industrial system for chemical inhibition of vapor cloud...

Industrial System for Chemical Inhibition of Vapor Cloud Explosions

Process Safety ConferenceDordrecht (The Netherlands)May 2019

Prof. dr. ing. lic. P. HoorelbekeProf. dr. ir. D. RoosendansTOTAL S.A., Paris, France

Project Characteristics

Process Safety Conference, Dordrecht (Nl), May 2019Industrial System for Chemical Inhibition of Vapor Cloud Explosions 2

o Start of project in 2004

o > 2 M€ of research (2005 – 2018)

o International collaboration with institutes & specialized organizations

o So far 3 PhD thesis dedicated to the project

o P. Hoorelbeke (2011)oO. Dounia (2017)o D. Roosendans (2018)

o Internal and external awards

Outline of Presentation


4. Industrial Application of Flame Inhibition Technology by Dry Powders of Alkali Metal Compounds

1. Context of Vapor Cloud Explosions

3. Experimental Study on Flame Inhibition by dry inhibitor powders

2. Principles of Combustion and Flame Inhibition

Context of Vapor Cloud Explosions






Context of Vapor Cloud Explosions 6 kg of natural gas


Context of Vapor Cloud Explosions


$$$$$$$$$$$

Refinery Units and Petrochemical Units in the World

NAMRefinery units: 1707Petrochemical units: 1693

SAMRefinery units: 634Petrochemical units: 619

EUROPERefinery units: 1869Petrochemical units: 2382

AFRICARefinery units: 460Petrochemical units: 219

FSURefinery units: 926Petrochemical units: 571

NE ASIARefinery units: 2065Petrochemical units: 5380

ASIA PACIFICRefinery units: 808Petrochemical units: 1732

MERefinery units: 450Petrochemical units: 737

Refinery units 8919Petrochemical units 13333TOTAL 22352

7

Process Safety Conference, Dordrecht (Nl), May 2019Industrial System for Chemical Inhibition of Vapor Cloud Explosions

Pasadena

(1989)

Exxon Valdez

(1989)

Chernobyl

(1986)

Piper Alpha

(1988)

Mexico City

(1984)

Sandoz

(1986)

Bhopal

(1984)

Macondo

(2010)

Amuay

(2012)

Fukushima

(2011)Tanjin

(2015)

8


Feyzin

(1966)

Umm Said

(1977)

Seveso

(1976)

Beek

(1975)

Antwerpen

(1975)

Los Alfaques

(1978)

Flixborough

(1974)

Ghislenghien

(2004)

Big Spring

(2008)

Texas City

(2005)

P36 platform

(2001)

Viareggio

(2009)

Skikda

(2004)

Buncefield

(2005)

La Mède

(1992)

Belpre

(1994)

Milford Haven

(1994)

Longford

(1998)

Deer Park, TX

(1997)

Erika

(1999)

Dakar

(1992)

Are Vapor Cloud Explosions a problem of the past ?

Major VCE

http://upload.wikimedia.org/wikipedia/commons/9/9d/Deepwater_Horizon_offshore_drilling_unit_on_fire_2010.jpg

http://upload.wikimedia.org/wikipedia/commons/9/9d/Deepwater_Horizon_offshore_drilling_unit_on_fire_2010.jpg

http://www.offshore-technology.com/projects/roncador/

http://www.offshore-technology.com/projects/roncador/

9


BayernOil (Vohburg, Germany), 1 September 2018

o Pressure drop reactor D-3401A gasoline desulfurization (140°C, 25 bar)

o Crack of approx. 1.5 m in length, starting directly at the support claw suspension (80 m³ reactor

o Gas alarm: in a few seconds increase to 20/40% LEL and beyond....44 seconds afterwards …vapor cloud explosion…

o CAT Cracker/gas plant, De-hexanizer , DIP , ETBE, gasoline desulfurization plant (OATS) field completely destroyed

o Proximity blast proof container for employees (nothing happened!) max 30-50 m

o Neighboring old control room destroyed (moved to BP Texas in new control room, but cable/junction over old control room)-> process control system completely lost

Vapor Cloud Explosions are not a problem of the past !

