ccps case study bp texas city an aiche technology alliance

CCPSCenter for Chemical Process Safety

An AIChE Technology Alliance Chemical Hazard Engineering Fundamentals – Case Studies

June 28, 2021 Slide - 1

Chemical Hazard Engineering Fundamentals (CHEF)Case Study – BP Texas City


An AIChE Technology Alliance

REFINERY EXPLOSION AND FIRE

Texas City, Texas

March 23, 2005



June 28, 2021 Slide - 2

Chemical Hazards Engineering Fundamentals (CHEF)For the Current Presentation, Please Join Us at:



June 28, 2021 Slide - 3

Case Study – BP Texas CityHazard Identification and Risk Analysis (HIRA) Study

We begin the study by Identifying the Equipment or Activity for which we intend to perform

an analysis. We will often use the operation of a specific equipment item containing a

specific chemical or chemical mixture to define the activity. For example, the operation of a

storage tank, a reactor, a piping network, etc. Inputs are chemical data, equipment designinformation, operating conditions, and plant layout.

What are the Hazards?

What can go Wrong?

How Bad could it Be?

How Oftenmight it

Happen?

Is the Risk Tolerable?

Identify Chemical

and ProcessHazards

Estimate Frequency

AnalyzeConsequences

Analyze Risk

ImplementAdditional

Safeguards as Needed

Develop Scenarios

Select Equipmentor Activity to be

Analyzed

Sustain Performancefor Life Cycle

of Facility



June 28, 2021 Slide - 4

Case Study – BP Texas City

Process Description

The ISOM unit provides higher octane components for unleaded gasoline, consists of four

sections: an Ultrafiner14 desulfurizer, a Penex15 reactor, a vapor recovery / liquid recycle

unit, and a raffinate splitter. At the BP Texas City refinery, the ISOM unit converted straight-

chain normal pentane and hexane into branched-chain isopentane and isohexane for

gasoline blending and chemical feedstocks.

We will start with the raffinate splitter section where a hydrocarbon mixture is separated into

light and heavy components. About 40 percent of the raffinate feed was recovered as light

raffinate (primarily pentane/hexane). The remaining raffinate feed was recovered as heavy

raffinate.. The raffinate splitter section could process up to 45,000 barrels per day

(approximately 1300 gallons/minute) of raffinate feed.

This is an illustrative example and does not reflect a thorough or complete study.



June 28, 2021 Slide - 5


Process DescriptionThe process equipment in

the raffinate splitter section

consisted of a feed surge

drum; a distillation tower; a

furnace with two heating

sections, one used as a

reboiler for heating the

bottoms of the tower and the

other preheating the feed;

air-cooled fin fan condensers

and an overhead reflux

drum; various pumps; and

heat exchangers. Figure 1: Raffinate Splitter Tower System of the ISOM Unit



June 28, 2021 Slide - 6

Liquid raffinate feed was pumped into the raffinate splitter tower near the tower’s midpoint.

An automatic flow control valve adjusted the feed rate. The feed was pre-heated by a heat

exchanger using heavy raffinate product and again in the preheat section of the reboiler

furnace, which used refinery fuel gas. Heavy raffinate was pumped from the bottom of the

raffinate splitter tower and circulated through the reboiler furnace, where it was heated and

then returned below the bottom tray. Heavy raffinate product was also taken off as a side

stream at the discharge of the circulation pump and sent to storage. The flow of this side

stream was controlled by a level control.

Light raffinate vapors flows overhead, is condensed by air-cooled fin fan condensers, and

then deposited into a reflux drum. Liquid from the reflux drum, was then pumped back into

the raffinate splitter tower above the top tray.


Process Description



June 28, 2021 Slide - 7


What can go Wrong?


How Oftenmight it

Happen?


Identify Chemical

and ProcessHazards

Estimate Frequency

AnalyzeConsequences

Analyze Risk

ImplementAdditional


Develop Scenarios


Analyzed


of Facility

The scope of this presentation is focused on the Raffinate Column. As this evaluation is

related to an incident investigation, little emphasis will be placed on Frequency Evaluation

(the incident has already occurred) or Risk Analysis (evaluation of protection layers).

However, a “worst case” consequence such as might be evaluated during risk analysis will be addressed.




June 28, 2021 Slide - 8


Raffinate Composition

Typical Raffinate composition per Refinery Explosion

and Fire, CSB Report No. 2005-04-I-TX page 259

For entry into CHEF, the mixture is

simplified to:

0.06 n-pentane (including isopentane)

0.45 n-hexane

0.31 n-heptane

0.18 n-octane



June 28, 2021 Slide - 9


Chemical DataA composition (weight fraction):

0.06 n-pentane

0.45 n-hexane

0.31 n-heptane

0.18 n-octane

was used as representative.

The operating pressure was 25

psig (172 kPa gauge) and the

operating temperature of 112 C

was estimated as the saturation

temperature or temperature

such that the estimated vapor

pressure matches the operating

pressure for a boiling liquid.

