mechanistic safety analysis of hydrogen based … · mechanistic safety analysis of hydrogen based...

1KIT – die Kooperation vonForschungszentrum Karlsruhe GmbHund Universität Karlsruhe (TH)

MECHANISTIC SAFETY ANALYSIS OF HYDROGEN BASED ENERGY SYSTEMS

W. BreitungInstitute for Nuclear and Energy Technologies

Karlsruhe Research CenterGermany

Second European Summer School on Hydrogen Safety, University of Ulster, Belfast,30 July- 8 August 2007


CONTENT OF PRESENTATION

• Presentation consists of two topics which are treated in parallel

Analysis of hydrogen releasein a private garage

Analysis tools and related physics

Combustion regimes

Gas transport and mixing

Turbulent deflagration

Detonation

Structural response

Mixture generation

Sequence of analysis steps

Combustion

Hazard potential

Consequences


OUR EXAMPLE FOR HYDROGEN ANALYSIS

• Oil peak behind us, hydrogen fueled cars in widespread use

• Returned from a trip late at night

• There was some small collision but apperentlyno domage to LH2-system

• Park car in private garage

• But at night the questions come ….

What would happen in case of a hydrogen leak?

What mixturescould develop?

Could they beflammable?

How fast couldthe burn be?

What would bethe pressure

loads?

What could be theconsequences?


GENERIC ARCHITECTURE OF AN LH2-TANK SYSTEM

Source: EU-Project EIHP-2, Final Report 2004

CRACK!


INVESTIGATED GARAGE SCENARIOS

222.3341000.34

1under-neath

trunk

22.334103.40Two times

10 x 20 cm²

70.2

Nr.Release Location

Release Temp.

(K)

Total Mass

(g)

Duration(s)

H2-Rate (g/s)

VentOpenings

Garage Volume

(m³)

CASEHYDROGEN SOURCEGEOMETRY

222.3341000.34

1under-neath

trunk

22.334103.40Two times

10 x 20 cm²

70.2

Nr.Release Location

Release Temp.

(K)

Total Mass

(g)

Duration(s)

H2-Rate (g/s)

VentOpenings

Garage Volume

(m³)

CASEHYDROGEN SOURCEGEOMETRY

• A thermal energy deposition of 1 Watt into a cryogenic LH2-tank leads to a boil-off of 170 g of gaseous hydrogen per day

• Assume here 5 release pulses per day, 34 g H2 each, with two different release rates


WHAT ARE THE IMPORTANT RISK DETERMINING PARAMETERS?

• Large spectrum of events possible, ranging from zero risk to destructionof garage

• What are the parameters influencing the outcome of such a leak scenario?

• Obvious first step is to understand mixture generation, defines initial and boundary conditions for further accident development

…. …. …. ….

…. …. …. ….

…. …. …. ….

…. …. …. ….


Analysis of a hydrogen releasescenario in a private garage





AN INITIAL ESTIMATE ON HYDROGEN CONCENTRATION

• We can make a first estimate on the hydrogen concentration in the garage by using a single volume approach

• Any risk?

• Why is result independent of release rate?

• Obviously the real situation is more complex

• Next approach is a CFD model

% 0.5 m 70

H g /2l 22.4 H g 34

garage of volumereleased H volume

H fraction volume 32222

≈⋅

=≈


T>1000KIRWST

Rekos

Physical models of 3d code GASFLOW (1)

• Conservation equations of fluid flow (fully compressible, 3-dim. Navier- Stokes)

• Thermophysical properties of components (JANAF)(internal energy, specific heats, 25 components including two-phase water)

• Molecular transport coefficients (CHEMKIN)(thermal conductivity, dynamic viscosity, binary diffusion coefficients)

• Convective and radiative heat transfer between gas and structure

• Heat conduction within structures

• Condensation and vaporization of water (film, droplets, sump)


T>1000KIRWST

Rekos

Physical models of 3d code GASFLOW (2)

• Lumped- parameter sump model

• Boundary layer model for wall shear stress

• Turbulence modeling (algebraic, k-ε)(effects on molecular transport coefficients)

• Accident mitigation measures(Recombiner and igniter models, containment inertisation)

• Ventilation systems(1-dim. ducts, pipes, junctions, blowers, dampers, valves, filters, etc)


GASFLOW EQUATIONS

Fully compressible Navier-Stokes, expressed in integral form for finite volume discretisation

Momentum conservation

∂∂t

( )dV dS pd dV dSV V S

ρ ρ ρ τu u A - S g A2

S S

= + −∫∫ ∫∫ ∫

( )- D AS

d mV

dS S dV+ ∫∫

momentumflux

pressuregradient

gravity viscousstresses

drag frominternal surfaces

& flow restrictions

momentumsources

Mass conservation

Total mass:

∂∂

ρ ρ ρtdV dS S dV

V V

= uAS∫∫ ∫+

Component α :

( )∂ ρ ρα α α ρ α∂tdV dS dS S dV

V VuA - J A

S S

= +∫∫ ∫∫ ,

convection diffusion chem. reactiontwo-phase change

convection sources (inflow, droplet depletion)

Internal energy conservation

( )edV e dS p d pV

Vt

dVV V

ρ ρ ∂∂

u A - uA SS S

= − ⎛⎝⎜

⎞⎠⎟∫∫ ∫∫

( )- qAS

dS S dVeV

+ ∫∫ e x e= ∑ α αα

pV work due to phase change

energy flux energy sources(thermal,conductivity,........)

