Test Methods and Apparatus for a Video-based System for Cargo Fire Verification I. Willms, W. Krüll, J. Shirer 1 Fourth Triennial International Fire and

Test Methods and Apparatus for a Video-based System for Cargo Fire Verification

I. Willms, W. Krüll, J. Shirer1

Fourth Triennial International Fire and Cabin Fourth Triennial International Fire and Cabin Safety Research ConferenceSafety Research Conference

Test Methods and Apparatus for a Video-based

System designed for Cargo Fire Verification Ingolf Willms, Wolfgang Krüll

Department of Communication Systems University Duisburg-Essen

Jeff Shirer

Goodrich Corporation

Vergennes, Vermont, U.S.A.

N T S



1 Motivation

2 Testing aspects of video-based systems

3 Fire and smoke tests

4 Dust tests

5 Summary and outlook



1 Motivation for a video-based Cargo Fire Verification System

• In case of smoke alarm: Crew must activate the extinguishing

system and land at the nearest suitable airport

• Pilot has no capability to verify if the alarm is real or false

False alarms of smoke detectors in cargo bays during long

range flights of passenger airplanes cause various problems:

• Diversion to a small airport may cause significant danger

• Consequences: High costs due to flight diversions,

landing costs, possible loss of the cargo load




Possible reasons for false alarms of traditional smoke detectors in cargo

compartments:

• Mist, dust and oil particles

• Additional problems are environmental conditions

(e.g. temperature variation, air pressure variation)

Objectives of the Cargo Fire Verification System:

• To provide the aircrew with images of the conditions in the cargo bay

• To detect fire earlier than conventional smoke detectors

• Thus to verify it or not

• Thus to greatly reduce false alarms



• not easily accessible during flight

• a difficult area to view with video cameras

The gap between the cargo and cargo bay ceiling can be only 4.3 cm

Cargo compartments are




MFS Multi-Function Sensor

ISM Illumination Sensor Module

FWD AFT BULK

Example of camera and illumination placement in a cargo bay of an

AIRBUS A340-500: FWD (in front of the aircraft)

AFT, BULK (behind the wing boxes)

For reference: The size of an A340-500 AFT compartment is

10.4 m long * 4.2 m wide * 1.7 m high




1 Motivation - Qualification Testing of traditional detection

systems

• Existing testing guidelines & standards for smoke detection systems

were formulated for traditional sensor technologies (e.g. optical-

based)

• EN-54 tests use different types of smoke sources at the centre of the

floor in a fire lab

• Essential measurements: Reference smoke density measurements

at the location of the detector & detector alarm time

• The geometry and contents of the test chamber are not critical to

smoke detectors if the smoke density increases within EN-54 specs



1 Motivation – Requirements for the test of video fire detection systems (CFVS)

• Similar procedures as in EN54 are reasonable (as far as possible)

• Comparison with EN54 qualified conventional detectors are needed

• Suitable stimulation of video sensors in field of view required

• No down scaling possibilities concerning room size

• No tests in a smoke duct are possible

• Increasing stimulus over time (as in EN54)

• Consequences:

New testing details needed

Modifications of EN54 are needed



Four views with different illumination, same geometric configuration

2 Testing aspects of video-based systems – Some CFVS images



2 Testing aspects of video-based systems

For a video-based detection system, scene geometry and image

contents are critical

• The scene should closely represent the actual application

• Usage of 1:1 scale – no test chamber

• The fire test room should resemble an actual cargo bay

• This includes cargo obstructions in cameras field of view

• Tests should retain the repeatability aspects of the EN-54 tests



2 Testing aspects - Possible candidates for testing of the CFVS

Special testing aircraft

• Many disadvantages:

- High costs, limited test times

- Restricted test options

(no open fires are permitted)

A reconstruction of a cargo compartment

• Main drawbacks (Trauen mock-up):

- Limited control of environm. conditions

- High construction effort

- Large required space

- Hard to change

- No comparative EN-54 tests are possible



2 Testing aspects - Possible candidates for testing of the CFVS

A modified EN-54 test room (Duisburg fire lab)

• EN-54 like test cell

• Option of changing the test cell height within minutes (2,90 …> 4 m)

Advantages of the fire lab of Duisburg University:

•Controlled environmental conditions

•EN 54 measurement conditions

•A wide range of measurement possibilities

•Good probability for “repeatable” tests



3 Fire and smoke tests - Performed in the A340-500 cargo bay

mock-up in Trauen and in the Duisburg fire lab

In the Duisburg Fire Detection Lab tests were performed in three major

types of scenarios:

• Standard EN-54 test fires

• Modified test fires without test cell modifications,

but with camera obstructions by containers

• Modified test fires with test cell modifications

(for Forward, Aft, and Bulk compartment configurations,

including camera obstructions by containers)

Test cell modifications for closely resembling an actual cargo

compartment

Fire material reduction to account for the reduced test cell volume



3 Fire and smoke tests – Performed in Duisburg and in Trauen

Modified EN-54 A340-500 cargo bay test cell in Duisburg mock-up in Trauen (without side walls)Volume: 293 m³ Volume: 103 m³ 10.5 m * 9.0 m * 3.1 m 15.0 m * 4.2 m * 1.71 m



