4
1
Factual Information
Events prior to 29 December 2006
On 1 November 2006, the aerodrome authority began a programme of runway
resurfacing and re-profiling as part of a major project to resurface the manoeuvring
area pavements. As a result, temporary ungrooved asphalt surfaces existed over
some parts of the runway during the period when the incidents took place. A
full width section, 295 metres long, around the mid-point of the runway, had a
surface of ungrooved Marshall Asphalt base course
1
, colloquially known as âthe
patchâ by airport staff, and referred to as such throughout this report. Figure 5
(page 34) shows the runway surface condition on 29 December 2006.
On 14 November 2006, the crew of a landing Embraer 145 reported to ATC
that the runway surface condition was conducive to aquaplaning and, shortly
after, the flight crew of a landing Airbus A319 reported aquaplaning
on the left
side of Runway 27, just west of the mid-point. The runway surface at the time
of both landings was assessed for each third of its length as
âdamp, wet, wetâ
.
The ATC watch log noted that
âwater is proving slow to drain from the parts
of the runway that have been worked onâ.
ATC staff discussed this with the
airport management, who advised that pilots should be told that due to runway
resurfacing, temporary areas may be wet.
Using the guidance contained in CAP 683, â
The Assessment of Runway Surface
Friction for Maintenance Purposesâ
the airport authority issued a NOTAM on
15 November 2006, which stated:
âDue runway maintenance, sections of the runway between the
intersections of taxiway foxtrot and delta may be slippery when wet.
Braking action co-efficient readings will be available if requiredâ.
Throughout the days that followed, until 5 January 2007, the Aerodrome Safety
Unit (ASU) routinely carried out Mu-meter runs on the runway. Mu-meter is
one type of Continuous Friction Measuring Equipment (CFME); the airport
had two Mu-meters for friction testing. Following each run, ASU staff passed
friction values to ATC staff by radio. ATC staff converted those values to
braking action descriptors, using the âsnow and ice tableâ
2
available to them in
the Visual Control Room (VCR). These braking action descriptors, and/or the
friction values, were subsequently passed to flight crews.
1
Marshall Asphalt base course is normally ungrooved; it is usually only the surface course laid on top that is
grooved.
2
A table which enables friction values, assessed by Mu-meter, to be converted to descriptions of braking action, but
is only relevant to operations on surfaces covered with snow and ice. The table is reproduced at Appendix F.
5
On the afternoon of 17 November 2006, when the runway surface was
âwet,
wet, wetâ
, the crew of a landing Embraer 145 reported that braking action in
the middle section of the runway was
âpoorâ
. The Mandatory Occurrence
Report (MOR) submitted by the crew described a
âtotal loss of braking actionâ
for about 3 seconds. That evening, a Mu-meter run was recorded in the ATC
watch log as showing
âgood braking actionâ
, and the runway surface was
assessed as
âdamp, wet, wetâ
. The flight crew of a Fokker 100 aircraft which
landed soon after the assessment reported that
âsome of the middle bits of the
runway have definitely not got good braking actionâ.
Between 24 November 2006 and 27 November 2006 there were eight incidents
reported verbally to ATC, of aircraft experiencing deceleration problems during
the landing roll with aquaplaning being reported by flight crew in two of these
events.
The Airport Authority did not receive any related MORâs prior to the incidents
that occurred on 29 December 2006.
1.1
History of the flights
1.1.1
G-XLAC (Boeing 737-800) 29 December 2006 at 1150 hrs
The crew operated a return non-scheduled public transport (passenger) flight
from BIA to Chambery Aix-les-Bains. They were aware of the NOTAM stating
that the runway at BIA
âmay be slippery when wetâ
and also that rain was forecast
for their return. The flight to Chambery was uneventful and the aircraft returned
to BIA with the commander as the pilot flying.
Approaching BIA, the flight crew received the 1120 hrs ATIS which stated that
the conditions were: surface wind of 190
°
/28 kt gusting to 39 kt, visibility of
8 km, one or two octas of cloud at 800 ft, aal and five to seven octas at 1,100 ft.
The runway state was described by ATC as â
DAMP, WET, DAMP
â and the braking
action as â
GOOD
â. The flight crew briefed for an ILS approach and a landing
with flap 30 selected, autobrake set to three and the use of full reverse thrust.
The commander carried out the approach to Runway 27 and disconnected the
autopilot when he was visual with the runway. The aircraft, however, drifted
off the extended runway centreline in the strong crosswind and the commander
executed a go-around. Although he did not expect the weather conditions to
change for a second approach, he stated that he felt the crew would be better
prepared for a second attempt at landing.
6
The commander flew another ILS approach and became visual with the runway
at about 400 ft agl. The wind readout on the Flight Management Computer
(FMC) was 190°/57-60 kt
3
. Conditions were described by the commander
as âvery turbulent with very heavy rainâ. He disconnected the autopilot and
continued the approach, positioning the aircraft slightly towards the upwind
side of the runway.
The commander reported that he made the flare âas short as possibleâ, as close
as possible to the upwind edge of the runway. Although he achieved a short
flare, the aircraft drifted slightly to the right, as he expected. Once all three
landing gears were on the ground, the commander selected full reverse thrust
and confirmed that the autobrake system was operating correctly. He assessed
that the aircraft was decelerating normally and began using the toe brakes; this
caused the autobrake to disconnect.
As the aircraft rolled onto the ungrooved centre section of the runway, the
commander sensed that the wheels had âlocked upâ and believed the antiskid
system was not functioning properly due to the slipperiness of the surface. The
aircraft veered towards the downwind (right) side of the runway, but remained
on the paved surface. When the aircraft reached the grooved surface, the wheel
brakes seemed to operate correctly again.