(combustion, phase change,heat transfer)

∂∂t

HO2

convection pV work

J.R. Travis et al, Report FZKA-5994 (1998)

xx

R6

R5

Ring Room R9

He- Source

10 Vol % HeIsosurface

R6

R5

Ring Room R9

He- Source

10 Vol % HeIsosurface

Vertical He- jet into air (42 m/s) for 200 sVertical He- jet into air (42 m/s) for 200 s

0

5

10

15

20

0 200 400 600 800Time (s)

9KP000K76FZK/GASFLOW

0

5

10

15

20

0 200 400 600 800Time (s)

9KP000K76FZK/GASFLOW

Vertical distance from injection location (m)

0

40

80

0 2 4 6

FZK/GASFLOW

6kp270M18

Vertical distance from injection location (m)

0

40

80

0 2 4 6

FZK/GASFLOW

6kp270M18

0

5

10

15

0 200 400 600 800Time (s)

6KP270M18FZK/GASFLOW

0

5

10

15

0 200 400 600 800Time (s)

6KP270M18FZK/GASFLOW

Distance from jet axis (m)-2 -1 10 2

0

40

80

x

x

xx

x

xx x

EXPERIMENTFZK/GASFLOWx

Distance from jet axis (m)-2 -1 10 2

0

40

80

x

x

xx

x

xx x

EXPERIMENTFZK/GASFLOWx


GASFLOW VERIFICATION

• 3d code GASFLOW used and developed at FZK for hydrogen distribution simulation. Large verification matrix:

Report FZKA-7085 (2005), www.fzk.de/hbm


EXAMPLE FOR RECENT GASFLOW VERIFICATION

• German national benchmark, test TH7 in Thai facility with condensation• Blind pressure prediction of CFD codes

lower source

Twall > 330K

Twall < 300K

steam

1.0

1.2

1.4

1.6

1.8

[bar

]

[sec]

CFX

STAR-CD

Fluent

Gothic

GASFLOWExperiment

upper inj35 g/s

lower inj 135 g/s

lower inj 25 g/s

Inj. points

P. Royl, IKET


Analysis of a hydrogen releasein a private garage




Mixture generation


GASFLOW SIMULATION OF GARAGE SCENARIO

Isosurface with 4 vol% H2, depicts flammable mixture in garage

• Case 1: release rate 3.4 g H2 / s for 10 seconds

volume fraction H2


GASFLOW SIMULATION OF GARAGE SCENARIO

Isosurface with 4 vol% H2 , depicts flammable mixture in garage

• Case 2: release rate 0.34 g H2 / s for 100 seconds

volume fraction H2


RESULTING HYDROGEN CLOUD IN GARAGE

• Computed dimension of combustibleH2-air cloud in garage (4…75% H2)

• Characteristic size of combustible cloudexpressed as dCC = (Vcc)1/3

• Combustible cloud size strongly dependenton release rate, is result of balance betweensource strength and sinks, or releaserate and mixing mechanisms

Cas

e2

0.34

g H

2/s fo

r 100

sC

ase

13.

4 g

H2/s

for 1

0 s

d (cm)cc

0

10

20

30

40

50

60

70

80

90

0 50 100 150 200 250

stable

transient

0

20

40

60

80

100

120

140

160

180

0 10 20 30 40 50 60

d (cm)cc

Time (s)

Time (s)


WHAT IS RISK FROM COMBUSTIBLE CLOUD?

Case 1 Case 2

• How would you judge the hazard in both cases?

• Who would switch on lights in the garage?

• What physical quantities determine hazard potential of a combustible H2-air cloud?




Combustion regimes

Transport and mixingMixture generation



COMBUSTION REGIMES

• H2 – air mixtures can burn in different modes / combustion regimes

Inert, no stableflame propagationvfl = 0

Laminar deflagrationvfl ≈ 1 m/s, Ma << 1

Fast turbulent deflagrationvfl ≈ 300 m/s, Ma ≈ 1

Detonationvfl ≈ 1500 m/s, Ma >> 1

• Change of mode possible by transition process

Ignition Flame acceleration Deflagration-to-detonation transition (DDT)

2222

KIT – die Kooperation vonForschungszentrum Karlsruhe GmbHund Universität Karlsruhe (TH)

PEAK OVERPRESSURES FROM HYDROGEN-AIR FLAMES

• The maximum flame speed generally governs the damage potential• Which combustion regime develops for given mixture and geometry?• How fast can it burn?

x

p pmax

0.01

0.1

1

10

100

7 10 100 1000flame speed (m/s)

overpressureratio (p-p)/p

max00

FZK-tube (closed) RUT ventedFLAME ventedplanar model,12%Hplanar model,28%H

2

2

Maximumacceptablestatic loadfor typical innercontainmentstructures(1 ton / m2)

(pm

ax-p

0)/p

0

p0

2323


IGNITION

• Combustion requires an ignition source and a burnable mixture

• Many potential ignition sources exist

• More than 90% of incidents with GH2 lead to ignition, cause often unknown

• Ignition difficult to exclude in a hydrogen safety analysis, conservatively the presence of an ignition source may be assumed

• Controlling factor is then flammability of mixture, well known for H2-air

G:1 G:7 G:5 G:4 G:3 G:0 G:2 G:8 G:6 G:90

5

10

15

20

25

30

35

40

45

Frac

tion

of ig

nitio

n so

urce

[%]

