of a

ccid

ent

cont

ribu

ting

fact

ors

Man

M

achi

ne

Met

hod

Mat

eria

l M

easu

rem

ent

Man

agem

ent

Per

spec

tive

s of

C

ount

erm

easu

res

Enf

orce

men

t

D

evel

op c

risi

s m

anag

emen

t man

ual

Rev

iew

sel

f di

sast

er

prev

enti

on m

anua

l

Rev

iew

ope

rati

on

man

uals

Env

iron

men

t

Rev

iew

dis

aste

r pr

even

tion

eq

uipm

ent.

Exa

mpl

e

Col

lect

and

mak

e fu

ll u

se

of th

e ex

tern

al in

cide

nt

case

stu

dies

and

tech

nica

l in

form

atio

n.

61

Afterword

The explosion and fire that occurred in the Nippon Shokubai’s Himeji Plant occurred from a chain of

events, starting with an increase in the amount of liquid stored in the intermediate tank that was used to

temporarily store the bottom liquid of the glacial acrylic acid rectifying column in the acrylic acid

production facilities. In spite of the increase in stored liquid, the Recycle to Top was not commissioned, so

that the cooling was insufficient in the tank. This accelerated the reaction to form acrylic acid dimer and

caused the temperature to rise, which in turn triggered an abnormality by which the polymerization of

acrylic acid occurred. Deficiencies in the temperature detection and temperature monitoring of the liquid

stored in the tank led to the explosion and fire.

Nippon Shokubai must take to its heart that the accident at the Himeji Plant triggered significant

casualties and return to its fundamental corporate commitment of “Safety takes priority over production”

and work on restoring the company to regain public trust as a chemical company. To achieve this, the

company must make certain to implement the recurrence prevention countermeasures recommended by the

Accident Investigation Committee. At the same time, it is strongly recommended that Nippon Shokubai

practically carried out the activities by third party evaluation in order to develop an effective framework for

safety management.

The recurrence prevention countermeasures recommended by the Accident Investigation Committee

should be considered to implement in the other Plants in addition to Himeji Plant. We urge the company to

laterally deploy these countermeasures at other production facilities, in order to improve the overall

framework for safety management at Nippon Shokubai.

Additionally, we hope that the accident investigation report will be utilized to prevent accidents in

similar and other processes and contribute in some way to improve the safety of the chemical industry.

Finally throughout the progess of investigating the accident, I would like to express my deep gratitude

to the members of the Accident Investigation Committee for their valuable input and the persons who

conducted the active investigation. I also wish to thank the following people and organizations for their

efforts in investigating the sequence of events that led to the explosion and fire and the impact of the

accident: Toyo Engineering Corporation for their assistance with fluid analysis, IHI Corporation for their

assistance with structural analysis and blast pressure investigation and Dr. Yoshio Nakayama of the

National Institute of Advanced Industrial Science and Technology (AIST) for verifying the validity of the

investigation and analysis results.

62

My deepest gratitude also goes to the personnel of the Fire and Disaster Management Agency,

National Research Institute of Fire and Disaster and Himeji City Fire Department for their extensive advice

in helping us to wrap up the accident investigation and accident investigation report.

March 2013

Masamitsu Tamura

Chairman

Accident Investigation Committee

Nippon Shokubai Co., Ltd.

63

Appendix 1 Confirmation by Adiabatic Reaction Test

(1) Purpose: Confirmed the following items concerning V-3138 liquid (T-5108 bottom liquid) using

adiabatic reaction test facilities.

a) Liquid composition curve at start temperatures of 80℃, 90℃, and 100℃.

b) Behavior of runaway reaction at start temperature of 90℃.

c) Difference in temperature curves due to start temperature and atmosphere

(2) Test methods and conditions

Tests were conducted by both the Nippon Shokubai laboratory and enclosed testing room in Asa Plant of

Kayaku Japan Co., Ltd. In both tests, sample was placed in an insulated container inside hot air oven and

the sample was heated to raise its temperature. The oven temperature was tracked in order to simulate the

adiabatic condition. (Temperature difference between the sample and the oven was kept within 1℃.) Refer

to Appendix 4/15 to 6/15 for the details of respective tests equipment and conditions.

a) Test at Nippon Shokubai laboratory (for purposes: a) and c))

Analyzed the initial and progressive liquid sample compositions took from the sampling nozzle.

