Captain Adams and other experts -Inflight fires (Air Safety Week)

Captain Adams assisted the CTSB in the investigation of sr111:

Potential In-Flight Fire Events Occur Daily; Many in Inaccessible
Areas

"Crews don't have the diagnostic tools to
locate the sources of smoke and fire."Jim Shaw, line
pilot and manager, Air Line Pilots Association
in-flight fire project

DAYTONA BEACH, Fla. - If a
pilot's greatest fear is an in-flight fire, aircrews in
the U.S. airline industry are facing this frightening
specter an average of three times a day. In one out of
every three of these cases, pilots were sufficiently
concerned to execute a precautionary landing.

More
than 960 smoke and in-flight fire events were recorded
in 1999. They triggered more than 350 unscheduled
landings. In many cases, evidence of electrical arcing was
discovered. After routine repairs, the aircraft were returned
to service. Even if these events were trivial,
although many were not, safety was affected. According to
Jim Shaw, manager of the in-flight fire project for
the Air Line Pilots Association (ALPA), the rate of
one unscheduled landing per day compromises safety.
"Every day there's at least one pilot and crew landing
at an airport they are not familiar with," he
declared. Shaw presented his findings at the Society of
Automotive Engineers (SAE) "Advances in Aviation Safety"
symposium.

Even at half the rate, Shaw's numbers are
staggering. They result from a compilation of Service
Difficulty Reports (SDRs) in the first 10 months of 1999. To
be sure, this period of time covers the months
following the Sept. 1998 fatal crash of Swissair Flight
111, in which an out-of-control electrical fire on the
accident MD-11 is the primary suspect. Pilots, sensitized
in the wake of the Swissair disaster, may have been
landing their aircraft with unusual alacrity at the first
whiff of a strange smell or the distinctive odor of an
electrical malfunction. The dread is universal, as evidenced
by a recent article in a Canadian publication for
bush pilots.

Many events not reported

Pilot
anxiety may be only a partial explanation. Shaw offers
another, more chilling point: even though the SDR database
represents the richest lode of such information, it may
represent a significant undercount of actual events. For
example, one SDR report cited "several air returns with
smoke odor" (e.g., unscheduled landings) for the
affected airplane. In addition, under U.S. regulations
concerning Immediate Notification (49 CFR 830.5), operators
are required to submit reports of in-flight fires,
irrespective of the outcome. However, smoke without fire is
not a reportable event, despite the folk wisdom
"where there's smoke there's fire." Consequently, the
Federal Aviation Administration's (FAA) incident database
contained only about 1/7th the number of reports Shaw
discovered in the SDR database. Of 155 SDR-reported events
that necessitated an aborted takeoff, return to the
block, emergency descent or unscheduled landing, only 21
such cases were found in the FAA incident database for
the same time period.

For these reasons, Shaw believes his tally
represents "minimum numbers," and that the actual count "can
reasonably be presumed" to be greater.

Sam Holoviak,
a pilot who also is a member of ALPA's central air
safety committee, observed with the vivid memory of
personal experience, "The sound of a popped circuit
breaker raises the hair on the back of your head." With
this observation in mind, highlights of Shaw's
findings should raise hair on the backs of many heads:

High numbers. An average of about 100 smoke and
fire-related events is being recorded per month, or roughly
three per day. An SDR database search based on entering
the words "Smoke and not False Warning" for the
Nature of Condition, and the words "insulation or
wiring, and char or burn or short" in the Summary turned
up 964 events over the 10 months from Jan. 1, 1999
to Nov. 2, 1999. Only 12 reports mentioned "fire,"
but flames in a hidden area (e.g., behind a panel,
under the floor) would more likely be recorded as
"smoke."

High temperatures. Most events involved
high temperatures: 578 out of the 964 total. A
high-temperature event was subjectively defined as one involving
reference to smoke from a solid material or an electrical
overload, short, bearing seizure, or whatever resulting in
a popped circuit breaker. A second category
involving air contamination was defined as a possibly
combustible event involving smoke or mist appearing to result
from leakage of oil or hydraulic fluid into the
ducting (i.e., an "air contamination event."). Some 60
percent of the events met the "high temperature"
definition.

Mostly electrical. Of 392 events that
were categorized, more than 300, or 80 percent,
related to aircraft electrical systems and components.
For example, an overheated and smoking fluorescent
light ballast was logically coded as an electrical
malfunction. On the other hand, smoke wafting out of an air
duct was not automatically presumed to be electrical
in nature. In order of descending frequency, the
following abridged list of failures illustrates the range
of reports:

Description Count

Fan 68
Connector
12

Wire 48
Circuit breaker
8

Light 27

Battery charger
4

Coffeemaker 17

Control panel
3

With 60
reports citing wiring, wire-harnesses and connectors,
some 11 percent of the high temperature events were
associated with wiring.

Mostly unprotected. About
380, or 66 percent, of the 578 high temperature events
occurred in or near the cabin, galley, lavatories and
cockpit (including 50 high temperature events in the
electronics and equipment, or E&E, bay immediately below the
cockpit). Of course, these are the areas where electrical
systems tend to concentrate, so the failure density
correlates with equipment density. Significantly, these also
are the areas, excepting the lavatories, where
neither built-in fire detection nor suppression is
provided.