CHEMICAL HAZARDS AND CHEMICAL MIXTURE INPUT INFORMATION

Inputs for one or more chemical components must be entered in shaded "yellow" fields if Table Data Value is not used

Process Inputs: Input Value Input Units

Temperature, T 112 112 C

Physical State of Contents Liquid Assumed liquid if blank

Estimated Vapor Pressure at Specified Temperature: 172.149 kPa gauge

Chemical Inputs: Table Name Wt FractionSecond Liq

Phase

Mol Wt Data

Table Value

Mol Wt Input

Value

Mol Wt for

EquationHealth Flammability Stability

0.06 72.15 72.15 2 4

0.45 86.2 86.2 2 3

0.31 100.2 100.2 1 3

0.18 114.23 114.23 2 3

1

Input Name

NFPA Hazard Ratings (Table Values)

Note that Weight Fraction, Molecular Weight and Vapor Pressure data, in addition

to physical State of Contents, must be entered to estimate Vapor Composition

Pentane

Hexane

Heptane

Octane

Clear Inputs

Temperature is adjusted

until the estimated vapor

pressure matches the

operating pressure.

Up to 5 chemicals with the

associated weight fraction may

be added to create a mixture.

Mixtures are assumed “ideal”.

The chemical property inputs for the various worksheets will then

be estimated from this composition at the appropriate temperature



June 28, 2021 Slide - 10


Chemical Hazards

Mol Fract Vapor

/ LFL

0.15581

0.53656

0.14130 -48.15 C

0.03813

S yi / LFLi = 0.87180

LFLMixture = 1 / (S yi / LFLi) = 1.15 Vol %

ERPG-2 (t /60)1/nMol Fract Vapor /

ERPG-2 (t /60)1/n ERPG-3 (t /60)1/n

Mol Fract Vapor /

ERPG-3 (t /60)1/n

32500.0 0.00001 190000.0 0.00000

2900.0 0.00020 8600.0 0.00007

800.0 0.00019 4900.0 0.00003

385.0 0.00009 5000.0 0.00001

S yi / ERPG-2'i = 0.00050 S yi / ERPG-3'i = 0.00011

ERPG-2'Mixture = 1 / (S yi / ERPG-2'i) = 2005.61 ppmv ERPG-3'Mixture = 9196.07 ppmv

ESTIMATED MIXTURE ERPG-2 and ERPG-3

Dose adjusted ERPG per equation 3-1 and mixture ERPG per equation 3-3

Chemical Name

The Minimum Flash Point for any

Component is:

ESTIMATED MIXTURE LFL and MINIMUM FLASH POINT

equation 2-1

Hexane

Heptane

Octane

Chemical Name

Pentane

Hexane

Heptane

Octane

Pentane

Note: n is assumed 2 for interpolation of exposure duration less than one hour and 1 for

extrapolation of exposure duration to greater than one hour if not entered

Based on Exposure Duration of 60 min

✓ An estimated mixture flashpoint of -48 C

(the minimum of any component) and

estimated lower flammable limit of 1.15

volume % indicates a potentially high

flammability hazard.

✓ An estimated ERPG-2 of >2000 ppm would

indicate a low to moderate toxicity hazard.

The NFPA Health Ratings for these

materials range from 1 to 2 also indicating a

low to moderate hazard.

✓ There may be other hazards to consider

such as thermal due to a maximum process

temperature greater than 60 to 80 C.



June 28, 2021 Slide - 11


What can go Wrong?


How Oftenmight it

Happen?


Identify Chemical

and ProcessHazards

Estimate Frequency

AnalyzeConsequences

Analyze Risk

ImplementAdditional


Develop Scenarios


Analyzed


of Facility




June 28, 2021 Slide - 12

Drain or Vent Valve

OpenFlow-Loss of Containment All

Procedure Failure (Human Error)

Flow Control FailureDrain or Vent Leak

Flammable Release

Toxic Release

Chemical Exposure

Excessive Heat Input -

Heat Transfer

Temperature-High

Pressure-High

Heat Input-High

All with Heat Transfer Capability

Temperature Control Failure

Residence Time Failure (Solids)

Feed rate Control Failure

Relief Venting

Equipment Rupture

Equipment Damage

Flammable Release

Toxic Release

Flash Fire or Fireball

Physical Explosion

Business Loss

Excessive Heat Input -

Pool Fire Exposure

Temperature-High

Pressure-High

Heat Input-High

AllScenarios involving spill plus ignition in nearby

liquid-containing equipment

Relief Venting

Equipment Rupture

Equipment Damage

Flammable Release

Toxic Release

Flash Fire or Fireball

Physical Explosion

Business Loss

Overfill, Overflow, or

Backflow

Level-High

Flow-BackflowAll Liquid Containing Equipment

Level Control Failure

Procedure Failure (Human Error)

Overflow Release

Equipment Damage

Equipment Rupture

Flammable Release

Toxic Release

Physical Explosion

Business Loss

Overflow - Flooding or

PluggingLevel-High

Adsorber/Scrubber

Distillation

Vapor Quench

Level Control Failure

Pressure Control Failure

Flow Control Failure

Overflow Release

Equipment Damage

Equipment Rupture

Flammable Release

Toxic Release

Physical Explosion

Business Loss

Relief Device Failure Flow-Loss of Containment All Mechanical Failure Relief Hole Size Leak

Flammable Release

Toxic Release

Chemical Exposure

Vacuum Damage Pressure-Low AllPressure Control Failure

Mechanical Failure

Full-Bore Leak

Equipment Damage

Flammable Release

Toxic Release

Business Loss

Case Study – BP Texas CityPartial List of Hazard Scenarios per CHEF Example Scenarios



June 28, 2021 Slide - 13


Suggested Scenarios for Raffinate Column

WORKING WITH YOUR EVALUATION TEAM:

❑ Review the suggested list of scenarios. Do these represent what you

would expect for a distillation column?