To prevent any abnormal reaction, an emergency used solvent was added to the insulated

container when the sample temperature reached 130℃ and stopped the test.

b) Tests at the enclosed testing room in Asa Plant of Kayaku Japan Co., Ltd. (for purpose: b) and c))

The enclosed testing room cannot be entered for safety reasons when sample temperatures

reached above 140℃, hence the oven temperature was not able to raise beyond this point. (oven

temperature unable to follow the rise in sample temperature)

(3) Test results (Acrylic acid: AA, acrylic acid dimer: DAA, acrylic acid trimer: TAA)

a) Liquid composition curve at start temperatures of 80℃, 90℃, and 100℃

Start temperature: 100℃ (under M-Gas, N2 atmospheres)

- There were slight differences in AA, DAA, and TAA composition but no polymers were

formed.

- There was little change in MQ concentration but PTZ concentration decreased over time for

both cases.

Start temperature: 90℃ (under N2 atmosphere)

- There was little change in AA, DAA, and TAA composition at the end of the test, compared

to the test with start temperature of 100℃. A little amount of polymers formed at the end of

the test.

- Inhibitor behavior was the same as the test with start temperature of 100℃ but the PTZ

concentration at the end of the test was lower than the test with start temperature of 100℃.

Probably this was due to the longer elapsed time.

Start temperature: 80℃ (N2 atmosphere)

- Same as the others tests, the DAA and TAA concentrations increased over time. However, the

DAA concentration peaked at about 40wt% and the TAA concentration increased to about

13wt%.

1/15

- Upon reaching 90℃, DAA has already reached about 21wt%. The behavior after this point

differed from the test with start temperature of 90℃ (The same observation as the test with

start temperature of 100℃).

- The PTZ concentration decreased significantly and polymers gradually formed starting from

the end stage of the test. Probably this was due to the elapsed time was much longer than

other tests (There was little change in the MQ concentration).

b) Behavior of runaway reaction at start temperature of 90℃.

Temperature curves and their comparison

- The temperature curve up to 130℃ was almost the same as laboratory test results.

- After this point, the temperature increased sharply but the rise in temperature stopped at

about 190℃ and began to decrease.

Changes in composition and weight before and after test

Weight Analyzed Composition [wt%] Inhibitor Concentration [ppm]

[g] AA DAA TAA Polymers PTZ MQ

Before 1,503 98.8 3.2 0.0 0.0 215 1,527

After 1,439 20.1 22.2 16.6 41.7 30 937

Weight decreased by 64g ⇒ 4.3% compared with weight before the test

- The weight decreased before and after the test was presumed to be due to the evaporated

sample released out from the system.

# It is presumed there was no pressure increased within the actual testing facilities

because any evaporated vapor due to temperature rise in sample is released out of the

system from the exhaust vent.

- After the test, about 40wt% of polymers was formed. Although the rise in temperature

stopped at about 190℃, it was obvious that polymerization itself has progressed.

- There were black and white polymers floating in the Dewar flask. Also polymers had

adhered to the vent tube and lid due to polymerization of condensate.

Observations

- From the composition and weight of the contents after the tests, the sample will reach an

estimated temperature of 260~270℃ if the heat generated from the reaction is entirely used

for increasing the temperature of sample itself.

- Reasons of the rise in temperature stopped before reaching the above temperatures was

presumed to be due to the reaction heat being consumed by the following factors.

• Heat dissipation at temperature range above 140℃

• Latent heat of liquid evaporation

# The released samples due to evaporation were 4.3% of the initial prepared amount.

Since condensate and its polymer were found in the vent tube, it is assumed that the

heat was consumed by the condensate returning to the Dewar flask and re-evaporating

2/15

into the vent tube.

c) Difference in temperature curves due to start temperature and atmosphere

Difference in temperature curves due to starting temperatures

- Although there were slight differences at each start temperature, the values calculated from

the actual temperatures and compositions curves were mostly matched.