Ignition sources

G:1 open fireG:2 mechanical sparkG:3 electrical sparkG:4 hot surfaceG:5 static discharge

G:6 catalytic surfaceG:7 self-ignitionG:8 othersG:9 unknownG:0 no ignition

Survey of 287 accidents with hydrogen

Kreiser et al, Report Univ. Stuttgart IKE 2-116 (1994)

2424


FLAME ACCELERATION

• Conservative conditions for flame acceleration in hydrogen mixtures were investigatedin closed obstructed tubes, e.g. FZK 12m-tube


RESULTS OF FLAME ACCELERATION EXPERIMENTS

D

flame velocity (m/s)

0 10 20 30 40 50 60 70x/D

0

200

400

600

800

1000

1200

v,m

/s

BR=0.6 H2-Air174mm

10%H211%H215%H2

520mm10%H211%H2

80mm10%H211%H213%H2

350mm9%H211%H217.5%H2

• Lean hydrogen mixtures in obstructedtubes with different tube diameters Dand 60% blockage ratio (BR)

• Two distinct regimes with slow and fast flame propagation are observed

H

0 5 10 15 20 25 30 35x/D

0

200

400

600

800

1000

1200

v, m

/s

350 mmBR=0.6

H2-air-CO2φ = 0.5

40%CO2

35%CO2

30%CO2

27.5%CO2

25%CO2

20%CO2

15%CO2

10%CO2

5%CO2

2.5%CO2

0%CO2

FA

x

2626


FLAME ACCELERATION CRITERION

• Summary of experiments with different H2-O2- dilutend (N2, Ar, He) mixturesin obstructed tubes of differentscales

• Each point represents oneexperiment

• Results of data evaluation: expansion ratio σ is mixtureproperty which governsflame acceleration limit

• No flame acceleration forσ < 3.75 ± 0.1

(10.5% H2 in dry air)

fast flames

unstable flamesslow flames

σ = 3.75

FZK

FZK

FZK

FZK

FZK / KI

KI

KI

KI

KI

Tube diameter L / Laminar flame thickness δ x 10-3

different length scales LFZK

FZK-tube: 350 mmKI

Driver tube: 174 mmTorpedo tube : 520 mm

RUT : 2400 mmFZK / KI

δ ~ 0.1 mm~

30.1 1.0 10.0

4

5

6

7global quenchingquenching-reignitionchoked flamesquasi-detonation

T 300 K~~

Expa

nsio

n ra

tio σ

(= ρ

ub/ρ

b)

fast flames

unstable flamesslow flames

σ = 3.75

FZK

FZK

FZK

FZK

FZK / KI

KI

KI

KI

KI

Tube diameter L / Laminar flame thickness δ x 10-3

different length scales LFZK

FZK-tube: 350 mmKI

Driver tube: 174 mmTorpedo tube : 520 mm

RUT : 2400 mmFZK / KI

δ ~ 0.1 mm~

30.1 1.0 10.0

4

5

6

7global quenchingquenching-reignitionchoked flamesquasi-detonation

T 300 K~~

Expa

nsio

n ra

tio σ

(= ρ

ub/ρ

b)D

D

In lecture notes


Fully obstructed tube (prototypic mode B)

accelerating flame DDT stable (quasi)detonation

spark ignition

pH / air2

burned gas

obstacle section flame precursor shock

conus

1 m 5-6 m

Partially obstructed tube with conus (prototypic mode A)

0,5 m

spark ignition

Shock tube with conus (idealized mode A)

p

He H / air2

high pressuresection

burst membrane

low pressure section conus

3 m/1m 9 m

0.35 m

3m

1m

DEFLAGRATION-TO DETONATION TRANSITION

• Two different modes of DDT have beenobserved

- shock focussing- detonation on-set in turbulent

flame brush

• Present here one example for DDTwith pressure wave emitted froman obstructed region and focussedin a conus


TURBULENT DEFLAGRATION EXPERIMENT WITHOUT DDT• Partially obstructed tube with conus, 15 % hydrogen in air

5 m

350 mm


TURBULENT DEFLAGRATION EXPERIMENT WITH DDT

• Partially obstructed tube with conus, 16.5 % hydrogen in air

• Result: focussing of pressure waves emitted from a fast turbulent flame can triggera detonation on other parts of the system

3030


CRITERION FOR DDT

• Correlation of all experimental data with given definitions of D and detonation cell size datashows that detonations areonly possible for D/λ > 7