❑ Are there scenarios that have been “screened out” (shown in gray) that

should be considered?

❑ Are there scenarios missing? (Possibly similar scenarios with different

Initiating Events)

❑ Do you agree with the “worst” Consequence (Tolerable Frequency

Factor) for the scenario listed?



June 28, 2021 Slide - 14


Suggested Scenarios for Raffinate Column

WORKING WITH YOUR EVALUATION TEAM:

❑ Utilize an Appropriate Hazard Evaluation Technique (HAZOP, What If, etc.)

to capture additional scenarios.

❑ Capture existing Safeguards and Recommendations for each Scenario.

Note the Dates and Names of participants in the Study.

❑ Select which Scenarios warrant more detailed Risk Evaluation (such as

Layers of Protection Analysis).



June 28, 2021 Slide - 15


What can go Wrong?


How Oftenmight it

Happen?


Identify Chemical

and ProcessHazards

Estimate Frequency

AnalyzeConsequences

Analyze Risk

ImplementAdditional


Develop Scenarios


Analyzed


of Facility




June 28, 2021 Slide - 16


Site Layout

Approximately 500 m

Several wooden trailers are located

approximately 200 m from the

raffinate splitter housing 20 people.

The trailers are “low strength”

construction. In addition, the process

area appears to be relatively “low”

equipment congestion.

The blowdown tank which receives

the discharge from the raffinate

splitter relief devices is located 60 m

from the wooden trailers and vents at an elevation of 36 m.



Estimated Pool Temperature = 74.9 C ( 348.05 K )

Vapor Pressure at Pool Temp, Psat

= 101.300 kPa

Assumed Wind Speed ( Location ) = 3 m/sec

(may be conservative as does not account for evaporative cooling)

Evaporation Rate (equation 11-27) , mP = 0.0021 Mw2/3

u0.78

Psat

/ TP

0.0021 ( 86 )^2/3 ( 3 )^0.78 ( 101.3 ) / ( 273.15 + 74.9 ) = 0.028096 kg/sec m2

Release Duration, tL = Inventory / L = / 55 = 3600 sec

Estimated Liquid Rate to Pool, L' = L ( 1 - Fv ) ( 1 - FD ) =

55 ( 1 - 0.334 ) ( 1 - 0.4409 ) = 20.4798 kg/sec

AP = 100 M [FV+(1-FV) FD] / rL

Estimated Pool Area (equation 11-24) , AP = L' / [ rL / ( 100 tL ) + mP / 2 ] =

20.4798 / [ 577 / { 100 ( 3600 ) } + 0.028096 / 2 ] = 1310 m2

Pool Evaporation = mP AP = ( 0.028096 ) ( 1310 ) = 20.4798 kg/sec

limited to pool fill rate

Liquid Release - Airborne Quantity = L [ Fv + FD ( 1 - Fv ) ] + mP AP =

55 [ 0.334 + 0.4409 ( 1 - 0.334 ) ] + 20.4798 = 55 kg/sec - or - 7260 lb/min

Liquid Release - Airborne Quantity for Equipment Rupture

If Initial Vapor / Total Vapor or 0 kg / 198000 kg < 0.01, Use Maximum Pool Evaloration Rate

or kg/sec - or - lb/min

1 hour Airborne Quantity = 198000 kg - or - 435600 lb

based on pool evaporation for 1 hour

LIQUID RELEASE - ESTIMATED AIRBORNE QUANTITY

LIQUID RELEASE - ESTIMATED POOL EVAPORATION

Flash Fraction (equation 11-19) , Fv = CS ( T-TB ) / DHV = ( 2.67 ) ( 112 - 74.9 ) / ( 297 )

Fv = 0.334 Initial Fraction Vapor = 0.334

Selected Liquid Release Rate, L = 55 kg/sec

Aerosol Temperature, T' = 74.9 C ( 348.05 K )

Vapor Pressure at Aerosol Temp, Psat

= 101.300 kPa

Flashing Fluid Density, r ' = 1 / [ Fv / r V + ( 1 - Fv ) / r L ] =

577 kg/m3

Liquid or Two-Phase Release Velocity (equation 11-5), vd = 1.27 L / [ d2 cd r' ] =

or vd = 1.524 [ 1000 (P0 - PA) / r' + 9.8 h' ]0.5

=

1.27 ( 55 ) / [ ( 0.087 )^2 ( 0.61 ) ( 577 ) ] = 26.3 m/sec

Estimated Droplet Diameter (equation 11-21), dd = 170 (1 - FV) / vd2 =

170 ( 1 - 0.334 ) / 26.3^2 = 0.1632 mm

Liquid Spray Distance (equation 11-4 ), s = vd t = 0.45 vd h1/2

=

0.45 ( 26.3406529539294 ) ( 1 )^0.5 = 11.9 m - or - 38.9 ft

Aerosol Fraction (equation 11-22) , FD = 0.043 vd2 Mw

2/3 P

sat h

1/2 / [ r

L T' ( 1 - Fv ) ] =

0.043 ( 26.34 )^2 ( 86 )^2/3 ( 101.3 ) ( 1 )^1/2 / [ 577 ( 273.15 + 74.9 ) ( 1 - 0.334 ) ]

FD = 0.4409

LIQUID RELEASE - ESTIMATED FLASH FRACTION

LIQUID RELEASE - ESTIMATED AEROSOL FRACTION

SOURCE MODELS INPUT INFORMATION

Required Inputs are Shaded "Yellow"