Mostly uncontrollable. "If the crew didn't know
where the smoke was coming from, they didn't have
access or controllability," Shaw observed. By this
measure, he concluded, "Control of the event, even at a
minimal level, was possible by the crew in less than 40
percent of the events, and access of some kind was
possible in less than 20 percent." It is his belief that
the best locations for in-flight smoke and fire
detection and suppression are "in the inaccessible areas of
the fuselage."

"The crews don't have the
diagnostic tools to locate sources of smoke and fire," Shaw
maintained. Although not mentioned at the SAE symposium, the
inability to control smoke/fire events is by no means
restricted to U.S. and Canadian skies. In 1993 a smoldering
electrical fire ignited by a short circuit on a Swissair
DC-9 completely filled the cockpit with dense, acrid
smoke. The pilots were barely able to see through the
windscreen, and at one point resorted to "wagging" the
checklist in a vain attempt to disperse the smoke as they
made an emergency landing at Munich. In its report of
this case, the German Aircraft Accident Investigation
Branch (FUS) suggested the use of an "inflatable
viewchannel" so crews in such circumstances could see their
instruments and gain at least a limited view through the
windscreen (see ASW, June 14, 1999). Such a device, known as
the Emergency Vision Assurance System (EVAS), is now
on the market (see ASW, Dec. 21, 1999).

Making it worse. In most cases, when crews reset popped
circuit breakers they made the situation worse. Shaw
found 12 events that reported the circuit breakers were
reset. In eight of those cases, additional smoke,
arcing, or damage occurred. In the remaining four cases,
a serious malfunction had occurred and resetting
the breakers injected the possibility of further
damage. Although the number of reset circuit breaker
cases is not that great, they lend weight to the recent
ALPA recommendation that circuit breakers should not
be reset for all but the essential systems needed to
get the airplane safely on the ground (see ASW, March
6).

Shaw's findings suggest much more in the
way of corrective action. Better reporting discipline
is needed to improve the quality of the data. Smoke
events should be added to the current requirement to
report in-flight fires. This action, Shaw believes,
would mitigate "the significant under-reporting of
smoke events."

A complete rethinking

Shaw
and other like-minded individuals who have studied
the hazard posed by in-flight smoke and fire have a
number of solutions in mind. At the SAE symposium they
presented a menu of actions to plug vulnerabilities and to
improve needed defenses against catastrophe. Their ideas
include:

Better detector placement.
Capt. Ken
Adams, another member of ALPA's safety committee,
believes a step as simple as placing detectors in each of
the main air conditioning ducts would greatly assist
pilots in isolating the source of smoke, enabling them
to better differentiate between "aerosolized" oil
leaking into the air conditioning system and the
potentially greater hazard posed by an electrical fire (see
expanded comments of Capt. Adams in ASW, Nov. 1, 1999).

Full coverage . Shaw believes coverage needs to be
expanded. "We can't keep living in the Stone Age. We need
detection and suppression throughout the aircraft," he
urged. Priority should be placed on inaccessible areas
of the fuselage.

Better detectors. Presently, even in locations on
the aircraft where smoke and fire detectors have been
placed, the industry is plagued with a high false alarm
rate. Tom Cleary, a fire researcher with the National
Institute of Standards and Technology (NIST), has found in
his work that the smoke detectors in aircraft cargo
compartments have a false alarm rate of about 100-to-1. This
ongoing epidemic of false alarms undermines pilot
confidence in the smoke detection equipment and certainly
complicates their decision- making. However, Cleary believes
that new technology offers the promise of reducing the
false alarm rate by "at least" a factor of 10, as
evidenced in recent NIST research. The techniques include
improved signal processing to eliminate spurious signals,
to compensate for "baseline drift" and to diagnose
faults in the detector itself. The integration of
photoelectric, ionization, smoke, and carbon monoxide detectors
also will contribute to that tenfold reduction in
false alarms.

Cleary believes the airline
industry will find the new detector affordable. After all,
it's being developed under the auspices of the
Department of Housing and Urban Development (HUD) for
low-income housing, he observed.

Better suppression.

New water-mist technology under development by the
U.S. Navy to protect its warships may find application
in the airline industry. According to Dr. Fred
Lawson at the Naval Research Laboratory, "Water mist is
better than Halon." Described as "God's gift to
firefighters," Halon production has been banned as a result of
its impact on global warming, and the airline
industry is searching for a substitute. Lawson believes
that a new, extremely fine misting technology ("Which
does not feel wet to the touch") may be the answer.

In trials aboard the U.S.S. Shadwell, an amphibious
warship converted to test vessel, Lawson said a mere 20
gallons of water converted to mist can kill a fire in a
60,000 cubic foot machinery space. The mist, he said,
"Brings down a totally untenable situation dramatically."
Lawson recounted that the water mist reduced the
temperature from 900? F to about 120? F "within seconds."

He dismissed the fear often expressed by airline
industry officials that the water mist would foul the
functioning of vital electronics. On the Shadwell, Lawson
said, "A 30- minute exposure posed little or no hazard
to energized switchboards, motor controllers or
electric motors."