- The required times to reach 130℃ were as follows:

• Start temperature 100℃: 7~8 hours

• Start temperature 90℃: about 24 hours

• Start temperature 80℃: about 63 hours

⇒ Up to 130℃, the increased temperature was assumed to be mainly due to heat of formation

from DAA and TAA.

Difference in temperature curves due to atmosphere (N2 atmosphere, M-Gas atmosphere)

- In the Kayaku Japan’s test, there was almost no difference in temperature curves for both

atmospheres beyond 130℃.

- In the Nippon Shokubai laboratory’s test, there was almost no difference in temperature

curves for both atmospheres. Hence the reproducibility was considered good.

⇒Through a series of reactions, it has not found any difference in temperature curves due to

different atmospheres. Therefore, it is presumed that the atmosphere inside the tank did not

influence the accident in the Himeji Plant.

3/15

Nippon Shokubai Adiabatic Reaction Laboratory Test Equipment

A. Thermostat equipment: Laboratory Constant Temperature Oven, DK240S by Yamato Scientific Co., Ltd.

B. Adiabatic container: Dewar flask, D-2000 (2 liter) by Thermo Scientific

*whenever sample temperature increased by 1℃, the oven internal temperature is also tracked

and increased by 1℃ (ΔT: Within 1℃)

C. Container lid

D. Thermometer Resistance temperature detector (Pt)

4/15

Kayaku Japan Co., Ltd. Adiabatic Reaction Test Equipment

A. Thermostat equipment: Laboratory Constant Temperature Oven, DK240S by Yamato Scientific Co., Ltd.

Weight: 23 kg, Power consumption: AC100V 4.5A (15A)

B. Adiabatic container: Dewar flask, D-2000 (2 liter) by Thermo Scientific

*whenever sample temperature increased by 1℃, the oven internal temperature is also tracked

and increased by 1℃ (ΔT: Within 1℃)

C. Container lid

D. Thermometer Resistance temperature detector (Pt)

5/15

Facility Layout of Kayaku Japan Co., Ltd. Adiabatic Reaction Test

A. Thermostat Equipment Testing Facility: Enclosed testing room in Asa Plant of Kayaku Japan Co.,

Ltd.

B. Thermostat equipment: Laboratory Constant Temperature Oven, DK240S by

Yamato Scientific Co., Ltd.

C. Adiabatic container: Dewar flask, D-2000 (2 liter) by Thermo Scientific

D. Other equipment: Video camera, lighting equipment, protection board,

ventilation equipment, etc.

6/15

i. Adiabatic Reaction Test Results (Composition Curves)

１．Start temperature: 100℃、M-Gas atmosphere * Sample: T-5108 bottom liquid (sample on 2012/10/22）

実験方法および条件

Elapsed Time Internal Temp.[℃] AA DAA TAA Polymers PTZ MQ

[min] [hr] Top Bottom [wt%] [wt%] [wt%] [wt%] [ppm] [ppm]

0 0.0 101.4 101.8 96.2 3.6 0.0 0.0 224 1525

115 1.9 104.4 105.0 88.1 11.4 0.3 0.0 217 1531

233 3.9 109.5 110.0 83.3 15.7 0.9 0.0 200 1529

314 5.2 114.4 115.0 78.8 19.8 1.3 0.0 192 1535

372 6.2 119.3 120.0 74.3 23.5 2.0 0.0 186 1544

414 6.9 124.2 125.1 69.9 26.8 3.1 0.0 180 1483

445 7.4 129.3 130.1 66.1 29.7 4.1 0.0 172 1537

※Polymers: molecular weight 5,000～50,000

0

20

40

60

80

100

0 1 2 3 4 5 6 7 8 9 10

Liq

uid

Com

posi

tion [w

t%]

Elapsed Time [hr]

AA

DAA

TAA

2．Start temperature: 100℃、N2 atmosphere * Sample: T-5108 bottom liquid (sample on 2012/10/22）

Elapsed Time Internal Temp. [℃] AA DAA TAA Polymers PTZ MQ

[min] [hr] Top Bottom [wt%] [wt%] [wt%] [wt%] [ppm] [ppm]