STEP 1 - Select Type of Release

"

STEP 2 - Enter Required Release Information

Release Inputs: Input Value Input Units Equation Input Equation Units

Hole Size, d 0.0867995 m

Coefficient cd (typically Square Edged Hole) 0.61 dimensionless

Heat Input Rate, q kWatt

Specified Rate (at either specified Hole 55 55 kg/sec Size or Release Pressure) Mismatch in Hole Size and Pressure Inputs, Equivalent Release Pressure = 172.369 kPa

Process Inputs: Input Value Input Units Equation Input Equation Units

Release Temperature, T 112 112 C

Release Pressure (gauge), P0 - PA 25 psi 172.36892 kPa gauge

Total Inventory (Leave Blank for 1 hour leak) kg

STEP 3 - Enter Chemical Properties (or Select Chemical Name from Pic List)

Cas No.

Chemical Name

(Leave Input Value Blank to accept Chemical Data Table Values)

Data Table Value Input Value Input Units Equation Input Equation Units

Physical State Liquid Liquid

Molecular Weight, Mw 86 86

Normal Boiling Point, TB 74.9 74.9 C

Liquid Properties at Release Temperature :

Vapor Pressure, Psat 273 273 kPa absolute

Liquid Density, rL 577 577 kg/cu m

Liquid Heat Capacity, CS 2.67 2.67 kJ/kg C

Heat of Vaporization, DHV 297 297 kJ/kg

STEP 4 - Enter Equipment and Plant Layout Information

Input Value Input Units Equation Input Equation Units

Equipment Location (Assumed Outdoor if Blank)

Equipment Volume, VEquip cu m

Estimated Contained Mass (for Reference if volume and density entered) kg

Liquid Height within Equipment, h' m

Diked Area (Leave Blank if No Dike) No Dike sq m

Leak Elevation above Surface, h 1 m

Input Units may be changed - Input Values in "blue" will be converted to appropriate equation units

Specified Rate

Square Edged Hole

Clear InputsClear Inputs

June 28, 2021 Slide - 17

Use a Specified Release rate equal to

feed rate of 55 kg/sec for Overfill at

the Release Temperature of 112 C

Chemical Properties at

Column Release

Temperature of 112 C

Leave Elevation Blank for now as Discharge was to a

Blowdown Tank which then Overflowed to the AtmosphereTotal Airborne Rate of 55 kg/sec

with a liquid pool of 1310 m2

Case Study – BP Texas CityConsequence Analysis – Source Models



per Correlation Details with Concentration in ppmv

Plume or Continuous: C m = 8.1E+08 1.76 )

Puff or Instantaneous: C m = )

per Correlation Details for 1.5 m/sec F Stability with Concentration in ppmv

Plume or Continuous: C m = 3.5E+08 1.74 )

Puff or Instantaneous: C m = )

Test for Plume versus Puff Model at Dispersion Conditions: (equation 12-20)

If Q > (a*/a) u Q* Xb-b* ---> Instantaneous Model (from equating Puff and Plume correlation at X)

Alternate weather Puff versus Plume Model

Estimated Exposure Duration - Continuous Dispersion (Equations 12-21, 12-23)

t = Q* / Q for Continuous or 2 (0.2) [Q* T amb / (Mw C)]0.37

/ Wind Speed for Instantaneous

t = 3600 = 3600.0 sec

Alternate weather Estimated Exposure Duration

t = 3600 = 3600.0 sec

Maximum Downwind Distance to Concentration of Interest (equations 12-9, 12-13)

Continuous (equation 37): X = a [ Q / ( u Mw Cm ) ]b - X0

r air = 1.18 Kg/m3 = 113800 [ ( 55 ) / { ( 3 ) 85 ( 11600 ) } ]^0.57 - 0 = 235 m

at ambient conditions Instantaneous (equation 41): X = a* [ (Q* / ( Mw Cm ) ]

b* - X0

m

( Q* / Mw ) / ( X ^

Continuous

for 3 m/sec Wind Speed, Class D Atmospheric Stability, and Residential Surface Roughness

( Q* / Mw ) / ( X ^

SIMPLE VAPOR DISPERSION

( Q / Mw ) / ( u X ^

Continuous

( Q / Mw ) / ( u X ^

Maximum Concentration at Ground Elevation

Effective Release Elevation

Release Elevation

Ground Elevation

Cloud Centerline

Concentration at Specified Distance

and Elevation

Maximum Concentration at Specified Distance

Outdoor Vapor Release

Initial Dilution

Distance Correction for

Initial Dilution, DXt

Show Correlation Details

Hide Correlation Details

VAPOR DISPERSION INPUT INFORMATION


STEP 1 - Select Location, Type of Release, Concentration and Distance of Interest

Release Location (Assumed Outdoor if Blank)

Type of Release

Use Averaging Time Correction for Flammable Release Yes


Concentration of Interest 1.16 vol % 11600 ppm

Concentration of Interest for Hazard Analysis is typically 1/2 LFL, LFL, ER-2, ER-3 or LC 50

Outdoor Downwind Distance of Interest, X m

Distance of Interest is typically to the property limit, to an unrestricted work area, or to an occupied building

Note that Concentration of Interest must be entered for estimation of Instantaneous (or puff) release


Cas No.

