Hydrogen Behavior in Containment

Calvin K. ChanExperimental Thermalhydraulics and Combustion Branch

Presented to US Nuclear Regulatory CommissionWashington DC

May 6-7, 2003

Pg 2

Outline

• Phenomenon and Definition• Limited and Severe Core Damage Accidents• Hydrogen Source Terms• R&D Programs on Hydrogen Behavior

− Hydrogen Test Facilities at AECL− Hydrogen Distribution (as attachment)− Hydrogen Combustion (as attachment)− Hydrogen Mitigation (as attachment)− Analysis Tools

• Summary

Pg 3

Hydrogen Behavior During Limited and Severe Core Damage Accidents

• During the limited or severe core damages accidents in CANDU reactors, hydrogen can be formed by zirconium-steam reactions, radiolysis and corrosion of metal. Hydrogen may migrate to the containment building creating a combustible atmosphere. Thermal and mechanical loads to containment structures resulting from ignition of the hydrogen are safety concerns.

Pg 4

Types of Accidents• Limited Core Damage Accidents

(e.g., LOCA + LOECC)− Cause fuel damage in a single or multiple channels − Reactor core is not severely damaged because moderator provides

a back-up to remove decay heat− Hydrogen is formed by Zirconium-steam reactions and eventually

accumulates in containment

• Severe Core Damage Accidents(e.g., LOCA + LOECC + unavailability of moderator heat sink)− Cause reactor core damage− While there can be loss of channel geometry, water in the calandria

vessel and shield tank prevents core expulsion− Larger hydrogen source term− An extremely improbable event in CANDU reactor

Pg 5

Examples of Events

• Limited Core Damage Accidents− Small LOCA + LOECC− End Fitting Failure + LOECC− Large LOCA + LOECC− Pressure Tube / Calandria Tube Failure + LOECC− Steam Generator Tube Rupture + LOECC− Stagnation Feeder Break− Severe Channel Flow Blockage

• Severe Core Damage Accidents− LOCA + LOECC + unavailability of moderator heat sink

Pg 6

Hydrogen Source Term

• Hydrogen release in containment− Short term (< 1 day)

• Zirconium-steam reactions• Hydrogen degassing

− Long term (> 1 day)• Radiolysis• Corrosion of metal

Pg 7

Typical Hydrogen Source Terms for LOCA + LOECC Accidents

(Typical hydrogen source term for CANDU 6)

• Zirconium-steam reactions− 65 kg of hydrogen (1500 m3)

• Hydrogen degassing− 0.3 kg of hydrogen (3.5 m3)

• Radiolysis− 30 kg/day from moderator water− 7 kg/day from sump water− 0.7 kg/day from moisture− Total 37.7 kg /day (890 m3/day)

• Corrosion of Aluminum− Maximum 40 kg after 5 days

Pg 8

Hydrogen Mitigation Strategies

• Gas Mixing - (distribution of hydrogen)− Natural convection− Local air coolers

• Passive Autocatalytic Recombiner, PAR - (removal of hydrogen)

Pg 9

Hydrogen Mitigation

• Containment design promotes hydrogen mixing and distribution by natural convection

• Local air coolers prevent local stratification of hydrogen

• Hydrogen catalytic recombiners remove hydrogen− Self start at ~2% H2− Passive device− Effective for long term hydrogen management

Pg 10

Impact of Hydrogen Burns in Containment

• Standing flames − H2 diffusion flame at the break created by auto-ignition or a

flash back− Thermal load

• Hydrogen explosion (deflagration)− H2 ignited accidentally by hot surfaces or electrical sparks− Static and dynamic pressures

• Detonation − Via a Deflagration to Detonation Transition− Blast waves

Pg 11

Understanding Hydrogen Behavior

• Based on an extensive research program in the past 30 years in studying hydrogen behavior for existing CANDU reactors, an understanding of the key phenomena has been achieved

• Computer codes have also been developed to analyze hydrogen behavior for various accident scenarios