0 0.0 96.7 101.8 95.7 4.0 0.0 0.0 218 1763

180 3.0 104.6 105.2 89.1 10.3 0.4 0.0 201 1759

360 6.0 117.0 117.6 78.8 19.4 1.6 0.0 187 1793

420 7.0 123.0 123.6 74.9 22.4 2.5 0.0 174 1790

466 7.8 129.4 130.0 70.5 25.4 3.9 0.0 170 1910

※Polymers: molecular weight 5,000～50,000

0

20

40

60

80

100

0 1 2 3 4 5 6 7 8 9 10

Liq

uid

Com

posi

tion [w

t%]

Elapsed Time [hr]

AA

DAA

TAA

7/15

3．Start temperature: 90℃、N2 atmosphere * Sample: T-5108 bottom liquid (sample on 2012/10/22）

Elapsed Time Internl Temp. [℃] AA DAA TAA Polymers PTZ MQ[min] [hr] Top Bottom [wt%] [wt%] [wt%] [wt%] [ppm] [ppm]

0 0.0 89.8 90.2 96.3 3.5 0.0 0.0 215 1527481 8.0 94.5 95.0 88.4 11.2 0.3 0.0 208 1524800 13.3 99.4 100.0 81.5 17.4 1.0 0.0 193 1522

1020 17.0 104.5 105.0 76.0 22.0 1.8 0.0 180 15181154 19.2 109.5 110.0 69.4 27.1 3.3 0.0 164 15191247 20.8 113.9 115.0 70.2 26.7 3.0 0.0 156 15851319 22.0 118.9 120.0 61.8 32.4 5.5 0.0 148 16091371 22.9 123.8 125.0 58.6 34.4 6.7 0.1 141 15581409 23.5 128.8 130.1 53.7 37.1 8.4 0.6 140 1521

※Polymers: molecular weight 5,000～50,000

0

20

40

60

80

100

0 5 10 15 20 25

Liq

uid

Com

posi

tion [w

t%]

Elapsed Time [hr]

AA

DAA

TAA

Polymers

4．Start temperature: 80℃、N2 atmosphere * Sample: T-5108 bottom liquid (sample on 2012/10/22）

Elapsed Time Internl Temp. [℃] AA DAA TAA Polymers PTZ MQ[min] [hr] Top Bottom [wt%] [wt%] [wt%] [wt%] [ppm] [ppm]

0 0.0 80.5 81.0 96.4 3.4 0.0 0.0 210 15101080 18.0 84.3 84.7 88.4 11.1 0.4 0.0 211 13581520 25.3 86.1 86.6 84.7 14.5 0.7 0.0 201 15191840 30.7 87.8 88.2 82.5 17.3 0.0 0.0 188 15172190 36.5 90.0 90.5 76.8 21.4 1.6 0.0 179 15493070 51.2 99.3 99.9 64.4 30.6 4.8 0.0 154 15943300 55.0 104.4 105.1 58.5 34.5 6.7 0.1 119 16253460 57.7 109.4 110.0 55.9 36.2 7.6 0.2 105 15673570 59.5 114.2 114.8 53.7 36.4 8.8 1.0 109 15523650 60.8 118.8 119.5 47.9 38.9 11.6 1.5 121 15633720 62.0 124.4 125.0 45.6 39.8 12.2 2.3 99 15543770 62.8 129.7 130.3 44.9 38.7 13.4 2.9 92 1565

※Polymers: molecular weight 5,000～50,000

0

20

40

60

80

100

0 10 20 30 40 50 60 70

Liq

uid

Com

posi

tion [w

t%]

Elapsed Time [hr]

AA

DAA

TAA

Polymers

8/15

Behavior of Runaway Reaction

1. Test conditions

A Sample: T-5108 bottom liquid (sampled on October 22, 2012) B. Liquid amount: 1503 g

C. Start temperature: 90℃ D. Atmosphere: N2

# The enclosed testing room cannot be entered for safety reasons when sample temperatures reached above

140℃, hence the oven temperature was not able to raise beyond this point.