• Current uncertainty in detonationcell size λ ≈ factor 2

2 3 5 2 3 5100 1000 10000Geometrical size L, mm

2

3

5

2

3

5

2

3

5

2

1

10

100

D/λ

No DDT, BR<0.5

No DDT, BR>0.5, rooms

DDT, BR<0.5

DDT, BR>0.5, rooms

L = 7λ

λ accuracy limits

d1

d1

d1

d1

d1

d1

d1

d1

c1

c1

t4

t4

t4

c2

d2

d2

d2

d2

d2

d2

d2

d2

g3

s1s1

s1

s1

s1

s1

s1s1

s1

s1

s1

s1

t3

t3

t3

t3a1

a1a1a1

a1

a1

a1

a1a1a1a1

a1a1

a1

a1a1

a1a1a1

f1

f3

r2

r2

r5

r5

r5

r5r5

r5r5

r5

r5

r5

r5

r5

r5

r5r5

r5r5

r6

r6

v2

b1b1

b1

b1

b1

b2

b2

b2b2

b2

b3b3

b3

b3

b3

b3

b4b4b4b4b4

b4

b4

b5b5

b5

b5

b5

b6b6

b6b6b6

b6

m3m3 m4m4m4 m5

c3

d3

d3d3

d3

d3

d3d3d3d3d3d3

d3

d3 g1

g1

g1

g1

g1

g1g2

g2g6

g6

g6

g6

g7

g7

g7

g7

g8

g8

g8

g9

g9r1s2

s2

s2s2s2

s2s2

s2

t1

t1

r3r4r4r4r4

r4

r4

r4r4

r4

r7

r7

v1

d4

riri

ri

ri ririri

ri

d1

d1

t4

d2d2g3

s1a1a1

a1a1

a1

a1

a1a1a1

a1

a1

a1a1a1a1

f1

f3

r2r2

r5r5

r6v2

b1b1

b2

b2

b3b3

b4

b5b5b6b6

m3

m3m3

m4

m4m4 m5

m1 m2m6

m6m6m7m7m7 m8m8m8

g1g1

g1

g7

g7

g8

g8

r1 r3r4

r4

r4r4

r4

r4

r4

r4

r7r7v1

ri

riri

ri

ri

riD/

λ

Characteristic geometrical size of reacting mixture D (mm)

keinkein

Fehlergrenze

• Experiments on DDT in differently sized and shaped facilities have shown that a certainminimum scale is required for DDT

• In accident scenarios D/λ can varyby orders of magnitude, criterionhas therefore predictive capability

3131


DETONATION CELL SIZES

• Detonation cell sizes (in cm) of H2-air-steam mixtures at 375 Kand 1 bar initial pressure.

Dry hydrogen concentrationis defined as H2 / (H2 + air)

H (trocken) (vol%)

2

H0 (vol%)2

State of the Art Report by a Group of Experts „Flame Acceleration and Deflagration – to –Detonation Transition in Nuclear Safety“, Nuclear Safety NEA/CSNI/R(2000)7, August 2000

3232


SUMMARY OF CRITERIA

• Criteria for possible occurrence of fast combustion regimes were evaluated from many experiments with various H2-mixturs on different scales

• Transition phenomena cannot be modeled numerically on large building scale• Criteria allow selection of fastest possible combustion mode from computed

H2-air cloud composition and scale

H 2

8%

x=16%0

4%

Luft

H 2

8%

x = 16%0

4%

air

σindex =

< 1 ausgeschlossen=

> 1 möglich

σ (Durchschnittsmischung)σ (T)critical

Flammenbeschleunigung

σindex=

< 1 excluded= > 1 possible

σ(average mixture)

σ (T)critical 7λ

Übergang zur Detonation

=

< 1 exluded= > 1 possible

(Cloud volume)1/3

σindex >1

7λ (average mixture)D

Detonation transitionFlame acceleration

3333




Combustion regimes

Distribution and mixingMixture generation


Hazard potential


COMPUTED HAZARD PARAMETERS FOR GARAGE SCENARIOSC

ASE

20.

34 g

H2/s

for1

00s

• Volume of cloud with potential for spontaneous flame acceleration (10.5 to 75 % H2)

• DDT index of cloud (10.5 to 75 % H2)

• Dimension of combustible cloud, 4 to 75 % H2, dcc = (Vcc)1/3

0

20

40

60

80

100

120

140

160

180

0 10 20 30 40 50 60

d (cm)cc

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0 10 20 30 40 50 60

V (m³)fa

0

0.5

1

1.5

2

2.5

3

0 10 20 30

DDT

no DDT

40 50 60

D7λ

CA

SE 1

3.4

g H

2/s fo

r10

s

no DDT

0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

0 50 100 150 200 250

D7λV (m³)fa

0

0.02

0.04

0.06

0.08

0.10

0.12

0 50 100 150 200 250

d (cm)cc

0

10

20

30

40

50

60

70

80

90

0 50 100 150 200 250

Time (s)

Time (s)Time (s)Time (s)

Time (s)Time (s)


HAZARD POTENTIAL FOR GARAGE SCENARIOS

• Risk parameters show strong dependence on H2 release rate

- Case 1:(3.4 g H2/s)

• Only Case 1 followed in further safety analysis

• Continuous potential for slow deflagration(≈ 20 g of 34 g)

• potential for supersonic combustion regimes (and ignition)during the release period

• high release rate not tolerable without mitigationmeasures

• only small potential for slow deflagrations, naturalmixing processes sufficient

• release rate (and mass) seems tolerable forpresent garage design

- Case 2:(0.34 g H2/s)


COMBUSTION EXPERIMENTS FOR CASE 1

• Up to 20 g of hydrogen would be in burnable concentrations

• A significant part of this could potentially burn with high flame speeds

• What would be pressure loads and consequences from a local explosion inthe garage?

