1

Statistical Analysis of Electrostatic Spark Ignition of Lean H2 -O2 -Ar Mixtures

Statistical Analysis of Electrostatic Spark Ignition of Lean H2 -O2 -Ar Mixtures

Sally P. M. Bane1, Joseph E. Shepherd1

Eddie Kwon2, Art C. Day2

1California Institute of Technology, Pasadena, CA2Boeing Research & Technology, Seattle, WA

3rd

International Conference on Hydrogen SafetyAjaccio, Corsica, FranceSeptember 16-18, 2009

2

Spark Ignition & Minimum Ignition Energy

•

determining risk of accidental ignition extremely important in industry and aviation safety

•

Minimum Ignition Energy (MIE) –

traditional basis for quantifying ignition hazards

•

experimental work using capacitive spark, tabulations of MIE values

Percent Hydrogen

Min

imum

Igni

tion

Ener

gy

MIE curves, Lewis and von Elbe, 1961

0 5 10 15 20 250

0.02

0.04

0.06

0.08

0.1

0.12

Test Number

Spa

rk E

nerg

y (J

)

GONO GO

Jet A ignition test data, Lee and Shepherd, 1999

•

New viewpoint –

ignition as statistical

phenomenon

•

previous statistical analysis of Jet A ignition, hot surface ignition

•

little work done on statistics of ignition of hydrogen

Data overlap region

3

Statistical View of Ignition

GOAL: isolate and examine statistical nature of ignition with respect to spark energy

Must quantify and minimize other sources of experimental variability

1) uncertainties in mixture composition

2) ignition detection method

3) turbulence

4) spark energy measurement

Small changes in composition lead to large change in combustion characteristics, MIE

false positives or negatives

effect on spark channel formation, flame propagation

SOME SOURCES OF VARIABILITY:

Previous work done to assess these sources,

improve experimental

design

Objective: examine statistical nature of lean hydrogen aviation test mixtures

4

Statistical Analysis of Ignition Test Data

Goal: probability distribution for ignition versus stimulus level (spark energy)

logistic distribution –

often used to analyze binary “failure”

data

Example: Jet A spark ignition (Lee and Shepherd, 1999)

0.0

0.3

0.5

0.8

1.0

0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16

Spark Energy (mJ)

Prob

abili

ty o

f Ign

ition

Logistic probability distribution

95% confidence envelope

50th

percentile (50% probability of

ignition)

test data points

5

Short, Fixed Spark Ignition Testing: Spark Ignition System

SS relay (opens

HV relay)

variable vacuum capacitor (~3-30 pF)

50 G

resistors

HV relay

current transformer

HV power supply input

acrylic tube filled

with dry air conical tungsten electrode tips

fixed length spark gap

6

Short, Fixed Spark Ignition Testing: Estimating Spark Energy

...21

22

RiEMRshockresidualthermalstored EEEEECVE

sparkE estimate neglect

residualstoredspark EEE

2

21

breakdownstored CVE

CQE residual

residual

2

21

dttiCVQQQ sparkbreakdownsparkstoredresidual

high-speed current transformer

Calculate the charge left in the capacitor

using the spark current

7

Experimental Setup

static pressure

gauge

fan mixer

gas fill line

vacuum line

thermocouple

dynamic pressure

transducer

spark circuit

spark gap

window for schlieren

Composition control

Reliable ignition detection

Turbulence

• vacuum chamber

• fill by partial pressures

•

static pressure measurement with 0.01 kPa

precision

• pressure transducer

• thermocouple

•

high-speed schlieren

visualization

• fan mixer

• wait time after turn-off

8

Flame Visualization

high-speed flame visualization using schlieren

optics

high-speed camera (1000 fps)

5% hydrogen test mixture (ARP), 2 other mixtures with 1% more H2

→

5% H2

–12% O2

–83% Ar, 6% H2

–12% O2

–82% Ar, 7% H2

–12% O2

–81% Ar

5% H2 6% H2 7% H2

9

Pressure Measurement

1.0

1.5

2.0

2.5

3.0

3.5

4.0

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0

time (s)

pres

sure

(bar

)

7% H2

6% H2

5% H2

1.0

1.5

2.0

2.5

3.0

3.5

4.0

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0

time (s)

pres

sure

(bar

)

7% H2

6% H2

5% H2

10

Short, Fixed Spark Ignition Testing: Ignition Probability

ignition tests in three H2

test mixtures

fixed spark gap length (1-2 mm), range of spark energies (vary capacitance)

0

0.2

0.4

0.6

0.8

1

0 300 600 900 1200 1500

Spark Energy (J)

Prob

abili

ty o

f Ign

ition Probability

95% ConfidenceData Points

Ignition –

a “go”

Binary Result = 1

No Ignition –

a “no go”

Binary Result = 0

5% H2 Mixture

11

Short, Fixed Spark Ignition Testing: Ignition Probability (cont.)