Vapor Cloud Explosion Probabilities

Average probability of VCE +/- 5.10-4/yrNumber of process units in TOTAL RC +/- 200Probability of major VCE in TOTAL RC Once every 10 years La Mède, 1992

Buncefield, 2005

10


Low probabilities ?

(once every10.000 years

per unit)

Barriers against Vapor Cloud Explosions

System Drawbacks

Isolation systemsDepressurization systems

Take time to be effectiveResidual vapor cloud to be considered

Water curtainsWater deluge

Limited efficiencyTime needed for full deployment

Blast wallsNot always practicalLimited efficiency

Blast designExpensiveSometimes difficult to achieve

Reinforced buildingsExpensiveSometimes difficult to achieve

11


Today, there are no effective active mitigating barriers to protect against the occurrence and

consequences of Vapor Cloud Explosions

Principles of Combustion and Flame Inhibition






Combustion Principles

13


Moses Gomberg, a chemistry professor at the University of Michigan, discovered organic free radicals in 1900

Radicals and Combustion Reactions

14


o Combustion reactions are radicular chain reactions . These chain reactions depend on the presence of flame radicals in sufficient amounts. Without these radicals, the chains are broken and combustion reactions cannot continue.

o Consider the combustion of hydrogen (H2) in oxygen (O2): 2H2 + 02 → 2H2O

H

O

H

C

H

H

H

H2 + M → 2H• + M

H2 + O2 → HO2• + H• + M

HO2• + H2 → H2O2 + H•

H• + O2 → O• + OH•

O• + H2 → O• + OH•

O• + H2O → OH• + OH•

OH• + H2 → H2O + H•

H• + H• + M → H2 + M

H• + OH• + M → H2O + M

Initiation(= creation of radicals)

Propagation & Branching(= multiplication of radicals)

Termination(= elimination of radicals)

Examples of reactions intervening in the combustion of H2

Inhibition of Vapor Cloud Explosions ? How ?

15


Vapor Cloud Explosion

But…oxygen available in ambient air

But…difficult to eliminate all hot

surfaces and ignition sources

But…large quantities of fuel in case of major leakage

Perturbation of Radicalar Reactions involved in Combustion

16


o Disturb chemical chain reaction ! But how ?

o Promotion of radical termination reactions !

→ Cooling of flames (but a lot of thermal agent needed)

→ Adding radical scavenging species (inhibitors)

→ Both (cooling + radical scavenging)

Experimental Study on Flame Inhibition by Dry Powders






Experimental Program 2008 - 2010

o Purpose:

To determine the relationship between inhibitor characteristics (concentration, type, etc.) and flame combustion velocity of a fuel-air mixture (20 l vessel)

To verify the effectiveness of inhibitors in medium scale conditions (50 m³)

o Number of tests:

316 experiments in a 20 l vessel

103 experiments in a 50 m³ modules

20 l vessel

50 m3 module

18


example of test.MOV

example of test.MOV

Inhibition Efficiencies

19


0.01 0.1 1 10

CO2

HCl

CF4

SO2

SF6

CH3Cl

SiF4

Cl2

SiHCl3

CF3Cl

CHF3

CF3CHFCF3

CHCl2F

BF3

BCl3

Si(CH3)4

HBr

CCl4

CH3I

CH3Br

i-C3H7I

CF3Br

Br2

SiCl4

GeCl4

CHBr3

AsCl3

(C2H5)3PO4

(CH3)3PO4

BBr3

SbCl3

POCl3

TiCl4

SnCl4

PSBr3

PCl3

PBr3

NaCl

PSCl3

CuCl

NaHCO3

K2SO4 (+Cab-O-Sil)

Na2CO3

KHCO3

CrO2Cl2

Pb(C2H5)4

Fe(CO)5

Relative efficiencies of flame inhibitorsBabushok & Tsang, 2000

Influence of Particle Size

20


Particle material Flame TypeParticle diameter

(µm)Reference

water Laminar stoechiometric premixed methane/air flame 10

Yang and Kee (2002)

