Upload
others
View
2
Download
0
Embed Size (px)
Citation preview
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
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
4
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