• Outcome uncertain, experiments performed in test chamber simulating thegarage




Combustion regimes



Combustion (Experiments)

Hazard potential

3838


COMBUSTION UNITS• To obtain conservative pressure loads, combustion units were developed providing the

fastest possible flame speed for a given H2 mass• Cubes were made for 0.5, 1, 2, 4, 8 and 16 g of H2, which can be inserted into each other• Wire grids 6.5 x 0.65 mm, 12 layers between cubes

4 g0.5 g

Hydrogen injection device cubes covered with plastic,filled with stoichiometric H2 - air mixture


0R

Pressure transducers

mH2pmax

i v ∫>Δ

+ Δ=0

)(p

dttpI

UNCONFINED TEST OF COMBUSTION UNIT

• Peak overpressure and impulsemeasured as function of distanceto characterize blast effects fromcombustion unit

4040


FLAME SPEEDS IN COMBUSTION UNITS

0.5 1 2 4 8 160

200

400

600

800

1000

1200

1400

1600

1800

2000

0 0.1 0.2 0.3 0.4 0.5

R, m

Flam

e sp

eed,

m/s

8 g

8 g16 g

16 g4 g

cube border

• The flame acceleration inside the combustion units was measured with photodiodes

• For 8 and 16 g H2 detonation speeds are obtained at the outer edge of the cube

center of the cube

4141


MAXIMUM OVERPRESSURE VS DISTANCE

• Measured peak overpressures Δp+ in unconfined tests with combustion units of 0.5 to 16 g H2

• Data are well reproducible

Δp+

t

I+

0

1

2

3

4

5

0 5 10Distance R from center of cube, m

Peak

over

pres

sure

Δp+

, bar 2g

2g1g

1g

0.5g

0

2

4

6

8

10

12

0 5 10

16g

16g

8g8g

4g

Distance R from center of cube, m

Peak

over

pres

sure

Δp+

, bar

4242


IMPULSE VS DISTANCE

Δp+

t

I+

• Measured positive impulse I+ values from unconfinedcombustion units

0

20

40

60

80

100

120

140

160

180

0 5 10

Distance R, m

Impu

lse,

Pa*

s

16g

16g

8g8g

4g

0102030405060708090

100

0 5 10

Distance R, m

Impu

lse,

Pa*

s

2g

2g1g

1g

0.5g

4343


SCALED PEAK OVERPRESSURES VS DISTANCE

• Use of Sachs scaling collapses measured peak overpressures to universal correlation for≥ 1 g H2, E = total energy of explosive charge

• Combustion units provide conservative and well defined overpressures

0.01

0.1

1

10

0 2 4 6

Scaled distance R (p0 / E)1/3

Scal

edpe

akov

erpr

essu

reΔ

P+/p

0

16g

16g

8g8g

4g

Gas detonationTNT (= energy)

0.01

0.1

1

10

0 5 10

2g2g

1g

1g0.5g

Gas detonation

TNT (= energy)

Scal

edpe

akov

erpr

essu

reΔ

P+/p

0

Scaled distance R (p0 / E)1/3

4444


TEST CELL FOR GARAGE SIMULATION

• Dimensions 5.5 x 8.5 x 3.4 m, about 160 m3

• Air flow ≤ 24.000 m3/h,up to 1 air exchange in 24s

• Controlled air flows in chamberpossible

• All ventilation systems explosion protected

• Test cell used for simulation of garage /confined volume

Ventilation ducts

Ventilation turbines

Test chamber(garage)

Lower compartment


0100

200300

400500

0

50

100

150

200

250

300

0

200400

6008000

100200

300400

500

0

50

100

150

200

250

300

0

200400

600800

DruckaufnehmerBeschleunigungs-sensoren

3A

3B

1B

13A

7B

2A

1A

4A5A

6A 7A

8A15A

14A

2B

4B

5B

6B8B

16A

Höhe [cm

]

Breite [cm] Länge [

cm]

TürZünd-ort

INSTRUMENTATION OF GARAGE

• The instrumentation included pressure and acceleration sensors at different locations,covering flat surfaces, (2d) edges and (3d) corners

Pressure sensorsAccelerationsensors

location of combustion unit

Height (cm)

Width (cm)Length (cm)


LOCAL HYDROGEN EXPLOSIONS IN A GARAGE

Experimentwith 8g H2

H2 - mass:

- 1g- 2g- 4g- 8g- 16g


0,188 0,190 0,192 0,194 0,196 0,198 0,200 0,202 0,204

-0,020

-0,015

-0,010

-0,005

0,000

0,005

0,010

0,015

0,020

Beschleunigungsaufnehmer 1B

1g H2Experiment2Experiment3

Vol

t

Zeit [s]0,188 0,190 0,192 0,194 0,196 0,198 0,200 0,202 0,204

-0,020

-0,015

-0,010

-0,005

0,000

0,005

0,010

0,015

0,020


1g H2Experiment2Experiment3

Vol

t

Zeit [s]

0,19 0,20 0,21 0,22 0,23

-0,03

-0,02

-0,01

0,00

0,01

0,02

0,03

0,04


1g H2 Experiment2Experiment3

Vol

t

Zeit [s]0,19 0,20 0,21 0,22 0,23

-0,03

-0,02

-0,01

0,00

0,01

0,02

0,03

0,04



Vol

t

Zeit [s]

0,20 0,21 0,22 0,23 0,24 0,25 0,26 0,27

-0,02

-0,01

0,00

0,01

0,02

0,03 Beschleunigungsaufnehmer 9A

1g H

2

Experiment2Experiment3

Volt

Zeit [s]0,20 0,21 0,22 0,23 0,24 0,25 0,26 0,27

-0,02

-0,01

0,00

0,01

0,02

0,03 Beschleunigungsaufnehmer 9A

1g H

2

Experiment2Experiment3

Volt

Zeit [s]0,192 0,194 0,196 0,198 0,200 0,202 0,204

-0,04

-0,03

-0,02

-0,01

0,00

0,01

0,02

0,03

0,04

0,05

0,06

0,07

0,08

0,09

Druckaufnehmer 16A


Übe

rdru

ck [b

ar]