Adams believes the existing
potable water supplies on an airliner may be more than
adequate for a water-mist fire suppression system to
protect the presently unprotected cabin area. A full-up
system of detection and suppression would weigh on the
order of 200 pounds. The Navy will be testing a
prototype water-mist system in the cabin of a retired B737
early next month (system performance will be duly
reported in this publication).

Better access.

Adams pointed out that fires can propagate behind
panels. Indeed, one such fire occurred behind the flight
engineer's panel on a Delta Airlines [DAL] (more) L-1011
enroute from Hawaii to California in late 1998 (see ASW,
July 12, 1999). The flight crew could see the flames
behind the narrow seam of the metal cover but could not
squirt fire-suppressing chemical from the hand
extinguisher into the interior. Adams believes that access
ports to insert extinguishers could be installed,
enabling crews to apply extinguishing agent without having
to waste time with screwdrivers or fire axes opening
panels.

The need for such emergency fire extinguishing
ports is growing, Adams maintained. More electrical
components are migrating into the cabin area "Little closets
with avionics stuffed in them," he said. Adams
recalled the recent case of a 767 on a Trans-Atlantic
flight where the electronics in one of these closets
overheated, and the crew was forced to divert to a
precautionary landing in Ireland.

Better procedures.

According to Holoviak, instances of in-flight smoke and
fire need to be treated "as equal to the complete loss
of one engine on a twinjet." The procedures, he
said, should be designed to "get the aircraft pointed
in the right direction immediately" for landing.

Emergency checklists, Adams declared, "should be designed
to relieve the crew of troubleshooting, focus on
emergency diversion, depower to essential items, and land
as soon as possible." As an example, the first three
items on the Delta 767-400 smoke/fire emergency
checklist consist of: (1) don masks, (2) establish
communications, (3) begin the divert.

With respect to
depowering flight essential systems, Bill Sell, an engineer
with German circuit breaker (CB) manufacturer E-T-A
Elektrotechnische Apparate GmbH, suggested a simple system of
standardized color-coding. The breakers for critical circuits,
he suggested, could be colored red. "If it pops, you
know immediately a flight essential circuit is
affected," he said.

"You could put a colored cap
over the button, or use a simple insert, e.g., color
code the rating label" on the CB, Sell explained. In
other words, the red-coded CBs would help pilots to
"sweat the red stuff," as it were. The color-coding,
Sell believes, would be a boon to pilots, who are
"under pressure anyway" when the CBs start popping.

Better design. Over the longer term, improved electrical
system design would help harden airplanes against the
ravages of electrical fires. There should be no "single
point" failures, Shaw maintained. "Separation of the
captain's and the first officer's instrument power wires
should be ensured," he said. Overall, he argued, "No
single wire bundle failure should result in loss of a
significant amount of electrical systems." Shaw,
e-mail jdshaw@compuserve.com; Adams, e-mail
kmuleadams@csi.com; Sell, e-mail wmfsell@aol.com

The Stark Findings in Brief
"We found a high number of
smoke and fire events that resulted in unscheduled
landings or were not controllable or accessible by the
crew." Jim Shaw

Lessons learned:

A
very high number of smoke or fire events occur on
transport aircraft in the US and Canada - 964 over 10
months.
478 were high-temperature events, vastly
outnumbering nearly 400 reports of air contamination (possibly
combustible).
A detailed look at 392 high temperature
events showed that 80 percent involved electrical
systems and components.
In the overwhelming number of
cases, the crew had limited ability to recognize, gain
access to, or to control the malfunction.
The
resetting of tripped circuit breakers with internal or
external short circuits generally made things worse.

There is an average of more than one unscheduled
landing a day due to smoke or fire based only on Service
Difficulty Report (SDR) data.
The SDR database
under-reports the significance of the problem.
Source: Shaw

Fire, Smoke and Toxic Gases on Board
"An
in-flight fire is one of the most dreaded situations for
flight crews. It's every pilot's recurring nightmare!
Most often electrical in origin, fires sometimes
smoulder for a long time before they break out. Fires are
deceitful, ruthless and difficult to control. In addition to
their flames and heat, their toxicity is fearsome...be
afraid of this ravenous and noxious enemy."

January 25, 1999
"The Swissair MD-11 accident has
opened a lot of eyes about many different aspects of
coping with fire in the cockpit. Of course, the
investigation is as yet incomplete but enough is known for us
to make some reasonable assumptions:

The
pilots were not sufficiently alarmed at the outset for
them to try and rush to land. In part this might be
because of a suspension of belief ('This can't be
happening to me') but also because large electrical fires
always start off as smaller, seemingly innocuous ones.
They get to be bigger catastrophic fires because,
under the current checklist philosophies (and
switchology), the power never comes off all the wires. The lack
of alarm could also be attributable to an induced
'society of success' attitude that prevails in simulator
training. That optimistic doctrine of positivism would have
it that, as long as a crew reacts per 'the book'
with the correct CRM approach, all will come right in
the end. Unfortunately, real life is not like that."

Source: John Sampson