0.01.02.03.04.05.06.0

240 300 360 420 480 540 600 660t [sec]

m [d

B/m

]modTF2_0607 * 3Trauen, 1_2Trauen, 1_3

Duisburg, modTF2_0607without side walls

• Standard EN-54 test fire material was reduced to account for

the reduced test cell volume

• Room geometries of both test cells are completely different

3 Fire and smoke tests – Performed in Duisburg and Trauen

3

Scalingfactor

• Results are comparable by using a simple scaling factor

to approximate the results from Trauen

MIR

EX

val

ues



MIC

MIREX

Camera

ISMSmoke detectorsFire

Test cell modifications:

• Aluminium side wall plates

were hung from the ceiling

to reduce the room width

• Fire and non-fire sources

were placed on a platform

• The distance between

the source and the fire

room ceiling then was 1.7 m

3 Fire and smoke tests - Test cell modifications



T

3 Fire and smoke tests - Test cell modifications

• Side walls were covered with

DuPont Tedlar PVF Film (as in

real cargo bay) for improved

realism

• Single 1 m² sheet metal plates

(also covered with the cargo

bay lining material) was

mounted on the ceiling for

further improved realism



0.0

1.0

2.0

3.0

4.0

5.0

6.0

240 300 360 420 480 540 600 660t [sec]

m [

dB

/m]

modTF2_0617modTF2_0614

modTF2_0614 * 1.5modTF2_0617 * 1.5Trauen, 1_2

0.01.02.03.04.05.06.0

240 300 360 420 480 540 600 660t [sec]

m [

dB/m

]

modTF2_0607 * 3Trauen, 1_2Trauen, 1_3

Duisburg, modTF2_0607

3

1.5

Scalingfactor

without side walls

with side wallswith side wallssm

oke

den

sity

via

th

e M

IRE

X



CFVS performance compared to conventional A340 detectors

• The CFVS was designed as a verification system, intended to

confirm or unconfirm alarms issued by the primary system.

• The CFVS has to detect fires earlier than conventional smoke

detectors and to confirm the primary system’s alarm.



4 Dust tests

• Many false alarms in cargo bays are caused by fog and dust

- Dust lifted from dirty containers or

- Perhaps produced by cargo content such as animals

• Dust is known to cause problems with optical smoke detectors -

and not only in cargo bays.

• Depending on the dust type and the dust production

it may look similar to smoke on video.

• Discrimination of dust and smoke in video-based systems thus is a

challenge in its own.

• Density of dust was chosen such that it caused the conventional

smoke detectors to alarm.



A B C D E F G H I K L M N

1

10

8

7

6

54

3

2

9

11

12

13

14

15

16

17

R 3.0m.

MIREX MIC

FWD

Observation Rooms

T

1.8 m.

0.92m.

LD-3LD-3

30°

DustGen

1.80 m

0.60 m

Particle Counter Inlet

4 Dust tests - Set-up for Bulk bay dust tests



4 Dust tests - Smoke detector alarm times

Different cargo compartment types (Aft bay / Bulk bay)

Different dust types (dust A1 / dust A2)

10

20

30

40

50

60

70

0 1 2 3 4 5 6 7 Samples

Tim

e af

ter

dust

ge

nera

tion

sta

rt [

sec.

] Aft Bay - A1

Aft Bay - A2

Bulk Bay - A1

Bulk Bay - A2Average



4 Dust tests - Particle counter values in the experiments

2000

4000

6000

8000

10000

0 60 120 180 time [sec]

Par

ticle

s

Dust Start Dust Stop

Fwd Alarm

Aft Alarm

• 2 minutes of dust generation

• Dust tests were considered valid

if the max. particle counter value was greater than 9000

Dust Start Dust Stop

Fwd Alarm

Aft Alarm2000

4000

6000

8000

10000

12000

0 60 120 180 time [sec]

Par

ticle

s



4 Dust tests - CFVS performance compared to conventional A340 detectors

• The success criterion for the CFVS was to diagnose a non-fire

case and to unconfirm the primary system’s alarm

• Thus to reduce false alarms



4 Conclusions and outlook

• The existing testing guidelines and standards for qualification of

current smoke detection systems were formulated mainly to

address the case of traditional detection technologies.

• So it was necessary to develop

- a suite of fire sensitivity tests,

- non-fire tests with dust and

- EN-54 test room modifications

applicable to video-based fire detection systems.

• Modified tests with room modifications in the Duisburg fire lab

show good match with tests performed in a cargo mock-up

for Forward, Aft and Bulk bay configuration.




• Recordings of

- point smoke detectors,

- video sensors, and

- reference measurement devices

were obtained in exhaustive test series in the Duisburg Fire Lab.

• Data were then used

- in the course of the algorithm development of the CFVS

by Goodrich Corporation, and

- for defining the performance criteria.




Possible improvements for future CFVS tests:

• Improvements in dust generation (reduction of charged particles)

• Use of additional dust types

• Tests with forced ventilation

Documents

Test Methods and Apparatus for a Video-based System for Cargo Fire Verification I. Willms, W. Krüll, J. Shirer 1 Fourth Triennial International Fire and