Chemical Name

Data Table Value

Lower Flammable Limit, LFL vol %

ERPG-3 Concentration ppm


Input Value Equation Input

Vapor Molecular Weight, Mw 85 85

Normal Boiling Point, TB C

STEP 3 - Enter Process Information

Process Inputs: Input Value Input Units Equation Input Equation Units

Airborne Rate, Q 55 55 kg/sec

Vapor Temperature, T 74.9 74.9 C

Total Release Quantity, Q* (Leave Blank if Unlimited) kg

Liquid or Two-Phase Release Velocity* (Flashing Liquid) m/sec

Initial Fraction Vapor*, f 0 (Flashing Liquid Only) Fraction

kPa absolute

*Note that these inputs may not significantly change results

STEP 4 - Enter Equipment and Plant Layout Information

Equipment and Plant Layout Inputs: Input Value Input Units Equation Input Equation Units

Diameter of Hole or Discharge Piping, d0 m

The hole size for vapor release estimate or diameter of relief system discharge piping. If blank, no jet mixing used.

Release Elevation, h (Blank assumed at Ground) 36 36 m

Release Direction (Assumed Horizontal if Blank)

Enclosed Process Area Volume, VBldg cu m

Enclosed Process Area Ventilation Rate Air Changes/Hr (Assumed 1 if Blank)

Vertically Upward

Vapor Pressure at Release Temp

(Subcooled Liquid Only)

Outdoor

Flashing Liquid


If "Yes" dispersion concentration

is approximately doubled

For reference in determining

concentration of interest

Leave Blank to accept

Chemical Data Table

Mw


June 28, 2021 Slide - 18

Use a Flashing Liquid Release

with Averaging Time Correction

for Flammable Evaluation


Blowdown Tank Release

Temperature of 74.9 C (estimated normal boiling point)

Elevation of Blowdown stack is 36 but release will

be assumes ground elevation for a flashing liquid

Distance to Concentration of

Interest (lower flammable

limit) is 235 m at the default

wind speed of 3 m/sec and

Class D Stability

Vapor Rate from Source

Model of 55 kg/sec. Leave

two-phase velocity and initial

fraction vapor blank for now

Case Study – BP Texas CityConsequence Analysis – Vapor Dispersion



June 28, 2021Slide - 19

Release

Location

Wind

Direction

Explosion

Distance

XLFL

Vapor Cloud

Explosion

Epicenter

The entire vapor cloud is considered

a single Potential Explosion Site with

epicenter at the center of the

flammable cloud (0.5 XLFL).

An single overall level of congestion

and confinement for the entire cloud

is used.

Wind direction is assumed toward

greatest population or building with

highest occupancy.

Vapor Cloud ExplosionSimple Modeling Approach within CHEF and RAST

Methodology is described in the CHEF Guide



June 28, 2021 Slide - 20

Site LayoutCongestion or Obstacle Density Categories

Low

High

Medium

Low – Only 1-2 layers of obstacles.

One can easily walk through the area

relatively unimpeded.

Medium – 2-4 layers of obstacles.

One can walk through an area, but it

is cumbersome to do so. Medium

Congestion is assumed in RAST if a

category is not entered by the user.

High – Many layers of repeated

obstacles. One could not possibly

walk through the area and little light

penetrates the congestion .

The method described in

CHEF is limited to

consideration of the entire

cloud volume as a single

Potential Explosion Site

(PES) at an overall or

average category of

process equipment

congestion. This technique

does not account for small

localized areas of higher

congestion where blast

overpressure will be higher.



EXPLOSION INPUT INFORMATION


STEP 1 - Select Type of Explosion and Distance of Interest

Type of Explosion:


Distance of Interest, X 60 60 m

STEP 2 - Enter Equipment Burst Pressure and Volume for Physical Explosion Skip Step

Physical Explosion Inputs: Input Value Input Units Equation Input

Burst Pressure (gauge), PB - P0 kPa gauge

Equipment Volume, VEquip cu m

Burst Temperature, TBurst C

Fraction Liquid Level (if Superheated), FF

Flash Fract during Depressurization, FV

STEP 3 - Enter Quantity and Heat of Reaction for Condensed Phase Explosion Skip Step

Condensed Phase Detonable Inputs: Input Value Input Units Equation Input

Mass of Material, M kg

Heat of Reaction per Mass, DHR kJ/kg


Cas No.

Chemical Name

Data Table Value User Value Equation Input

Vapor Molecular Weight, Mw 85 85

Liquid Density, rL (at Burst Temperature) kg/m3

Lower Flammable Limit, LFL 1.16 1.16 vol %

Fuel Reactivity based on Fundamental Burning Velocity Medium(Leave User Value Blank to accept Data Table Value)

Ideal Gas Vapor Density, rV (at Burst Temperature) 3.78 kg/m3

STEP 5 - Enter Information for Building or Head Space Explosion Skip Step

Building or Head Space Explosion Inputs: Input Value Input Units Equation Input

Building or Head Space Volume, VB cu m

Degree of Internal Congestion Assumed "Medium" if Blank

1 D Confinement? (such as Fire Tube Boilier) Assumed "No" if Blank

STEP 6 - Enter Information for Vapor Cloud Explosion

Vapor Cloud Explosion Inputs: Input Value Input Units Equation Input

Distance to LFL from Dispersion Model , XLFL 235 235 m

Vapor Rate 55 55 kg/sec

Wind Speed (Assumed 3 m/sec if Blank) 3

Degree of Outdoor Congestion Low Assumed "Medium" if Blank

Vapor Cloud Explosion


Clear Inputs

Vapor Cloud Explosion (based on 3 m/sec wind speed)