0

0.2

0.4

0.6

0.8

1

0 200 400 600 800 1000 1200 1400

Spark Energy (J)

Prob

abili

ty o

f Ign

ition

7% H2 6% H2 5% H2

Data Overlap Region

0

0.2

0.4

0.6

0.8

1

0 200 400 600 800 1000 1200 1400

Spark Energy (J)

Prob

abili

ty o

f Ign

ition

0

0.2

0.4

0.6

0.8

1

0 200 400 600 800 1000 1200 1400

Spark Energy (J)

Prob

abili

ty o

f Ign

ition

7% H2 6% H2 5% H2

Data Overlap Region

12

Short, Fixed Spark Ignition Testing: Ignition Probability (cont.)

0

0.2

0.4

0.6

0.8

1

0 200 400 600 800 1000 1200 1400

Spark Energy (J)

Prob

abili

ty o

f Ign

ition

Prob

abili

ty o

f Ign

ition

Comparison with original MIE data (Lewis & von Elbe 1961):

952 J

(50% probability of ignition) determined from experiments

MIE for 5% H2

Mixture ~ 200 J

(from extrapolation)

5% H27% H2

MIE for 7% H2

Mixture = 100 J

13

Long, Variable Spark Ignition TestingQUESTION: In addition to the spark energy, is the spark length important too?

Is spark energy density (spark energy/spark length) a more appropriate parameter?

Teflon tube

Isolated capacitor plate, charged (-)

Charged electrode

Movable grounded electrode

Motor-driven linear stage

Movable electrode

Current transformer

developed ignition system to vary both spark energy and spark length

14

Long Spark Ignition Testing: Results (cont.)

6 mm

3 mm

NO IGNITIONSpark Energy: 1000 J

Spark Length: 6 mm

Energy Density = 167 J/mm

IGNITIONSpark Energy: 740 J

Spark Length: 3 mm

Energy Density = 247 J/mm

0

1

0 500 1000 1500 2000Spark Energy (J)

Test

Res

ult

Long Spark Data Points

MIE for Fixed 1.5mm Spark

Median Energy for Long Spark

351 J at 1.5 mm 741 J at 3.0 mm

0

1

0 500 1000 1500 2000Spark Energy (J)

Test

Res

ult

Long Spark Data Points

MIE for Fixed 1.5mm Spark

Median Energy for Long Spark

351 J at 1.5 mm 741 J at 3.0 mm

Test

Res

ult

Spark Energy (J)

15

Long Spark Ignition Testing: Results

0

0.2

0.4

0.6

0.8

1

0 100 200 300 400 500

Energy Density (J/mm)

Prob

abili

ty

95% ConfidenceData PointsProbabilityFixed 1.5mm Spark

50% probability of ignition using fixed 1.5 mm spark

0

0.2

0.4

0.6

0.8

1

0 100 200 300 400 500

Energy Density (J/mm)

Prob

abili

ty

95% ConfidenceData PointsProbabilityFixed 1.5mm Spark

50% probability of ignition using fixed 1.5 mm spark

second set of tests in 6% H2

mixture

vary both spark energy and length →

range of energy densities

50% Probability of Ignition:

Short, Fixed Length Sparks:

351 J

at 1.5 mm

234 J/mm

Longer, Variable Length Sparks:

154 J/mm

reduced effect of quenching

instability of spark channel

16

Conclusions

developed low-energy capacitive spark system to produce short, fixed length sparks

used to perform ignition tests in 3 hydrogen-based aviation test mixtures

analyzed statistically –

probability distributions for ignition vs. spark energy

• results statistical in nature –

contradict traditional MIE view

• small change in composition → large change in flame propagation, MIE

second spark ignition system –

sparks of variable energy and lengths

ignition tests in 6% H2

mixture, varying spark energy density

probability distributions for ignition vs. spark energy density

• longer sparks = more energy to ignite

• spark length must be considered, NOT just energy

• spark energy density is more appropriate parameter