• These tools can be used to analyze postulated accidents in ACR

Pg 12

CANDU ContainmentACR CANDU 6

Pg 13

R&D Program on Hydrogen Behavior at AECL

• AECL has a comprehensive R&D program to acquire understanding of key phenomena and to develop tools for accident analysis since early 1970s. Areas of research include:

− Hydrogen Distribution− Hydrogen Combustion− Hydrogen Mitigation− Containment Response

Pg 14

Goals of R&D Programs

• Acquire fundamental understanding of key combustion phenomena relevant to postulated CANDU reactor accident scenarios

• Develop computer models for predicting gas distribution and combustion pressure

Pg 15

Hydrogen Programs at AECL

• Recent research programs include:− Gas distribution under accident conditions− Mechanisms and dynamics of standing flames − Mechanisms and dynamics of vented combustion − Flame acceleration and transition to detonation− Dynamics of flame jet ignition − Burn model development and validation

Pg 16

Research Facilities for Hydrogen

• Large Scale Vented Combustion Test Facility• Containment Test Facility• Diffusion Flame Facility• Large Scale Gas Mixing Facility

Pg 17

Large Scale Vented Combustion Test Facility (LSVCTF)

The Large-Scale Vented Combustion Test Facility (LSVCTF) is a 10-m long, 4-m wide, 3-m high rectangular enclosure with an internal volume of 120 m3. The test chamber, including the end walls, is electrically trace-heated and heavily insulated to maintain temperatures in excess of 100oC for extended periods of time. The combustion chamber can be subdivided into 2 or 3 compartments. Variable sizes of vent openings are available between compartments and to the outside.

Pg 18

Large Scale Vented Combustion Test Facility

Pg 19

Diffusion Flame Facility (DFF)

The Diffusion Flame Facility (DFF) consists of a burner with associated gas supply lines and instrumentation housed within a modified grain silo (5 m diameter and 8 m height), which is insulated to retain heat for experiments that involve an air / steam environment. Tests with H2 / steam jet flames (up to 15 cm in diameter) in air / steam atmosphere (up to 30% steam by volume) can be performed in this facility.

Pg 20

Hydrogen Diffusion Flame

Pg 21

Containment Test Facility (CTF)

The Containment Test Facility (CTF) consists of a 6-m3 sphere and a 10-m3 cylinder, both rated for pressures up to 10 MPa and trace-heated for operation at temperatures up to 150oC. The large vessels may be inter-connected by 30 cm and 50 cm diameter ducts. The CTF is designed to investigate the fundamentals of combustion phenomena. These include flammability limits, ignition, turbulent combustion, flame acceleration, detonation, detonation transition.

Pg 22

Containment Test Facility

A 28cm-diameter and 9m-long combustion pipe with a design pressure of 10MPa. Obstacles can be mounted inside this pipe to induce flame acceleration. This apparatus has been used to determine the run-up distances for supersonic flames and DDT.

Pg 23

Large Scale Gas Mixing Facility (LSGMF) at Whiteshell Laboratories

AECL’s Large-Scale Gas Mixing Facility (LSGMF) is a 10.3 m by 11.0 m by 8.2 m concrete enclosure with an internal volume of approximately 1000 m3. Helium and steam can be injected into the enclosure at various locations to simulate a break in the primary cooling system inside a reactor containment building following an accident. Internal partitions can also be added to the facility to simulate sub-compartments inside a reactor building.