The commander reported that he brought the aircraft to taxi speed approximately
200 m from the end of the runway and considered that had the touchdown not
been carried out towards the upwind edge of the runway, the aircraft would
have departed the downwind side of the runway and run onto the grass. He
thought the information received from ATC that the braking action was âgoodâ
was misleading.
1.1.2
G-BWDA (ATR-72) 29 December 2006 at 1215 hrs
The flight crew were flying their third sector of the day, from Guernsey to BIA,
having obtained the necessary pre-flight information, including NOTAMs and
meteorological forecasts and reports. The NOTAMs, however, did not include
any information about the condition of the runway at BIA, as the NOTAM
filtering system in use by the company excluded NOTAMs more than 15 days
old. The briefing pack provided to the flight crew included a list of the reference
numbers of 138 NOTAMs which had been excluded by the search criteria.
3
This is not an accurate âinstant windâ value, but is a weighted average of very recent wind values computed from
the inertial reference systems on board the aircraft. It gives an indication of the wind the aircraft has experienced
in the last few moments.
7
The aircraft had been loaded in accordance with the companyâs standard loading
instructions for the flight to BIA, with the centre of gravity position at landing
calculated to be 24.2% of mean aerodynamic chord. This meant the aircraft
was trimmed with the centre of gravity very slightly forward of the centre of its
range at that mass.
The flight proceeded uneventfully and the flight crew received the 1150 hrs ATIS
which stated that the wind was 180
°
/24-35 kt, visibility 3,500 m in slight rain
and drizzle, one or two octas of cloud at 700 ft and overcast cloud at 1,000 ft.
The temperature was 10
°
C and the runway condition was described as â
DAMP,
WET, DAMP
â. In accordance with company procedures, the co-pilot flew the ILS
approach to Runway 27 with 30
°
of flap and the propellers at 100% rpm.
Whilst the aircraft was carrying out the approach, the ASU team conducted
friction assessments of âthe patchâ using a Mu-meter. They reported a friction of
0.44 in the westbound direction and 0.49 in the eastbound direction. They also
measured the average friction over the entire runway length on the southern side
of the runway; this gave an average value of 0.72.
During the last few miles of the approach, the tower controller broadcast the
following wind information;
180
°
/23 kt
(landing clearance given at this point)
(83 seconds later)
190
°
/24 kt
(21 seconds later)
200
°
/26 kt
(18 seconds later)
170
°
/25 kt
(9 seconds later)
190
°
/34 kt
This final wind report was transmitted when G-BWDA was at a radio altitude of
70 ft and 15 seconds prior to touchdown.
The aircraft touched down normally at 1218 hrs, and the co-pilot gradually
applied reverse thrust. The landing roll was without incident until, at a speed
of approximately 75 kt, the aircraft yawed right slightly and the co-pilot applied
left rudder. Subsequently the aircraft began to drift left of the runway centreline
and both pilots recalled applying right rudder to correct this.
8
The co-pilot handed control to the commander
4
but retained control of the
ailerons which he continued to apply into the wind. The commander recognised
that the aircraft was still yawing and drifting to the left and, as the aircraft
slowed, he applied a nosewheel steering input on the tiller in an attempt
to correct this. However, the yaw and drift continued and the commander
perceived that the aircraft was hydroplaning. It departed the paved surface
onto the grass to the south and came to rest on a heading of approximately
227°M. The commander stated that the control inputs he had made in the
latter part of the landing roll had no effect on the aircraft.
With the aircraft at rest, the commander spoke to the cabin crew and passengers,
and ascertained that no injuries or damage had occurred in the cabin. At
the suggestion of ATC, he attempted to make contact with the Rescue and
Fire Fighting Service (RFFS) on 121.6 MHz
5
but experienced considerable
difficulty in achieving adequate communication.
1.1.3
G-EMBT (Embraer 145) 29 December 2006 at 2001 hrs
This aircraft suffered an event during landing at 2001 hrs, which has been
investigated separately and reported by the AAIB. The full report was
published in Bulletin 3/2008 on 13 March 2008. The flight crew experienced
difficulties keeping the aircraft straight as it rolled over âthe patchâ during
landing in strong crosswind conditions. The synopsis of the report stated:
âDuring the landing roll, in a strong crosswind, the aircraftâs
rudder hardover protection system (RHPS) tripped, which resulted
in the loss of both rudder hydraulic systems and reversion to the
rudderâs mechanical mode. Despite the loss of hydraulic power
to the rudder, the commander was able to maintain directional
control using a combination of asymmetric braking and rudder.
There was no fault found in the aircraft and no evidence of a rudder
ârunawayâ; high rudder pedal or brake pedal force application by
the commander, or incorrectly adjusted pedal force microswitches,
may have triggered the RHPS.
4
The ATR aircraft type is fitted with one steering tiller for use on the ground, thus the pilot in the left seat must have
control after landing.
5
121.6 MHz is a VHF communication frequency promulgated for use on the ground between flight crews and
aerodrome fire and rescue services in the UK.
9
1.1.4
G-EMBO (Embraer 145) on 29 December 2006 at 2133 hrs
The flight crew reported for duty to fly return scheduled public transport
(passenger) flights from BIA to Paris Charles de Gaulle. They were aware of the
NOTAM stating that the runway
âmay be slippery when wetâ
, and that rain was
likely. Although the commander was aware of the runway resurfacing work at
BIA and had read NOTAMs describing it, he was not aware of difficulties other
aircraft had experienced. However, on aircraft handover from the previous crew,
he was told that an ATR-72 (G-BWDA) had been involved in a runway excursion
earlier that day.