Zeit [s]

0,185 0,190 0,195 0,200 0,205 0,210

-0,10

-0,05

0,00

0,05

0,10

0,15

0,20

0,25

Druckaufnehmer 2B

1g H2

Experiment1Experiment2Experiment3

Übe

rdru

ck [b

ar]

Zeit [s]

0,190 0,195 0,200 0,205 0,210 0,215

-0,05

-0,04

-0,03

-0,02

-0,01

0,00

0,01

0,02

0,03

0,04

0,05

0,06Druckaufnehmer 5A1g H2


Übe

rdru

ck [b

ar]

Zeit [s]

0,192 0,194 0,196 0,198 0,200 0,202 0,204

-0,04

-0,03

-0,02

-0,01

0,00

0,01

0,02

0,03

0,04

0,05

0,06

0,07

0,08

0,09

Druckaufnehmer 16A


Übe

rdru

ck [b

ar]

Zeit [s]0,192 0,194 0,196 0,198 0,200 0,202 0,204

-0,04

-0,03

-0,02

-0,01

0,00

0,01

0,02

0,03

0,04

0,05

0,06

0,07

0,08

0,09

Druckaufnehmer 16A


Übe

rdru

ck [b

ar]

Zeit [s]

0,185 0,190 0,195 0,200 0,205 0,210

-0,10

-0,05

0,00

0,05

0,10

0,15

0,20

0,25

Druckaufnehmer 2B

1g H2


Übe

rdru

ck [b

ar]

Zeit [s]

0,190 0,195 0,200 0,205 0,210 0,215

-0,05

-0,04

-0,03

-0,02

-0,01

0,00

0,01

0,02

0,03

0,04

0,05



Übe

rdru

ck [b

ar]

Zeit [s]0,190 0,195 0,200 0,205 0,210 0,215

-0,05

-0,04

-0,03

-0,02

-0,01

0,00

0,01

0,02

0,03

0,04

0,05



Übe

rdru

ck [b

ar]

Zeit [s]

REPRODUCIBILITY OF MEASURED DATA

• The experiment with 1 g H2was performed three times

• Acceleration and pressuresensors show very good reproducibility of measuredsignals

• Complex, but reproduciblepressure waves are createdin confined local explosionsof H2-air mixtures

2


0,182 0,183 0,184 0,185 0,186 0,187 0,188

-0,4

0,0

0,4

0,8

1,2

1,6 1g2g4g8g16g

Übe

rdru

ck[b

ar]

Zeit [s]0,182 0,183 0,184 0,185 0,186 0,187 0,188

-0,4

0,0

0,4

0,8

1,2

1,6 1g2g4g8g16g

Übe

rdru

ck[b

ar]

Zeit [s]

0,193 0,194 0,195 0,196 0,197

-0,1

0,0

0,1

0,2

0,3

0,41g2g4g8g16g

Übe

rdru

ck [

bar]

Zeit [ s]

COMPARISON OF OVERPRESSURES

• Pressure sensor 2 B,floor near combustion unit

• Pressure sensor 8 A,back wall, half wall height

• Pressure signals very consistent in timing, amplitudes increase systemarically with H2 mass,reproducible pattern of reflected pressure waves in confined volume.

Ove

rpre

ssur

e[b

ar]

Ove

rpre

ssur

e[b

ar]

Time [s] Time [s]




Combustion regimes




Combustion

Hazard potential

5050


Re =t u'L t

ν

Turb. integral length scale Lt / Laminar flame thickness δ

Evolution inobstructedlarge scalecombustion

Turb

ulen

cein

tens

ityu‘

/ La

min

arfla

me

velo

city

SL Da < 1

Da = 1

Da > 1, Ka >1

Ka = 1

Ka < 1

Ret < 1

Da = Turb. Transport time (makro)

Laminar reaction timeLt / u'

δL/ SL=

Ka = Laminar reaction time Turbulent transport time (Kolmogorov scale)

δL / SLlΚ / u 'K

=

Flameshapes

PDFs

laminarflame

foldedflame

wrinkledflame

homogeneousreaction

Unburned gas

Burned gas

Reaction zoneKa < 1 Da > 1 Da = 1 Da << 1

thickened flame

TURBULENT DEFLAGRATION REGIMES

5151


COM3D EQUATIONS

A. Kotchourko, IKET

• COM3D under development at FZK for simulation of turbulent deflagration

5252


COM3D VERIFICATION (1)

• Large scale experiments performed in RUT facility near Moscow (FZK, CEA, partly NRC), H2-air, H2-air-steam

- Total length 62 m- Total volume 480 m3

- First channel with obstacles- Second part without obstacles

5353


A. Kotchourko, IKET

COM3D VERIFICATION (2)

Distance (m)

Overpressure(bar)

0,2 0,3 0,4 0,5 0,6

Time (s)

0

5

10

15

20

25

30

35

Experiment RUT13H=11%BR=0,3p=1barT=283K

2

0

0

COM-Code c=7 f

Experiment

0

2

4

6

8

10

12

Distance (m)

Overpressure(bar)

0

2

46

8

10

12

14

16

100 150 200 250 300

Time (ms)

0

5

10

15

20

25

30

35

Experiment RUT21H=12,5%BR=0.6p=1barT=283K

2

0

0

COM-Code c=7 f

Experiment

Distance (m)

Overpressure(bar)

Distance (m)

Overpressure(bar)

0

2

4

6

8

10

150 200 250 300 350

Time (ms)

0

5

10

15

20

25

30

35

Experiment RUT23H=11,2%BR=0.6p=1barT=283K

2

0

0

Experiment COM-Code, c=6 f

• Numerical simulation of large scale RUT experiments with hydrogen-air and hydrogen-air steam mixtures. Standard k-ε and Eddy-Break-up model.