A new facility is being constructed at AECL’s Chalk River Laboratories (CRL)

Pg 24

AECL Passive Autocatalytic Recombiner

• recombine hydrogen with oxygen in a controlled fashion

• based on AECL’s wet-proofed catalyst technology

• have been qualified with tests in the large-scale vented combustion facility

Pg 25

R&D Programs on Hydrogen

• Details of selected R&D programs are included as attachments to this presentation

−Hydrogen distribution−Hydrogen diffusion flames−Vented combustion in a complex geometry−Flame jet ignition−AECL PAR

• Analysis tools and analysis strategy are discussed in this presentation

Pg 26

Analysis ToolsGOTHIC (mechanistic approach)• GOTHIC (Generation of Thermal-Hydraulic Information for Containment)

is a general purpose computer code for thermal hydraulic and combustion calculations (in 3-dimensional or hybrid mode)

• GOTHIC, with addition of CANDU-specific models for hydrogen behavior, is used to model containment thermal hydraulics and hydrogen transport

• It also calculates the combustion pressure in the event of an ignition• It cannot predict supersonic flames and DDT

DDTINDEX (upper-bound approach)• Calculate a set of parameters for assessing the possibility of flame

acceleration to supersonic velocities and subsequently trigger atransition to detonation

Pg 27

Potential Hazards for Supersonic Flames and Detonations

• The hydrogen cloud can be ignited accidentally creating a gas explosion inside the containment building

• An expanding flame is intrinsically unstable• In the presence of obstacles, a flame can potentially

accelerate to supersonic velocities and subsequently lead to Deflagration to Detonation Transition (DDT)

• Impact of pressure waves associated with supersonic flames (>500 kPa) and detonations (>1500 kPa) need to be understood

Pg 28

Assessing Potential for Supersonic Flames and DDT

• Criteria (necessary conditions in the form of a set of non-dimensional parameters) for flame acceleration to supersonic velocities and DDT have been experimentally determined

• If these criteria are not met, supersonic flames and detonations can be ruled out

• DDTINDEX calculates these parameters based on gas distribution information predicted by GOTHIC

Pg 29

Experimental Data for DDT

100 1000 10000Geometrical size D, mm

0

4

8

12

16

20

D/λ

Detonation

Deflagration

D/ λ = 7

Summary of DDT Conditions in D/λλλλ, where D is the Characteristic Dimension of a Room and λλλλ is the detonation cell size

Source: OECD Nuclear Energy Agency State-of-the-Art Report on Flame Acceleration and DDT in Nuclear Safety (AECL is one of the contributor to this report)

Pg 30

Criteria for DDT (the λλλλ criteria)

Transition to detonation requires L > 7λλλλ,

where L is the cloud size

λλλλ is the detonation cell width

DDT Index, φφφφDDT=L/L*

DDT is not possible for φφφφDDT < 1

where L*= 7λλλλ

Pg 31

Experimental Data for Flame Acceleration

Summary of flame acceleration conditions in σσσσ, where σσσσ is the expansion ratio

Source: OECD NEA State-of-the-Art Report on Flame Acceleration and DDT in Nuclear Safety

Pg 32

Flame Acceleration Index δδδδ = σσσσ/σσσσ*where σσσσ is the expansion ratio (ρρρρu/ρρρρb)

Flame acceleration to supersonic velocitiesis impossible if δδδδ<1

σσσσ* = 3.75 for XH2 = or >2XO2 , X is the gas mole fraction

σσσσ* = 3.75 - 0.0115(T-25) + 0.00002(T-25)2 for XH2 < 2XO2

where T is average temperature of the gas cloud in oC

Criteria for Flame Acceleration to Supersonic Velocities (the σσσσ criteria)

Pg 33

DDT Potential (a second criterion for DDT)

Since DDT cannot occur if the cloud size is less than the minimum run-up distance for DDT, a “DDT Potential”can also be defined as a second criterion for DDT.

DDT Potential, γγγγDDT = L/LDDT,

where L is the nominal diameter of thecombustible gas cloud and LDDT is the minimumrun-up distance for DDT.

DDT is not possible for γγγγDDT <1.

Pg 34

Flame Acceleration (FA) Potential (a second criterion for SF)

Since a flame requires physical distance (run-up distance) to accelerate to supersonic velocities, a “FA Potential” can also be defined as a second criterion.

The FA Potential, φφφφFA = L/LSF,

where L is the nominal cloud size and LSF is the run-up distance for a supersonicflame.