The commander operated the flight back to BIA as pilot flying; company
standard operating procedures required flight crew to carry out monitored
approaches, so control would pass from the commander to the co-pilot for the
descent and initial approach, and the commander would re-take control for
landing. The flight crew briefed for an ILS approach to Runway 27 and added
5 kt to the normal approach speed because of the anticipated turbulence.
The commander made contact with the approach radar controller at 2043 hrs
and discussed the wind conditions. The controller stated that the wind had
âMAINLY BEEN AROUND TWENTY-FIVE KNOTS AND ABOVE CONSTANTLYâ
. At
2053 hrs, the aircraft entered the hold at BIA and the commander reported that
he wished to monitor the wind for a few minutes before making the decision
whether to land or divert. At 2059 hrs, the approach controller transmitted that
the instant wind was 180°/20 kt and the two minute average was 180°/19-26 kt.
The commander replied that he wished to make an approach. The approach
controller vectored the aircraft for an ILS approach to Runway 27. At
2101 hrs the approach controller advised the flight crew that the runway was
now flooded throughout its length and the commander stated that he wished
to return to the hold pending a report on the braking action. The approach
controller advised that it would not be possible to provide braking action until
the runway ceased to be flooded. The commander then enquired as to whether
a shower was affecting the runway, and the approach controller confirmed that
it was.
At 2106 hrs, the approach controller advised that the runway surface was
âWATER PATCHES ON ALL THIRDS OVER 65% OF THE RUNWAY SURFACE, MEAN
DEPTH IS 3 MMâ,
and that braking action would shortly be provided. At 2116 hrs,
the approach controller informed the commander that the instant wind was
now 190°/20 kt, and the commander replied that he could make an approach
but needed braking action information. The approach controller confirmed
10
that the braking action check was in progress, and the commander began the
approach in anticipation of receiving the braking action prior to touchdown.
At 2122 hrs, the approach controller passed an update of the runway surface
condition, stating that all thirds were wet, and that
âBRAKING ACTION OVER
THE WHOLE OF THE RUNWAY IS zERO DECIMAL SEVEN TWO [PAUSE] WHICH
IS GOOD [PAUSE] THE MID THIRD OF THE RUNWAY WHICH IS UNGROOVED
BRAKING ACTION IS DECIMAL FIVE zERO WHICH IS ALSO GOODâ
. The
commander replied that he wished to continue the approach.
At 2131 hrs, the tower controller cleared the aircraft to land and stated that the
wind was 190°/19 kt. At this point the ASU passed Mu-meter readings of 0.52
and 0.53 for the un-grooved section of the runway. The co-pilot acknowledged
the clearance to land, and the tower controller passed information based on the
figures from the Mu-meter, stating that the braking action was
âGOODâ
.
ATC broadcast the wind as from 190°/17 kt shortly before the aircraft
landed. The aircraft broke out of cloud at 500 ft aal, and the commander
disconnected the autopilot at 300 ft. The touchdown occurred with the left
wing down and, in the commanderâs recollection, was âa little longâ. The
commander began braking soon after touchdown concentrating on applying
âeven and symmetricalâ braking whilst applying full left wing down aileron
and full aircraft nose-down elevator. He very quickly experienced difficulties
maintaining the runway centreline. The left main landing gear ran off the
runway pavement and onto the grass area south of the runway before the
commander gradually regained control of the aircraft and steered it back
towards the runway centreline. The left main landing gear had run on the grass
for 85 m; the nose and right main landing gear had remained on the runway.
The aircraft came to a halt with all the landing gear back on the runway.
The tower controller transmitted to the aircraft
â[CALLSIGN] JUST CONFIRM
YOUâRE OKâ
and the co-pilot replied
âYEAH WEâRE FINE NOW BUT WE DID
GO THROUGH THE GRASS [PAUSE] WE SKIDDED AWAY COMPLETELY ON THE
MIDDLE SECTIONâ.
The crew stated their intention to taxi to their stand but the controller suggested
they remain stationary awaiting aircraft inspection. This was conducted by
the RFFS and although they assessed the aircraft as being undamaged, they
identified that there was considerable grass and mud on the runway and that
a runway edge light had been damaged. The aircraft was also inspected by
a ground engineer, pins were inserted in the aircraftâs landing gear and the
11
aircraft taxied to a parking stand under its own power where the passengers
disembarked normally.
The commander commented that the combination of moderate rain and the wet
runway had made it difficult to see precisely where the edge of the runway
surface was during the landing roll.
The aircraft operator carried out an investigation and the interim report stated:
âprior to [this incident], company air safety reports detailing three
occasions of transient loss of braking action during the landing roll
on Runway 27 had been received. No reports involving any loss
of directional control had been received. Whilst the most recent of
these three reports were still the subject of correspondence between
[the company safety department] and Bristol Airport ATC/airport
on the day of the incident, Company Flight Operations had not
indicated any significant safety concern.â
The investigation made three initial safety recommendations regarding takeoff
and landing distance at BIA, use of crosswind limits applicable to a âslipperyâ
runway for operations at BIA whilst any part of the runway was âwetâ, and both
a one-off and ongoing review of NOTAMs throughout the companyâs network
of destinations, to ensure that the company took action to introduce temporary
operating restrictions where appropriate.
1.1.5
Events between 29 December 2006 and 3 January 2007
At 1346 hrs on 30 December 2006, an RJ-100 aircraft landed on Runway 27,
which was reported to be â
wET, wET, wET
â at the time. A Mu-meter run completed
20 minutes earlier gave an average friction value of 0.42 in the middle portion of
the runway. The flight crew reported the middle section of the runway as
âvery
slipperyâ
, and asked for it to be inspected.