Low Medium High

High 0.5 >1 >1

Low-Medium 0.35 0.5 1

Flammable Cloud Volume (equation 13-4), VC = 2440 Q XLFL / ( F u Mw CLFL) =

2440 ( 55 ) ( 235 ) / [ 2 ( 3 ) ( 86 ) ( 1.16 ) ] = 30000 m3

Distance to Explosion Epicenter, X EE = 0.5 X LFL = 0.5 ( 235 ) = 117.5 m

Potential Explosion Site Volume limited to 30000 cu m

Explosion Energy (equation 13-3), QE = 3500 VPES = Note: P A = 101.3 kPa

3500 ( 30000 ) = 105000000 kJ

Scaled Overpressure at 1 psi = 0.068

Scaled Distance, R = X / (2 QE / P0 )1/3

= 1.5

Distance to 1 psi = R (2 QE / P0 )1/3

+ XEE =

1.5 [ 2 ( 105000000 ) / 101.3 ]^1/3 + 117.5 = 307.8 m

From Graph, Scaled Distance, R = ( X - XEE) / (2 QE / P0 )1/3

=

( 50 - 117.5 ) / [ 2 ( 105000000 ) / 101.3 ]^1/3 = 0.01

Scaled Overpressure = 0.22

Overpressure at 50 m = 3.2 psi or 22.3 kPa

Distance of Interest is less than the distance from leak source to explosion epicenter

Fuel

Reactivity

Obstacle Density or Congestion

Use Mach 0.35 for Low-

Medium Fuel Reactivity and

Low Confinement

BAKER-STREHLOW-TANG MODEL

0.01

0.1

1

10

100

0.1 1 10

Sca

led

Ove

rpre

ssu

re,

PS

= P

0/

P a

Scaled Distance, R = X / (2 QE / Pa)1/3

Baker-Strehlow-Tang Overpressure Curves

Mach > 1

Mach 1.0

Mach 0.7

Mach 0.35

Mach 0.5

1 psi overpressure

June 28, 2021 Slide - 21

Case Study – BP Texas CityConsequence Analysis – Explosions

Use Vapor Cloud Explosion and

Distance of Interest as 60 m from

Release Point to Project Trailers


Blowdown Tank Release

Temperature of 74.9 C (estimated normal boiling point)

Estimated Distance to 1 psi

Blast Overpressure is 308 m

and Estimated Overpressure

at 50 m Distance of Interest

is 3.2 psi

Enter Distance to LFL from

Vapor Dispersion estimate,

vapor rate and “low” degree

of congestion



A simplification in RAST is wind

direction toward the highest population.

This is quite reasonable in Risk Analysis

where the wind direction is unknown.

In the actual incident, the wind direction

was toward the southeast rather than

west toward the wooden trailers.

Wind Direction represents a key

difference between estimates for Risk

Analysis versus Incident Investigation.

Blast overpressure at the wooden

trailers would likely have been higher if

the wind direction was toward the

trailers.June 28, 2021 Slide - 22


Consequence Analysis

CHEF estimated

explosion epicenter as

the center of LFL cloud

REPORT NO. 2005-04-I-TX , US Chemical Safety Board,

Figure H-2 Blast Overpressure Map

Release Point –

Blowdown Stack

CHEF estimated maximum 308

m 1 psi overpressure distance

from release point assuming

“low” equipment congestion..

CHEF estimated 235 m

distance to LFL concentration

using default 3 m/sec wind

speed and class D

atmospheric stability



June 28, 2021 Slide - 23


Consequence Analysis – Impact Estimation

Flash Fire

Distance to 1/2 LFL of 349 m Exceeds 3 m Threshold for Severe Impact.

Impact Area (Equation 14-1), = 0.3 XMultiple LFL2

= 0.3 ( 349 )^2 = 36540.3 m2

N = ( 36540.3 sq m ) 0.0002 people/sq m + 1 = 8.308

Vapor Cloud Explosion (VCE)

Total Airborne Release of Unlimited Kg Exceeds 100 Kg Thresold for Vapor Cloud Explosion

Overpressure at Nearest Wall of 3.2 psi Exceeds 1.3 psi Thresold for Building Damage

Impact to Occupied Building (Figure 14.11), Vulnerability = 0.918

N = 0.918 ( 20 ) = 18.36

Total Personnel Severely Impacted = Flash Fire + Occupants Impacted = 26.67

OUTDOOR FLAMMABLE IMPACT

Flash Fire Personnel Seriously Impacted, N = Impact Area times

Population Density plus Personnel in Immediate Area

Number of Building Occupants Severely Impacted = Vulnerability times

Total Building Occupants =

ONSITE IMPACT ASSESSMENT INPUT INFORMATIONRequired Inputs are Shaded "Yellow"

STEP 1 - Select Location and Scenario Outcome

Release Location (Assumed Outdoor if Blank)

Scenario Outcome

Equation Input

Number of Personnel Routinely in the Immediate Vicinity? 1.00


Cas No.