Flame acceleration to supersonic velocity is not possible for φφφφFA <1.

Pg 35

DDTINDEX

• DDTINDEX calculates DDT Index, DDT Potential, FA Index and FA Potential for a non-uniform combustible gas cloud (predicted by GOTHIC)

• If all these parameters are less than 1, both supersonic flame and DDT can be ruled out

• The use of DDT Index and FA Index has been recognized internationally (OECD NEA State-of-the-Art Report) as a sound methodology for assessing the impact of hydrogen inside containment

Pg 36

GOTHIC Simulation

Accident Scenario (hypothetical):100% Header break infueling machine vault

Header

Fuelling machine vault

Pg 37

• Hydrogen distribution in containment during a large LOCA (header break)

GOTHIC Simulation

Animations

Pg 38

DDTINDEX Outputs

Animation

Pg 39

Strategy for Evaluating Hydrogen Hazards

σσσσ Criteria

Scenario

Sources

Distribution

Flammable Mixture?

Flame Acceleration?

DDT?

Slow deflagration

Fast flame

Local Detonation

Consequences•Mechanical and thermal loads•Equipment Qualification•Structural response•Fission product interaction•Secondary events (e.g. cable fires, etc.)

λλλλ Criteria

FA Index < 1

DDT Index < 1

Standing flameAnalysis(3D GOTHIC)

Analysis(3D GOTHIC)

Recombiner

Analysis(DDTINDEX) FA Potential < 1

DDT Potential < 1

No Impact

Engineering correlations

Pg 40

Summary• AECL has a mature and widely-recognized hydrogen

program• The dynamics and mechanisms associated with

hydrogen combustion behavior in CANDU containmenthave been understood

• Models to capture the key phenomena have been developed and validated

• AECL has developed an accepted mitigation device (PAR) for CANDU and other reactor designs

• Knowledge base acquired in the past can be used to analyze reactor accidents relevant to ACR

Pg 41

Pg 42

Research Activities on Hydrogen Behavior at AECL

Hydrogen Distribution• Buoyancy driven mixing• Gas mixing in a partitioned enclosure• Containment code for gas

distribution analysis (GOTHIC)

Pg 43

CANDU Containment

Dome region

Fuelling machine vault

Reactor

Possible H2release paths

Pg 44

Large Scale Gas Mixing Facility (WL) - Interior

Pg 45

Buoyancy Driven Gas Mixing

0 100 200 300 400 500 6000

2

4

6

8

10

P2

P3

P4

P5

0 100 200 300 400 500 600

0

10

20

30

40

50

60H

eliu

m C

once

ntra

tion,

%

P1

P2

P3

P4

P5

Time, s

0.6m from helium jet

3.2m from helium jet

5.5m from helium jet

Pg 46

Buoyancy Induced Convection in Partitioned Volumes

Experimental facility• a 2.44 m x 2.44 m x 6.10 m enclosure • Helium injection at bottom• Helium concentration measurement• Velocity measurement at inlet and outlet

Hel

ium

Jet

x

x

x

x

x

x 1.20 m

2.40 m

0.60 m

2.38

m1.

29 m

1.16

m

0.90 m0.30 m

0.30 m

0.43 m0.48 m

0.30 m

1.21

m

0.54 m

0.30 m

0.07 m

Helium Probe #1

Helium Probe #2x

Helium Probe #3

Helium Probe #4

Helium Probe #5

Velocity Probe

Velocity Probe

Offset Helium ProbeLocations

0.60

m

Pg 47

Buoyancy Induced Convection in Partitioned Volumes (cont.)

0 200 400 600 800 1000 12000.00.20.40.60.81.01.21.41.61.82.02.22.42.62.83.0

0.47 g/s

0.91 g/s

1.39 g/s

1.83 g/s

Inle

t Vel

ocity

, m/s

Time, sec

Inlet Velocity

Pg 48

Buoyancy Induced Convection in Partitioned Volumes (cont.)