Twenty minutes later, the flight crew of a departing Airbus A319 reported that
the runway was
âvery slipperyâ
in the middle section. A Mu-meter run carried
out immediately after this report found the friction values to be 0.38 and 0.42 in
the middle portion.
At 1530 hrs on 30 December 2006, the airport authority issued a NOTAM
concerning the runway condition:
12
âdue to rwy maint the rwy sfc btn the int of twys delta and foxtrot
will be slippery when wet. variable friction co-efficient readings will
be experienced throughout the rwy length and are avbl on request.
acft handling difficulties may be experienced during crosswind
conditions.â
This NOTAM stated that the runway âwill beâ slippery when wet; the previous
stated that it âmay beâ.
1.1.6
G-XLAC (Boeing 737-800) 3 January 2007
The aircraft had flown from BIA to Fuerteventura on a non-scheduled public
transport (passenger) flight and was returning to BIA with the commander as
the pilot flying. The flight crew were aware of the NOTAM stating that the
runway
âwill be slippery when wetâ
; forecasts indicated that rain was likely.
As the aircraft approached BIA at 1805 hrs, a Mu-meter friction assessment
was carried out on âthe patchâ. This gave a friction value of 0.52, and this
information and the verbal description
âGOODâ
were passed to the flight crew by
the approach radar controller.
The METAR issued at 1820 hrs stated that the wind was 210°/15-25 kt, visibility
was 10 km or more in moderate showers of rain, there were one or two octas of
cloud at 500 ft, three or four octas at 800 ft, and five to seven octas at 1,000 ft.
The temperature was 10°C and the dewpoint 9°C, and the runway was wet
throughout its length.
The commander briefed and flew an ILS approach to Runway 27 using flap
40 and
MAXIMUM
autobrake to ensure the minimum stopping distance. He
described the approach as
âdemandingâ
and stated that touchdown, at 1832 hrs,
occurred in the landing zone and the automatic speedbrake system operated
correctly.
The commander applied
âfullâ
reverse thrust and stated that he felt no response
from the brakes and that the aircraft began to skid. He maintained directional
control and, at about 110 KIAS, applied maximum pedal braking, which caused
the autobrake to disconnect. However, he did not perceive a speed reduction
as expected. He was concerned that the aircraft may overrun the runway end
but recognised that as the thrust reversers had been deployed, it would not be
possible to go around. As the aircraft ran from âthe patchâ onto the grooved
section of runway, the deceleration became more rapid and the aircraft was
brought safely to a stop prior to exiting the runway at the runway end.
13
The commander stated that he had, in the past, made landings on contaminated
runways with braking actions described as both âmediumâ and âpoorâ, but that
he had never experienced the lack of braking effectiveness which occurred on
this occasion. After the landing, another Mu-meter run was carried out, and this
also gave friction values between 0.45 and 0.52 in the ungrooved section.
1.2
Injuries to persons
1.2.1
Injuries to persons â G-XLAC 29 December 2006
Injuries
Crew
Passengers
Others
Fatal
â
â
â
Serious
â
â
â
Minor / None
9
186
â
1.2.2
Injuries to persons â G-BWDA 29 December 2006
Injuries
Crew
Passengers
Others
Fatal
â
â
â
Serious
â
â
â
Minor / None
4
52
â
1.2.3
Injuries to persons â G-EMBO 29 December 2007
Injuries
Crew
Passengers
Others
Fatal
â
â
â
Serious
â
â
â
Minor / None
4
13
â
1.2.4
Injuries to persons â G-XLAC 3 January 2007
Injuries
Crew
Passengers
Others
Fatal
â
â
â
Serious
â
â
â
Minor / None
7
187
â
1.3
Damage to the aircraft
The only damage was to G-BWDA which suffered damage to a blade on its left
propeller, which was replaced.
14
1.4
Other damage
During the G-EMBO runway excursion a runway edge light was damaged.
1.5
Personnel information
Personnel information for each flight is included in Appendix E.
1.6
Aircraft information
1.6.1
G-XLAC aircraft information
Manufacturer:
The Boeing Commercial Airplane Group
Type:
Boeing 737-81Q
Aircraft Serial Number:
29051
Year of manufacture:
2000
Number and type of engines:
2 CFM56-7B26 turbofan engines
Total airframe hours:
22,339 hrs
Certificate of Registration:
Issued on 15 March 2006 and valid
Certificate of Airworthiness:
Valid until 14 March 2008
1.6.1.1
Boeing 737-800 general description
The Boeing 737-800 is a short to medium range twin engine jet airliner (see
Figure 1). It seats up to 188 passengers and is powered by two CFM56 turbofan
engines.
Figure 1
Boeing 737-800, G-XLAC (photo courtesy of Ian Meadows)
15
1.6.1.2
Boeing 737-800 control system description
The Boeing 737-800 has a conventional flight control system with mechanically
commanded Power Control Units (PCUs) using hydraulic pressure to move
control surfaces. The rudder is a single conventional rudder without tabs. A
main and a standby rudder PCU control the rudder with mechanical inputs via
cables and control rods from the rudder pedals.
The aircraft is equipped with a nosewheel steering system which is normally
powered by hydraulic system A, but can be powered by hydraulic system B in
the event of a failure. The primary means of controlling the nosewheel at low
speed is via the nosewheel steering wheel (also known as the tiller), with limited
steering control available using the rudder pedals.
1.6.1.3
Boeing 737-800 brake system description
Each of the four main gear wheels has a multi-disc hydraulic powered brake.