Fuss et al (2002)

Fleming et al (2002)

water Laminar non premixed counter flow methane/air flame 15 - 20 Seshadri (1978)

water Laminar non-premixed counter flow propane/air flames 14Chin and Lefebvre (1983)

Fleming et al (2002)

water Turbulent flames 10 Van Wingerden (2000)

water Turbulent flames 30 Acton et al (1990)

water Turbulent flames 18 Sapko et al (1977)

water Counter diffusion methane flame ～ 20 Lentati and Chelliah (1998)

NaHCO3

KHCO3

Laminar propane/air counter flow non-premixed flame < 38 Fleming et al (1998)

NaHCO3 Heptane pool fire < 16Ewing et al (1984)

Ewing et al (1989)

NaHCO3 Laminar non-premixed counter flow methane/air-flame 0 – 10 Chelliah et al (2002)

Turbulent flame inhibited by NaHCO3, KHCO3, Na2CO3, K2CO3 < 20 Hoorelbeke (2011)

NaHCO3 Counter flow diffusion methane flame < 20Linteris et al (2001)

Wanigarathne et al (2000)

NaHCO3 Counter flow diffusion flame, liquid heptane 0 – 10Hamins et al (1994)

Trees et al (1997)

NaHCO3Simulation of hydrogen/air flame at φ = 0.5 (burning velocity of 50 cm/s) and φ = 1.4

(burning velocity of 300 cm/s)

< 17 µm

< 3 µm Mitani (1983)

NaHCO3 Premixed methane/air flame (φ=1.2) 12 - 28 µm Iya et al. (1975)

NaOH solution Laminar non-premixed methane/air flame 10 – 20 Lazzarini et al. (2000)

NaOH solution Laminar premixed and non-premixed opposed flow flames 10 – 20Wanigarathne et al. (2001)

Chelliah et al. (2002)

NaOH solution Counter diffusion methane flame (modeling) 15 Lentati and Chelliah (1998)

Flame Inhibition by Aqueous Solutions of K2CO3

21


For effective inhibition, the size of droplets must be very small (< 10 – 20 µm)

Human hair 10 µm droplet

Experimental Program: tests in 50 m³ module

22


example of test.MOV

example of test.MOV

Industrial Application of Inhibition Technology






Development of Inhibition Technology

2011

• First powder dispersion tests

• Desktop study of existing technologies

24

• Large scale powder dispersion tests

• Desktop study of existing technologies

2015-2016

2013-2014

2012

2017-2018

• Development of industrial implementation in the context of projects

• Preparation and execution of large scale explosion tests (California, US)

• Further development of engineering solution


Development of Inhibition Technology: Technology Selection

25


Development of Inhibition Technology: Dispersion Testing

26


Power jet

3 test campaigns in 2012

Development of Inhibition Technology: Dispersion Testing

27


30 m

40 m

4 meters high

Development of Inhibition Technology: Final Solution

o Inhibitor: KHCO3 or K2CO3, D50 = 20 - 30 µm

o Volume protected per skid : ~ 1250 m³

o Storage pressure: 16 barg

o Nitrogen volume : 15 Nm³ @ 300 barg

o Surface plot: ~1 m²

o Continuous injection of inhibitor (2 kg/s)

o Pre-ignition strategy

o Dispersion of the cloud of inhibitor : 20 s after the start of the leak

o Concentration of the inhibitor in the volume to be protected: 100 g/m³

o Sustained cloud of inhibitor during at least 5 minutes

28


Large Scale Testing (Livermore, US)

29



30


Slow deflagration, no inhibition

Detonation

31



32


Detonation, inhibition with KHCO3


33


Deflagration, inhibition with KHCO3


34


50 mbar

15.7bar

Industrial Application of the Technology

35


Industrial Application of the Technology

o 2 confirmed projects:

Steam cracker project in South Korea

Steam cracker project in United States

o Several potential other projects

36


powder skid

37


For more information, please contact:

o Pol Hoorelbeke ([email protected])o Dirk Roosendans ([email protected])

industrial system for chemical inhibition of vapor cloud...

Documents