0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.00

10

20

30

Volu

me

Flow

rate

Mag

nific

atio

n Fa

ctor

Mass Flow Rate g/s0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0

0

20

40

60

80

100

120

140

160

180

200

Mas

s Fl

owra

te M

agni

ficat

ion

Fact

or

Mass Flow Rate, g/s

Pg 49

Hydrogen Diffusion Flame Facility

Pg 50

Hydrogen Diffusion Flame

0 100 200 300 400 500 600 700 800 9000

200

400

600

800

1000

1200

1400

1600

1800

XGas T/C from nozzle: 5 mmMin. XFlame T/C from nozzle: 38 cmFlame T/C's: 4 cm apart

Dry H2 GasUpstream Pressure: 800 kPaNozzle dia.: 6 mm

Oven Temp. H2 Gas Inlet Temp. H2 Gas Outlet Temp. Flame Temp. #4 Flame Temp. #5 Flame Temp. #6 Flame Temp. #7

Tem

pera

ture

, o C

Time, sec

Flame Temperature of a Dry Hydrogen Jet Preheated to 500C

Pre-heaterBurner

Hydrogen

Pg 51

Large Scale Vented Combustion Test Facility

Pg 52

LSVCTF: Combustion in Interconnected Chambers

Specific Tests Series• Variation in H2% Between Chambers• Variation of Vent Area• Variation of Vent Location • Variation of ignition location

ignition

Inter-connected chambers exist in containment

Pg 53

Combustion Pressure

0 1 2 3 4 5 6 7 8 9 10 11 120

10

20

30

40

50

10% H 2 in Rear Chamber

Rear Chamber

Front Chamber

Pres

sure

(kPa

)

H 2 in Front Chamber (vol% )

7 8 9 10 110

10

20

30

40

50

Rear Chamber

Front Chamber

Pres

sure

(kPa

)H2 in rear and front chambers (vol%)

Unequal H2% (C1 > C2) Equal H2% (C1 = C2)

C2 C1

Igniter

Pg 54

Variation of Vent Area

0 1 2 30

10

20

30

40

50

60

70

80

10% H2 in Front and Rear Chambers

Inner Vent Area = 0.38m2

Rear Chamber

Front Chamber

Pres

sure

(kPa

)

External Vent Area (m2)

Pressure

Vent area

0.38 m2

Pg 55

Containment Test Facility

Pg 56

Flame-Jet Ignition Tests

Top View of Flange

Orifice Plate

Igniter Spark

( Channel 4) (PCB)

(Channel 3) (Kulite)

(Channel 1) (Kulite )

(Channel 2) (PCB)

Pg 57

Single-Flame-Jet Ignition Tests

7 8 9 10 11 12 13 14 150

1

2

3

4

5

6

EBVvsH3.opj

Effe

ctiv

e B

urni

ng V

eloc

ity (m

/s)

Hydrogen Concentration (%)

Pg 58

Single-Flame-Jet Ignition Tests (cont.)

7 8 9 10 11 12 13 14 150

2

4

6

8

10

12

BVRvsH9.opj

Enha

ncem

ent F

acto

r (S j/S

l)

Hydrogen Concentration (%)

Pg 59

Hydrogen Mitigation by Catalytic Recombiner

• Passive Autocatalytic Recombiner (PAR)− Passive device− Self-start at 2% hydrogen at 20C− Wet-proofed catalyst− Suitable long-term hydrogen management

Pg 60

Continuous Hydrogen Injection

Test conditions:

25°C

100% relative humidity

0.029 kg/h injection rate

Injection at centre bottom of side wall

60 120 180 240 300 3600.0

0.5

1.0

1.5

2.0Temperatures

Catalyst Plate Recombiner inlet

Hydrogen At ceiling Mid-height of chamber Recombiner inlet Recombiner outlet

Hyd

roge

n Co

ncen

tratio

n (v

ol%

)

Time (min)

0

50

100

150

200

Temperature ( oC)

Pg 61