• Venting in experiments, no venting in simulation

Dis

tanc

e (m

)

Dis

tanc

e (m

)

Dis

tanc

e (m

)

5454




Combustion regimes




Combustion simulation

Hazard potential


SIMULATION OF UNCONFINED TESTS

• The unconfined tests with different combustion units were simulated with COM3D

• The COM3D combustion model was fitted to the measured flame speed in the combustion units

• The calculated peak overpressures agree with the experimental values and follow Sachs scaling

Δp+

/p0

R (p0 / E)1/3


COM3D COMBUSTION SIMULATION

• 3d pressure field, calculated isosurface for 1.1 bar

• Test with 8g H2


COMPARISON OF OVERPRESSURES

• Good agreement, remaining differences are due to geometry simplification and rigid wall modelin simulation

over

pres

sure

[bar

]

over

pres

sure

[bar

]ov

erpr

essu

re[b

ar]

over

pres

sure

[bar

]

Time [s]

Time [s] Time [s]

Time [s]




Combustion regimes


(Detonation)

Structural response




Hazard potential


STRUCTURAL RESPONSE

• What are effects of blast loads on the structure?

6161


SINGLE-DEGREE-OSCILLATOR MODEL FOR STRUCTURAL RESPONSE

• Simplest model for structural response is SDO model

• Describes ground mode (first harmonic) of structural element which is represented by lumped values for mass, stiffness and damping of motion

• Tool to understand basic effects of transient pressure loads on global displacement of element

• In FEM analysis also higher modes included, but superposition of different effects, results not so transparent

xmax

k

p(t)

mt

D

6262


BLAST LOADED ELASTIC OSCILLATOR (1)

0 0) (t x 00) x(twith

ep kx xm

ep kx -

F xm

load

load

Tt

Tt

i i

====

Δ=+

Δ+=

=

−+

−+

∑

&

&&

&&

( )( ) ⎥

⎦

⎤⎢⎣

⎡+ω−

ω+

=Δ

−+

loadt/T

loadload

load etcosωT

t sinωT

ωT/kp

x(t)2

2

1

/kp x0x

max+Δ=

=&&

( )osc

1/2

T2πperiod oscillatork/mω ===

• Equation of Motion

Static maximum deflection

• Solution

where

• Damage is determined by maximum displacement xmax,can be found from solution by setting x(t) = 0

• Scaled displacement = f(scaled loading time)

)Tf( /kp

xload

max ω=Δ +

k

m t

x

loadT/tep)t(p −+Δ=

t

p(t)

Δp+

. scal

edde

form

atio

nx m

axscaled loading time T

mkT =ω

T = 1 ms 5 ms

scal

edde

form

atio

nx m

axscaled loading time T

mkT =ω

/ (P

* / k

)

6363


BLAST LOADED ELASTIC OSCILLATOR (2)

• Asymptotes for maximum deflection /deformationcan be computed from energy balances

• Quasistatic loading realm (T load >>Tosc)- strain energy = work on structure

½ kx2max = Δp+ · xmax

2k/p

xmax =Δ +

• Impulsive loading realm ( T load << Tosc)- initial kinetic energy = strain energy

2max

2

2max

20

kx21

2mI

kx21mv

21

=

=

loadload/max T T)

mk(

k/px

ω==Δ +

21or I)

km1(x 1/2max ⋅=

maximum deformation is proportional to blast wave impulse I

Impulsiveasymptote

Quasistaticasymptote

Scaled loading time ωT

loadT/t Tp epI load ⋅Δ=Δ= +∞ −+∫0

dynamic maximum deflection is two times staticdeflection (DLF = 2) S

cale

ddi

spla

cem

ent

2

W.E. Baker, P.A. Cox, P.S. Westine, J.J. Kulesz, R.A. Strehlow; Explosion Hazards And Evaluation; Fundamental Studies in Engineerings, 5; Elsevier

6464


max load

loadmax

21loadmax

x~ I T pT

1 x(km) p

T1

kxp

=Δ

=Δ

ω=

Δ

+

+

+

OSCILLATOR RESPONSE: ANOTHER VIEW

• Often oscillator response is presented with inverted ordinate and unscaled load parametersΔp+ and Tload

• Quasistatic asymptote

Maximum deflagration xmax is only proportional to applied peak overpressure Δp+, independent of loadduration• Impulsive asymptote

Maximum deflection xmax is proportionalto applied impulse

positive peakoverpressure

logΔp+

log of positive overpressure duration Tload

I2I1

xmax1

xmax2

xmax2

xmax1

Δp+2

Δp+1

2kx p

21

kxp

max

max

=Δ

=Δ

+

+

6565




Combustion regimes


(Detonation)

Structural response




Hazard potential

Consequences


STRUCTURAL DAMAGE FROM CONFINED LOCAL H2 EXPLOSIONS IN GARAGE

10000Duration of positive overpressure T [ms]

Pos

itive

pea

kov

erpr

essu

reΔ p

+[P

a]