2. Test results

A. Comparison for the temperature curves

B. Weight and liquid compositions before and after the test

Analyzed Composition [wt%] Weight

AA DAA TAA Polymers [g]

Before 98.8 3.2 0.0 0.0 1,503

After 20.1 22.2 16.6 41.7 1,439

Weight Decreased: 64g or 4.3% (compare to before test)

C. Conditions of Polymers Adhesion after the Test

Internal of Dewar Flask Exhaust vent Lid of Dewar Flask (silicone stopper)

Floating Polymers (black, white): 31g Condensed Polymers White polymers at the back of the lid

Elapsed Time (hr)

Upper Temperature

Lower Temperature

Oven Temperature

NS Laboratory Data (90～130℃)

9/15

iii. Difference in temperature curves due to start temperature and atmosphere

* Sample: T-5108 bottom liquid (sample on 2012/10/22）

１．Difference in temperature curves due to start temperature and atmosphere

実験方法および条件

※ During the liquid storing operation in V-3138, it took about 60 hours to reach 60 m3 from 25 m3 and about

　 80 hours to discover the white smoke. The experiment data with comparatively long elapsed time (start

　 temperature: 80℃) was used to estimate the tank internal temperature.

70

80

90

100

110

120

130

140

0 10 20 30 40 50 60 70

温度

[℃]

経過時間 [hr]

温度推移実績（開始温度100℃） 組成による演算値（開始温度100℃）

温度推移実績（開始温度90℃） 組成による演算値（開始温度90℃）

温度推移実績（開始温度85℃） 組成による演算値（開始温度85℃）

温度推移実績（開始温度80℃） 組成による演算値（開始温度80℃）

Actual temperature (start temperature: 100℃ )

Actual temperature (start temperature: 90℃ )

Actual temperature (start temperature: 85℃ )

Actual temperature (start temperature: 80℃ )

Calculated compositions (start temperature: 100℃)

Calculated compositions (start temperature: 90℃)

Calculated compositions (start temperature: 85℃)

Calculated compositions (start temperature: 80℃)

Elapsed Time [hr]

Liqu

id T

empe

ratu

re [℃

]

10/15

Appendix 2 Estimated Tank Internal Temperature and Pressure Based on Reaction Analysis

(1) Estimation results

Estimated internal pressure at the time that crack initiated on tank: about 0.274MPaG (structural analysis

result: 0.24 to 0.29MPaG)

(2) Premises for the calculation

-1) The temperature of the upper portion liquid above the coil was assumed to increase and the

composition and temperature were assumed to be evenly distributed.

-2) It was assumed that after the tank liquid volume reached 60m3, DAA formation has caused the

upper portion liquid above the coil to reach an average temperature of 120℃ . After which the

polymerization has caused the temperature to rise.

-3) At the point when the tank internal temperature reached 120℃, the DAA concentration was set at

60wt% and AA concentration at 40wt%.

-4) The amount of heat generated associated with polymerization was estimated based on the rate of

increased temperature which corrected the adiabatic reaction test results. Also, the amount of polymers

formed was calculated by dividing the heat of polymerization into amount of heat generated calculated

from the rate of increased temperature.

-5) The vapor pressure of the liquid inside the tank was presumed to be the vapor pressure of mixed

AA/DAA liquid. The tank internal pressure was calculated from the pressure drop of the generated vapors

which pass through the tank’s vent and sealed gas piping (seal pot).

-6) Confirm the presence of choke flow

Choke flow refers to phenomenon when the gas flow rate released from the vent reached and maintained at

sonic velocity, the pressure at the vent outlet increased. Choke flow will significantly affect the calculation

results and hence we need to confirm its presence.

⇒ The gas velocity at the vent outlet when the tank ruptured was estimated at 248m/sec. On the

other hand, the sonic velocity of acrylic acid gas at temperature of 233℃ was estimated at

250m/sec. We therefore confirmed that it was not within the range of affecting the estimated

internal pressure.

-7) Correction of adiabatic reaction test data

The highest temperature reached during the tests was about 190℃ but the reaction rate (temperature rise)

has slowed down before reaching 190℃. This was presumed to be due to the heat loss during the adiabatic

reaction tests. The time to reach the corrected temperature was calculated by the compositions after tests.

Under no heat loss situation, this time was faster than the time to reach the highest temperature recorded

during the tests.

⇒ The correction range for the rate of temperature increased was calculated right up to the point

that the reaction has slowed down.