The left and right brake pedals (part of the rudder pedals) provide independent
control of the left and right main gear wheel brakes. The normal brake system
is powered by hydraulic system B and can be powered by system A (alternate
brake system) in the event of a failure. The brake system also comprises antiskid
protection, locked wheel protection, touchdown/hydroplane protection, and
an autobrake system. The maximum brake pressure is 3,000 psi.
1.6.1.3.1 Boeing 737-800 antiskid system - normal brakes
The antiskid system protects the aircraft from a skid condition by releasing
brake pressure to a wheel which is about to skid. The wheel speed is measured
by a speed transducer in each wheel and there are four antiskid valves which
control the brake pressure to each wheel brake. If the antiskid system detects
that one wheel is slowing down too quickly, it will command the antiskid valve
to release some of the brake pressure to that wheel, until the wheelâs speed
starts to increase again. It will then allow brake pressure to be reapplied.
Antiskid does not operate at an aircraft groundspeed below 8 kt.
Locked wheel protection is another part of the brake system which compares
the wheel speeds of the two outboard or the two inboard pair of wheels. If
the slower wheelâs speed decreases to less than 30% of the faster wheelâs
speed, the locked wheel protection releases brake pressure from the slower
wheel. Locked wheel protection does not operate at a groundspeed of less
than 25 kt.
16
Touchdown/hydroplane protection is a system that compares wheel speed to
ground speed. If a wheelâs speed reduces to 50 kt less than the ground speed,
this system releases pressure to that wheelâs brake. Only the left outboard and
right inboard wheels are protected by this system.
1.6.1.3.2 Boeing 737-800 autobrake system
The autobrake system provides automatic braking at pre-selected deceleration
rates immediately after touchdown and for a rejected takeoff. Antiskid protection
will reduce autobrake commanded brake pressure if a skid is detected. The
autobrake select switch has six positions:
RTO, OFF
, 1, 2, 3, and
MAX
. The target
deceleration rates and brake pressures for each of the autobrake settings are
shown as follows:
The autobrake system will apply brake pressure up to the pressure limit listed in
the table in order to achieve the target deceleration rate. If the autobrake is set
to
MAX
for touchdown, the system will not apply the maximum 3,000 psi if a
deceleration rate of 0.435 g can be achieved using less than 3,000 psi. The pilot
can override the autobrake system at any time by depressing the brake pedals
sufficiently to command at least 750 psi. If the autobrake pressure is above
750 psi when the pilot commands a pressure of 750 psi, then the brake pressure
will reduce to the pilot commanded level. However, if a pilot were to rapidly
apply a pressure that is the same or higher than the autobrake pressure, then a
pressure drop would not occur.
1.6.1.4
Boeing 737-800 tyre pressures
The main gear tyre pressures for the Boeing 737-800 are 205 Âą 5 psig, but
there is a variable chart in the aircraft maintenance manual which permits this
pressure to be reduced when operating below maximum gross weight.
Autobrake
setting
Deceleration
rate (ft/s
2
)
Deceleration rate
(g)
Brake pressure
(psi)
1
4
0.124
1285
2
5
0.155
1500
3
7.2
0.224
2000
MAX/RTO
12 if below 80 kt
14 if above 80 kt
0.373 if below 80 kt
0.435 if above 80 kt
3000
17
1.6.1.5
Boeing 737-800 performance
Performance information relating to flight planning for the aircraft (ie, that used
prior to departure to assure the safety of a proposed operation) was contained
in the manufacturerâs Flight Planning and Performance Manual. Performance
information for use in flight was contained in the Quick Reference Handbook,
available on the flight deck. The two sets of information differed in that
information to be used in flight was presented largely in an un-factorised form,
whereas that used prior to flight incorporates safety factors. Data was presented
for operations on dry, damp, wet, and contaminated runways, and for braking
action good, medium, and poor, but no data was presented for wet runways
which had been notified âslippery when wetâ.
The manufacturer shows a maximum demonstrated crosswind component of
33 kt for G-XLAC for landing.
1.6.2
G-BWDA aircraft information
Manufacturer:
Avions De Transport Regional (ATR)
Type:
ATR 72-202
Aircraft Serial Number:
444
Year of manufacture:
1995
Number and type of engines:
2 Pratt & Whitney Canada PW124B
turboprop engines
Total airframe hours:
19,488 hours
Certificate of Registration:
Issued on 29 August 2003 and valid
Certificate of Airworthiness:
Valid until 28 August 2008
1.6.2.1
ATR 72 general description
The ATR 72 is a twin-turboprop short-haul regional airliner (see Figure 2). It
seats up to 72 passengers and is powered by two Pratt & Whitney PW124B
turboprop engines.
18
1.6.2.2
ATR 72 control system description
The ailerons, elevators and rudder on the ATR 72 are all mechanically
actuated without hydraulics. Two spoilers augment roll control and these
are hydraulically actuated. The rudder has a spring tab which moves in the
direction opposite of rudder movement to help reduce rudder pedal forces. The
tabâs travel increases with airspeed so that pedal force is reduced more when
aerodynamic forces are high. The rudder is linked to the aircraft structure by a
damper which regulates rudder travel speed and also reduces rudder movement
on the ground as a result of wind gusts. A rudder Travel Limitation Unit
reduces maximum rudder deflection at airspeeds above 180 kt. The maximum
rudder deflection below this speed is ¹27°.