Chemical Name

(Leave Input Value Blank to accept Chemical Data Table Values)

Data Table Value Input Value Equation Input

Lower Flammable Limit, LFL 1.16 1.16 vol %

Fuel Reactivity based on Fundamental Burning Velocity Medium



LC1 as Multiple of EPPG-3 Concentration 2 2 typically 1 to 2

LC50 as Multiple of EPPG-3 Concentration 5 5 typically 3 to 5

Dose Exponent 2 2 typically 1 to 2

Fraction Indoor/Outdoor Concentration 0.5 0.5 typically 0.2 to 0.5

Dermal Toxicity Risk Phrase

Does Process Temperature Represent a Thermal Hazard?


Release or Exposure Duration (leave blank if not limited) 1.0000 hr

Total Airborne Quantity, AQTotal kg

Vapor Cloud Explosion

Outdoor

Personnel in the Immediate Vicinity include those associated with procedures requiring operator attendance such as unloading a tank truck,

sampling, in addition to personnel using nearby walkways, etc.

(Assumed 1 if

Blank)


STEP 3 - Enter Plant Layout Information


Distance to Property Limit or Offsite Population (Fence Line) m

Fraction of Offsite Population Outdoors 0.1 (Assumed 0.1 if Blank)

Fraction of time for Night Weather 0.2 (Assumed 0.2 if Blank)

Offsite Population Density 0.0015

Occupied Building Type

Maximum Number of Occupants in Occupied Building 20

Onsite Outdoor Population Density 0.0002

Enclosed Process Area Volume, VBldg cu m

Maximum Number of Personnel within Enclosed Process Area 2(Assumed 2 if Blank)

STEP 4 - Enter Hazard Distances


Concentration at Distance to Fence Line or Public ppm

Concentration at Distance to Occupied Building ppm

Explosion Overpressure at Distance to Occupied Building psi

Explosion Overpressure at Center of Occupied Building 3.2 3.2 psi

Spray Distance m

Distance to Multiple of Lower Flammable Limit 349 349 m

Distance to LC50 Concentration - Daytime Conditions m

Distance to LC50 Concentration - Nighttime Conditions m

Distance to LC1 Concentration - Daytime Conditions m

Distance to LC1 Concentration - Nighttime Conditions m

Distance to Severe Blast Impact m

Maximum Fragment Range m

Number of Fragments (typically 3 to 10)

Concentration within Enclosed Process Area ppm

Outdoor Population Density should account for maintenance and other personnel who may occasionally be in nearby outdoor

process areas.

(Assumed 0.0015

people/m2 if Blank)

Low Strength (Assumed Typical

Construction if Blank)

(Assumed 0.0002

people/m2 if Blank)

Use Outdoor Release Location

and Vapor Cloud ExplosionEnter Low Strength

Building with 20 Occupants

Enter Blast Overpressure at

Distance to Center of Occupied

Building from Explosion Worksheet

Enter Distance to Fraction of LFL

concentration (0.5 in this example) for Flash

Fire Impact from Vapor Dispersion Worksheet

Estimate of 18 fatalities within Occupied

Building and addition 6 to 7 outdoors from

Flash Fire at default population density



June 28, 2021 Slide - 24


What can go Wrong?


How Oftenmight it

Happen?


Identify Chemical

and ProcessHazards

Estimate Frequency

AnalyzeConsequences

Analyze Risk

ImplementAdditional


Develop Scenarios


Analyzed


of Facility




Risk Analysis Screening Tools (RAST)

Risk Matrix

June 28, 2021 Slide - 25

To understand the Consequence

Severity and Tolerable Frequency, the

values for key Study Parameters and a

Risk Matrix may be viewed on the

Workbook Notes worksheet. These

values may be updated on hidden

worksheets and should reflect the

company’s specific risk criteria.

For this case study, the Risk Matrix

(right) has been used. The Human

Harm criteria is based on an estimated

number of people severely impacted

(severe injury including fatality).

2 3 4 5 6 7

Description Human Harm Environment Business Loss 10^-2/year 10^-3/year 10^-4/year 10^-5/year 10^-6/year 10^-7/year

Reportable Incident to Environmental Agency OR

< 10 kg Very Toxic to Waterway OR < 100 kg NFPA-H4 to Soil

< 100 kg Toxic to Waterway OR < 1000 kg NFPA-H3 to Soil

< 1000 kg Harmful to Waterway OR < 10000 kg NFPA-H2 to Soil

Environmental Contamination Confined to Site OR




Environmental Contamination of Local Groundwater OR




Incident Requiring Significant Off-Site Remediation OR



> 100000 kg Harmful to Waterway OR > 100000 kg NFPA-H2 to Soil

Incident with Significant National Media Attention OR


> 100000 kg Toxic to Waterway OR > 1000000 kg NFPA-H3 to Soil

Acceptable

Tolerable - Offsite

Tolerable - Onsite

Unacceptable

Low

Con

sequ

ence

Hig

h

Con

sequ

ence

Low

Frequency

High

Frequency

Consequence Severity Description Frequency

Severity Level-1

Minor Injury On-site

(or < 0.01 Person Severely Impacted On-site)

Potential for Adverse Local Publicity

Property Damage and

Business Loss < $50M2 Orange Yellow

Green

Green

Yellow> 10 People Severely Impacted On-site

> 1 Person Severely Impacted Off-site

Property Damage and

Business Loss > $50 MM6 Red

Red Red Orange Yellow GreenSeverity Level-41 to 10 People Severely Impacted On-site

0.1 to 1 People Severely Impacted Off-site

Property Damage and

Business Loss $5 MM to

$50 MM

Legend

6

Yellow Green GreenSeverity Level-2

Major Injury On-site

(or 0.01 to 0.1 Person Severely Impacted On-site)