11/15

80

100

120

140

160

180

200

220

240

260

280

0 5 10 15 20 25 30

Liqu

id T

empe

ratu

re[℃

]

Elapsed Time [hr]

Test Results

Corrected Transition Temperature

(3) Calculation results

※The estimated liquid temperature was calculated by using Aspen Plus

Time To About 10:00 ⇒ About 13:20 ⇒ About 13:40 ⇒ To About 14:35

Estimated liq. temp. ℃ 120 ⇒ 167 ⇒ 175 ⇒ 233

Estimated press. MPaG - ⇒ 0 ⇒ 0.025 ⇒ 0 .274

Estimated liq. composition （upper section on above coil）

AA wt% 40 ⇒ 35.7 ⇒ 34.2 ⇒ 17.0

DAA, etc. wt% 60 ⇒ 53.8 ⇒ 52.8 ⇒ 46.2

Polymers wt% 0 ⇒ 10.6 ⇒ 13.0 ⇒ 36.8

Polymers Amount kg 0 ⇒ 4,248 ⇒ 5,178 ⇒ 12,814

Evaporation Amount kg 0 ⇒ 59 ⇒ 420 ⇒ 5,522

Vent gas linear velocity m/sec 0 ⇒ 20 ⇒ 86 ⇒ 248

Analysis Starting

Point

Reched Boiling

Point

Level Gauge

Reached Out-

of-Range Limit

Cracks Appeared

on Tank's Shell

Plate

i. Estimated tank internal pressure ※About 14:35, cracks initiated on tank

タンクのき裂発生時点の内部圧力 推定結果：約0.274MPaG　（構造解析結果：0.24～0.29MPaG）

タンク液量が60m3到達後、DAA等の生成によりコイル上の液領域平均温度が120℃に

-6)0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

13:00 13:15 13:30 13:45 14:00 14:15 14:30

Tan

k in

tern

al p

ress

ure

[M

PaG

]

Time （2012/09/29）

Estimated tank internal pressure

Value converted from liquid level gauge

Tank internal

pressure start to

increase

Level gauge out-

of-range limit

Corrected Temperature Curve

12/15

ii. Estimated released gas flow rate ※About 14:35, cracks initiated on tank

0

50

100

150

200

250

13:00 13:15 13:30 13:45 14:00 14:15 14:30

Rele

ased g

as f

low

rate

[kg/m

in.]

Time （2012/09/29）

Tank internalpressure start to

increase

iii. Estimated tank internal temperature ※About 14:35, cracks initiated on tank

タンクき裂

100

120

140

160

180

200

220

240

260

10:00 10:30 11:00 11:30 12:00 12:30 13:00 13:30 14:00 14:30

Tank inte

rnal te

mpera

ture

(℃)

Time （2012/09/29）

Tank internalpressure start

to increase

13/15

Appendix 3 Observation of V-3138 Fracture Surfaces

(1) Extensive observation of the fracture surfaces at the V-3138 fractured sections.

As shown in the figures below, the fracture surfaces are slanted to the through-thickness direction and

formed through shear destruction caused by glide plane decohesion were extensively observed. The fact

that shear fracture was extensively observed did not contradict the presumption of V-3138 destruction were

developing at a fast rate.

Fracture surfaces slanted to the through-thickness direction Destruction due to shear force

Plane stress condition （thin plate, thin pipe, etc.）

Slanted type (shear type) fracture

surfaces easily formed

14/15

(2) Observation of fracture surfaces at cooling water coil outlet nozzle

As shown in the amplified photos, the dark gray areas seen in the outer surface resemble ductile

fractures that are formed from voids linkage when there is destruction from tensile deformation of ductile

metallic materials. Also, the fine veined patterns in the through-thickness direction observed in light gray

areas are assumed to be traces, when the crack formed due to ductile destruction propagated in the

through-thickness direction with the increased load.

Based on the above observations, it is presumed that crack initiated due to ductile destruction on the

exterior of flange outer circumference of cooling water coil outlet nozzle. The crack propagated vertically

on the shell plate escaping from the flange outer circumference.

Veined pattern in through-thickness direction Fracture surfaces due to crack propagation

Ductile fracture surfaces without shear fracture

(dark area) Fracture surfaces formed due to tensile deformation (voids linkage)

Outlet nozzle of cooling water

(observed from outside)

15/15