The nosewheel steering system is mechanically controlled and hydraulically
operated. The nosewheel steering position is controlled by a steering control
hand wheel mounted on the commanderâs left console. During takeoff, landing
and taxiing operations the nosewheel steering angle is limited to ¹60°. During
towing operations, with no hydraulic pressure in the system, the nosewheel
can be deflected up to ¹91°. There is no connection between the rudder pedals
and the nosewheel steering system. There is no maximum speed limit for
nosewheel steering operation, and the Flight Crew Operating Manual (FCOM)
advises that the commander take control of the nosewheel steering at a speed
no lower than 40 kt.
Figure 2
ATR 72 aircraft, G-BWDA (photo courtesy of Stuart Lawson)
19
1.6.2.3
ATR 72 brake system description
The four main gear wheels are equipped with multidisc carbon brakes which
are each operated by a set of hydraulically powered pistons. Normal braking
is controlled by brake pedals which are part of the rudder pedals and permit the
use of differential braking to assist with steering. The aircraft is fitted with an
antiskid system which operates on all four main wheels at speeds above 10 kt.
The system measures each wheel speed and moderates the pilot commanded
brake pressure to obtain maximum stopping performance without skidding.
The ATR 72 is not equipped with an autobrake system but, maximum braking
is possible without restriction down to a stop, regardless of runway condition,
provided that antiskid is operative.
1.6.2.4
ATR 72 reverse thrust description
The four-bladed propellers on the ATR 72 are variable pitch and can be set
to negative blade pitch angles for reverse thrust operations. Reverse thrust is
commanded by moving the power levers aft of Ground Idle into the reverse
thrust range. Maximum reverse thrust can be used down to a stop if required,
although to minimise flight control shaking, it is advised that reverse thrust is
reduced to Ground Idle below 40 kt.
1.6.2.5
ATR 72 performance
Performance information relating to flight planning for the aircraft (ie, that
used prior to departure to assure the safety of a proposed operation) and to
enable calculations in flight was contained in the FCOM and other documents.
Information to be used in flight, was presented in an un-factorised form,
whereas that used prior to flight incorporated safety factors. Both factorised
and un-factorised data was presented for operations on dry, damp, wet, and
contaminated runways, but no data was presented for wet runways which had
been notified âslippery when wetâ.
The operatorâs operations manual Part B included the following information in
the âLimitationsâ section:
âMaximum Crosswind Component for Take-Off and Landing
The following maximum crosswind components apply:
Dry Runways
30 kt
Wet Runways
25 kt
Contaminated Runways 15 kt with Braking Action Medium
Contaminated Runways 5 kt with Braking Action Medium/Poorâ
20
1.6.2.6
ATR 72 operations in strong crosswinds
The operatorâs operations manual part B included the following instruction:
âTo increase nose-wheel steering effectiveness in strong cross-wind
conditions, load the aircraft to obtain a forward CG.â
The operator had not established a procedure by which this could be
accomplished and some members of the operatorâs staff, when interviewed,
were not aware of this instruction. The commander of G-BWDA was not aware
of this instruction.
The âcrosswind landingâ section of the operations manual part B also stated:
âAny reluctance to use sufficient into wind aileron will lead to the
airframe listing away from the wind direction due to the close tracked
main undercarriage. This must be avoided to ensure no additional
directional control difficulty.
If reverse is required, apply reverse slowly and symmetrically. If
problems with directional control occur reduce reverse or select
ground idle.â
With regard to normal landing technique, the manual stated:
âF/O holds control column fully forward and aileron into wind as
required to keep wings level.â
1.6.3
G-EMBO aircraft information
Manufacturer:
Empresa Brasileira De Aeronautica SA
(Embraer)
Type:
EMB-145EU
Aircraft Serial Number:
145219
Year of manufacture:
2000
Number and type of engines:
2 Allison AE3007 turbofan engines
Total airframe hours:
14,156 hrs
Certificate of Registration:
Issued on 18 February 2003 and valid
Certificate of Airworthiness:
Valid until 13 March 2008
21
1.6.3.1
Embraer 145 general description
The Embraer 145 is a 50-seat regional jet powered by two Allison AE3007
turbofan engines (see Figure 3).
1.6.3.2
Embraer 145 control system description
The rudder on the Embraer 145 is split into two sections in tandem: forward
and aft. The forward rudder is driven by the control system while the aft
rudder is mechanically linked to the forward rudder and is thus deflected as
a function of forward rudder deflection. The forward rudder is driven by
two rudder actuators connected to a PCU in the rear fuselage. The PCU is
commanded by the rudder pedals via control cables that run from the pedals
in the flight deck to the PCU. The maximum rudder deflection on the ground
is ¹15° and in the air is ¹10°. The corresponding rudder pedal deflection on
the ground is ¹9° and in the air is ¹6°.
The nosewheel steering system is electronically controlled and hydraulically
operated. The nosewheel steering position can be controlled by the rudder
pedals or by the steering handle (also known as the tiller) on the commanderâs
left console. There is no steering handle on the co-pilotâs side. The pedals can
command up to ¹5° of nosewheel steering angle, and the steering handle can
command up to ¹71°. If the pedals and steering handle are used in combination,
then a maximum of ¹76° of nosewheel steering angle can be obtained. The
steering handle is normally only used below a speed of 40 kt.
Figure 3
Embraer 145 aircraft, G-EMBO (photo courtesy of Michael Brazier)
22
1.6.3.3
Embraer 145 brake system description
The Embraer 145 has two main landing gears, with two wheels on each gear.
Each wheel has a disc brake and an associated hydraulic brake control valve.