Public Required to Shelter Indoors

(or Minor Injury Off-site)

Property Damage and

Business Loss $50 M to

$500 M

3 Red

Red Orange Yellow GreenSeverity Level-3

Potential Fatality On-site

(or 0.1 to 1 Person Severely Impacted On-site)

or Potential Major Injury Off-site

Property Damage and

Business Loss $5 MM to

$50 MM

4 Red

Severity Level-5

6

Red Orange

5 Red

Risk Matrix: Risk = Consequence Severity times Frequency

Red Red

Green Green Green Green

Orange



June 28, 2021 Slide - 26


Risk Analysis / Layers of Protection Analysis (LOPA)

Initiating Event – either Basic Level Control System failure (instrument was not

indicating the correct level) or Human Performance Failure as “operators routinely

deviated from written procedures maintaining a high liquid level during startup to

minimize the potential for costly damage to the fired heaters”. From the CCPS Book,

Layers of Protection Analysis, a frequency of 0.1 per year may be appropriate.

Loss Event – Overfill of the Raffinate Splitter Column to the Blowdown Tank with

subsequent release of flammable material to the atmosphere.

Incident Outcome and Consequence – Vapor Cloud Explosion resulting in greater

than 10 potential fatalities.



June 28, 2021 Slide - 27


Risk Analysis / Layers of Protection Analysis (LOPA)Conditional Modifier – the probability of ignition for an unknown source assuming

“good “ ignition controls

UK HSE Research Report 226, Development of a method for the

determination of on-site ignition probabilities (2004)

Cloud Area ~ 0.3 (235 m)2 = 16568 m2 = 1.66 hectares

~0.15

CCPS, Guidelines for Determining the

Probability of Ignition of a Released Flammable Mass, (2014)

A reasonable probability

of ignition for this case

may be 0.1 to 0.15 as

the process area is

likely electrically

classified with “good

controls”.



June 28, 2021 Slide - 28


Risk Analysis / Layers of Protection Analysis (LOPA)Scenario Number: Equipment Identification:

Raffinate Column

Date:

Consequence Description/

Category

Column Overfill resuting in flammable

release to the atmosphere and vapor

cloud explosion.

Risk Tolerance Criteria

(Category or Frequency)Potential for greater than 10 fatalities 1.E-06

Initiating Event

(typically a frequency)

Failure of Level Control Loop or Human

Performance Failure for filling the colmn

above level reading during startup

1.E-01

Enabling Condition or

Event

Probability of Ignition 1.E-01

Probability of Personnel in Affected Area

Probability of Fatal Injury

Others

1.E-02

Description ProbabilityFrequency

(per year)

Scenario Title:

Overfill Scenario during Startup

Conditional Modifiers (if applicable)

Frequency of Unmitigated Consequence

BPCSHigh Column Pressure shut off of heat to

the column reboiler1.E-01

Human InterventionBlowdown Tank High Level with

Procedure to close manual Feed valves1.E-01

SIFRaffinate Column High Level Alar with

Automated shutoff of Heat and Feeds1.E-01

Pressure Relief Device

Other Protection Layers

(must justify)

Restriced Access near Isom Unit during

Operation or Startup1.E-01

Safeguards (Non-IPLs)

1.E-04

1.E-06

Actions Required to Meet

Risk Tolerance Criteria

Notes

References (link to

originating hazard review,

PFD, P&ID, etc.)

LOPA Analysis (and Team

Members)

Independent Protection Layers (IPLs)

Total PFD for all IPLs

Ensure all Protections are Inspected, Tested, and Maintained to company standards. Restrict

placement of temporary occupied buildings beyond blast contour unless blast resistent

buildings are used.

Meets Tolerable Criteria for Harm to On-Site Personnel

Risk Tolerance Criteria Met? (yes/no)

Frequency of Mitigated Consequence



June 28, 2021 Slide - 29


Sustaining Performance

The existing safeguards were close to sufficient for managing this scenario to a tolerable risk

level had they been adequately maintained and some actions automated rather rely only on

operator response to an alarm. In addition to those listed in the LOPA worksheet, several other

alarms existed (such as high pressure) that may have contributed to reducing the overall

scenario frequency if the potential for column overfill would have been recognized.


What can go Wrong?


might it Happen?


Identify Chemical

and ProcessHazards

Estimate Frequency

AnalyzeConsequences

Analyze Risk

ImplementAdditional


Develop Scenarios


Analyzed


of Facility



June 28, 2021 Slide - 30

Risk Analysis Screening Tools (RAST)

Case Study – BP Texas CityRisk Analysis and Incident Investigation often use similar methods to better understand

the scenario. Risk Analysis “anticipates” what could go wrong and what the potential

consequences may be. For Incident Investigation, the Incident Outcome and

Consequences are known in addition to the actual weather conditions and wind

direction.

For the Raffinate Splitter, a column overfill would be one of many scenarios to consider.

It would need to be recognized that a Vapor Cloud Explosion could be a feasible

Incident Outcome for an Overfill loss event. CHEF was conservative in estimating blast

damage as actual wind direction was not toward the wooden trailers and a simplified

single Potential Explosion Site of “average” congestion and confinement was assumed.

However, the “order of magnitude” estimate of consequences seems reasonable.



Questions?

June 28, 2021 Slide - 31

ccps case study bp texas city an aiche technology alliance

Documents