103

104

105

106

Δp+I+

t

T+=2I+/Δp+

T+

Unconfined gaseous detonations

1,4m

1,4m

2,4m

2,4m

2,8m

5.6m

4,0m

6,3m4,0m

FZK Exp. 2g H2

FZK Exp. 16g H2

0.1 1 10 100 1000

Partial demolition, 50%- 75% of walls destroyed

Major structural damage, some wrenchedload bearing members fails

Minor structural damage, wrenchedjoints and partitions

2g H2

1 m

glass breakage

Area 1m2

4.7 mm thick

Area 3m2

4.7 mm thick

16g H2

2 m

5 m

450 Pas

300 Pas110 Pas

10000[ms]

Δ+

[Pa]

103

104

105

106

Δp+I+

t

T+=2I+/Δp+

T+


1,4m

1,4m

2,4m

2,4m

2,8m

5.6m

4,0m

6,3m4,0m

FZK Exp. 2g H2

FZK Exp. 16g H2

0.1 1 10 100 1000




2g H2

1 m

glass breakage

Area 1m2

4.7 mm thick

Area 3m2

4.7 mm thick

16g H2

2 m

5 m

10000[ms]

Δ+

103

104

105

106

Δp+I+

t

T+=2I+/Δp+

T+


1,4m

1,4m

2,4m

2,4m

2,8m

5.6m

4,0m

6,3m4,0m

FZK Exp. 2g H2

FZK Exp. 16g H2

0.1 1 10 100 1000




2g H2

1 m

glass breakage

Area 1m2

4.7 mm thick

Area 3m2

4.7 mm thick

16g H2

2 m

10000+ [ms]

Δ+

103

104

105

106

Δp+I+

t

T+=2I+/Δp+

T+

Δp+I+

t

T+=2I+/Δp+

T+


1,4m

1,4m

2,4m

2,4m

2,8m

5.6m

4,0m

6,3m4,0m

FZK Exp. 2g H2

FZK Exp. 16g H2

FZK exp. 2g H2

FZK exp. 16g H 2

0.1 1 10 100 1000




2g H2

1 m

glass breakage

Area 1m2

4.7 mm thick

Area 3m2

4.7 mm thick

16g H2

2 m

5 m5 m

450 Pas

300 Pas110 PasResults: - windows and light garage components (door) wold break

- damage to masonary walls only from nearly explosion of 16 g H2- wooden framework construction would be destroyed


0.1 1 10 100 1000 10000

NASA

16g H2

Pos

itive

pea

k ov

erpr

essu

re Δ

p+[P

a]

103

104

105

106

lung rupture thresholdΔp+

I+

t

T+=2I+/Δp+

T+


1,4m

1,4m

2,4m

2,4m

2,8m

5.6m

4,0m

6,3m4,0m

FZK Exp. 2g H2

FZK Exp. 16g H2

increasinggrade ofinjury


50 % eardrum rupture

1 % eardrum rupture

Baker

Duration of positive overpressure T+ [ms]

1 m

2 m

2g H25 m

0.1 1 10 100 1000 10000

NASA

16g H2

Pos

itive

pea

k ov

erpr

essu

re Δ

p+[P

a]

103

104

105

106

lung rupture thresholdΔp+

I+

t

T+=2I+/Δp+

T+


1,4m

1,4m

2,4m

2,4m

2,8m

5.6m

4,0m

6,3m4,0m

FZK Exp. 2g H2

FZK Exp. 16g H2

FZK Exp. 2g H2

FZK Exp. 16g H2



50 % eardrum rupture

1 % eardrum rupture

Baker

Duration of positive overpressure T+ [ms]

1 m1 m

2 m2 m

2g H25 m5 m

HUMAN EFFECTS FROM CONFINED LOCAL H2 – EXPLOSIONS IN GARAGE

Results: - high probability of ear drum rupture- no long damage


SUMMARY OF MECHANISTIC SAFETY ANALYSIS OF

HYDROGEN BASED ENERGY SYSTEMS


PROCEDURE FOR HYDROGEN SAFETY ANALYSIS

• A complete, self-consistent and mechanistic analysis procedure has been developed whichaddresses all important physical phenomena of hydrogen behaviour in accidental release scenarios

CONSEQUENCEANALYSIS

Mechanical andthermal loads

Structural

Human effects

CRITERIA FORHAZARD

POTENTIAL

Flammability

y

y

Flame Acceleration

y

COMBUSTIBLEMIXTURE

GENERATION

GASFLOW

FLAME3D

Fast turbulentdeflagration

COM3D

DetonationDET3D

Slow deflagration

CONSEQUENCEANALYSIS

responseSDO

ABAQUS

Detonationtransition

COMBUSTIBLEMIXTURE

GENERATION

Problem geometry

Mitigation

Scenario

Sources

DistributionGP - Program

COMBUSTIONSIMULATION


MITIGATION MEASURES• The proposed analysis procedure allows identification of possible mitigation measures

for risk reduction

• Exlude severe scenarios by design changes

• Limit hydrogen sources

• Support hydrogen dispersion and mixing processes

• Exclude ignition sources

• Suppress flame acceleration(low confinement and turbulence generation)

• Avoid detonation transition processes(lean mixtures, small scale)

• Confine consequences(strong enclosure)

Accidentprogression

If one level of defencehas beenoptimized, workon next barrierfor accidentprogression

mechanistic safety analysis of hydrogen based … · mechanistic safety analysis of hydrogen based...

Documents