Normal braking is controlled by toe brakes on the rudder pedals. The aircraft
is fitted with an anti-skid system which is designed to provide the maximum
allowable braking effort for the runway surface in use, while preventing
skidding. This is accomplished by measuring each wheel speed. If one wheel
speed drops significantly below the aircraftâs average wheel speed, a skid is
probably occurring, so the brake pressure is relieved to the appropriate wheel
brake until its speed recovers. The wheels and corresponding brakes are
numbered sequentially from one to four (left outboard is number one and right
outboard is number four).
The anti-skid system does not apply pressure on the brakes, but only relieves
the pilot-commanded pressure to avoid a skid. Therefore, in order to steer the
aircraft using asymmetric braking, during a heavily braked landing, the pilot
needs to reduce brake pressure on the side opposite to the direction of turn,
instead of applying pressure to the desired side.
The Embraer 145 does not have an autobrake system and G-EMBO was not
fitted with the optional thrust reverser system.
1.6.3.4
Embraer 145 performance
The operator of G-EMBO had published performance information in the
operations manual, and also operated a computerised performance calculation
system at its head office; flight crew could request calculations to be made to
meet operational needs. The operations manual gave advice and information
about operation on slippery runways, but did not define how flight crews should
make performance decisions on wet runways notified as
âmay be slippery when
wetâ
. The slippery runway table in the operations manual required knowledge
of braking action before slippery runway calculations could be made.
1.7
Meteorological information
1.7.1 Meteorological information relating to 29 December 2006
Analysis of the relevant meteorological data showed that at 0600 hrs on
29 December 2006, a complicated, multi-frontal situation existed over the UK.
A cold front over Bristol separated a moist, south-westerly warm sector and a
returning polar maritime air mass. A second frontal system moved rapidly into
23
the area, bringing further moist, warm sector conditions to the Bristol area
from 1200 hrs. By 1800 hrs, the south-west was still under moist, warm sector
conditions, with a second enclosed warm sector over St Georgeâs Channel
at 1800 hrs. In summary, from the early hours of the morning the area was
mostly affected by a strong and gusty, south-south-west to south-westerly
surface wind.
Two of the Terminal Aerodrome Forecasts (TAFs) that cover the period of the
aircraft incidents are as follows:
290413 16015G25KT 9999 SCT010 BKN020 TEMPO 0407 7000
BKN014 TEMPO 0713 6000 RA BKN008 BECMG 0811 18022G
35KT PROB30 TEMPO 1013 4000 +RA BKN004=
291601 17018KT 5000 -RA BKN008 TEMPO 1619 18020G32KT
9999 NSW BKN012 TEMPO 1901 20023G35KT 3000 +RA RADz
BKN004=
The relevant Meteorological Actual Reports (METARs) for the incidents are as
follows:
291120z 19028G39KT 8000 FEW008 BKN011 10/09 Q1007
291150z 18024G35KT 3500 -RADz FEW007 OVC010 10/09 Q1008
291220z 18023KT 3500 -RA SCT006 BKN008 10/09 Q1008
291820z 18021G31KT 7000 -RA BKN004 11/10 Q1004
292120z 18023G35KT 8000 RA SCT005 BKN007 11/11 Q1001
1.7.2
Meteorological information relating to 3 January 2007
Analysis of the relevant meteorological data showed that at 0600 hrs, a broad
warm sector covered the UK with a moist, tropical maritime airflow. An
additional warm front was over Pembrey, south-west Wales, at 0600 hrs, and
crossed Bristol a little less than three hours later. At about 1800 hrs, a cold front
was over Cardiff and crossed Bristol about an hour later.
The TAF that covers the period of the aircraft incident was:
031601 22020G35KT 9999 -RA BKN012 TEMPO 1619 4000
-RADz BKN005 PROB30 TEMPO 1719 23025G45KT +RA
BECMG 1820 25022G37KT SCT015 PROB30 TEMPO 1901 7000
SHRA BKN012=
24
and the actual report was
031820z 21015G25KT 9999 SHRA FEW005 SCT008 BKN010
10/09 Q1010
1.7.3
Reported runway conditions for 29 December 2006 and 3 January 2007
Date
29/12/2006
Time Incident
Reported Runway Condition
11:12
Damp-Wet-Damp
11:50 G-XLAC
12:30 G-BWDA
13:56
Wet-Wet-Wet
14:03
Wet-Wet-Wet
14:51
Wet-Wet-Wet
19:23
Damp-Wet-Damp
20:01 G-EMBT
21:06
Flooded-Flooded-Flooded
21:12
Water Patches-Water Patches-Water Patches
21:15
Water Patches-Wet-Wet
21:25
Wet-Wet-Wet
21:35 G-EMBO
Date
03/01/2007
18:05
Wet-Wet-Wet
18:32 G-XLAC
18:43
Wet-Wet-Wet
1.8
Aids to navigation
Not applicable.
1.9
Communications
1.9.1
Runway state reporting means and methods
The runway state is typically assessed by airport authority staff often in
vehicles moving around the airport or by ATC staff from the visual control
room. Instructions regarding assessment of runway state were included in
both the UK Aeronautical Information Publication (AIP) and the Manual of Air
Traffic Services Part 1 (MATS Part 1)(CAP 493). At BIA, the responsibility
for the runway surface state assessment rests with the aerodrome authority.
25
MATS Part 1 included the following definitions regarding reporting runway
conditions:
âDry
The surface is not affected by water, slush, snow, or ice.
NOTE: Reports that the runway is dry are not normally to be passed
to pilots. If no runway surface report is passed, pilots will assume
the surface to be dry.
Damp
The surface shows a change of colour due to moisture. NOTE: If
there is sufficient moisture to produce a surface film or the surface
appears reflective, the runway will be reported as WET.
Wet
The surface is soaked but no significant patches of standing water
are visible.
NOTE: Standing water is considered to exist when water on the
runway surface is deeper than 3mm. Patches of standing water
covering more than 25% of the assessed area will be reported as
WATER PATCHES.
Water patches
Significant patches of standing water are visible.
NOTE: Water patches will be reported when more than 25% of the
assessed area is covered by water more than 3mm deep.
Flooded
Extensive patches of standing water are visible.
NOTE: Flooded will be reported when more than 50% of the assessed
area is covered by water more than 3mm deep.
Water depth on runways is typically measured using washers
of known thickness, placed on the runway surface by a human
observer.
The AIP noted that âfor JAR-OPS performance purposes, runways
reported as DRY, DAMP or WET should be considered as NOT
26
CONTAMINATEDâ and that âfor JAR-OPS performance purposes,
runways reported as WATER PATCHES or FLOODED should be
considered as CONTAMINATED.â
1.9.2
Braking Action
The process by which braking action should be assessed and communicated to
pilots is described in MATS Part 1. It stated that when reports of braking action
are passed to pilots by radio, they should be in plain language and an assessment
should be given sequentially for each third of the runway to be used. It stated:
âin conditions of slush or thin deposits of wet snow, friction
measuring devices can produce inaccurate readings. Therefore,
in conditions of slush, or uncompacted snow, no plain language
estimates of braking action derived from those readings shall be
passed to pilots.â
No reference was made to braking action reports on dry, damp, or wet,
runways.
In explaining the terms used, it stated:
âthe word âgoodâ is used in a comparative sense and is intended
to mean that aircraft generally, but not specifically, should not
experience undue directional control or braking difficulties, but
clearly a surface affected by ice and/or snow is not as good as a
clean dry runway.â
The MATS Part 2 in use at the airport included the following instruction:
âContamination by water
The measurement of the runway friction value will not normally be
required but if requested by a pilot this value will be measured by
mu-meter (MATS Part 1 Section 9 Chapter 3 pages 2-3 refer).â
The reference to MATS Part 1 referred to above was erroneous, as that section
was removed some years before the events described in this report; the
investigation was not able to identify in exactly which revision it was removed.
Following the incidents investigated in this report, this MATS Part 2 instruction
was removed.
27
Prior to 16 November 2006, this practice was not in place at BIA but after
this date it became common for controllers to pass either braking action
descriptions in words, or the figures output by the Mu-meter, or both. It also
became common for controllers to pass such information both for the runway
as a whole, and specifically for the âpatchâ.
On 5 January 2007, conversations took place between the airport authority,
Bristol ATC and the CAA regarding the passing of braking action reports. The
CAA subsequently directed ATC staff at the airport that:
âthe passing of braking action on a wet runway is to cease. If an
aircraft requests any braking action information [controllers] are
to advise that âbraking action is unavailableâ.â
MATS Part 1 instructions regarding braking actions reports.
MATS Part 1 includes instructions to controllers regarding:
âEssential aerodrome informationâ, which is defined as âconcerning
the state of the manoeuvring area⌠which may constitute a hazard
to a particular aircraft.â
Such information must
be issued to pilots in sufficient time to ensure the safe
operation of aircraft.
Essential aerodrome information includes:
âreports on the estimated braking action determined either by
CFME
6
or by reports from pilots of aircraft which have already
landed.â
The information must include a description of the prevailing conditions, the
time of measurement or report, and the type of aircraft if an aircraft report.
1.9.3
UK AIP
The UK AIP stated:
âwhen a runway is contaminated by water (i.e. more than 3 mm), wet
snow or slush, a braking action report will not be available due to
the limitations of existing friction measuring equipmentâŚ, however,
6
Continuous Friction Measuring Equipment.
28
a runway surface condition report will normally be available stating
the type of contaminant and its respective depth.â
Although the AIP stipulated that braking action reports would not be made
available on a runway âcontaminatedâ by water, it did not specifically state that
CFME was not to be used on a runway which was simply âwetâ. The CAAâs
position was that there was no intention that such friction assessments in the wet
be used to determine braking action.
No table existed in the UK to enable the interpretation of CFME readings on wet
surfaces into braking action reports relevant to aircraft operations.
1.9.4
âSlippery when wetâ
The definition of the term
âslippery when wetâ
is described in the MATS Part 1.
It stated:
âwet surface friction characteristics of the runways at certain
aerodromes have been calibrated to ensure that they are of an
acceptable quality. If the quality deteriorates below an acceptable
level the particular runway will be notified as liable to be slippery
when wet.â
This information is repeated in AIC 15/2006 which also states that when a
runway is notified
âmay be slippery when wetâ
, aircraft operators may request
additional information relating to that notification from the aerodrome operator,
and that:
âany performance calculations or adjustment made as a result of
this information is the responsibility of the aircraft operator.â
A
âslippery when wetâ
NOTAM needs to be issued if a friction survey determines
that the friction level has dropped below the Minimum Friction Level (MFL) (in
accordance with CAP 683).
1.9.5
Notice to Aerodrome Licence Holders (NOTAL) 9/2006
NOTAL 9/2006, âWinter operationsâ laid out the requirements placed upon
aerodrome authorities regarding safe operations in adverse weather conditions.
It stated:
29
âTo provide for safe operations in adverse winter conditions,
appropriate information must be made available to pilots
and aircraft operators. However, this information has to be
reliable and relevant to the aircraft operation or movement.
The Aerodrome Licensee is responsible for the determination,
measurement and dissemination of information on the condition
of the movement area for use by aircraft, particularly if there is
any contamination by water, snow, slush or ice. Similarly, the
Licensee is responsible for the treatment of any contamination
or the withdrawal of any part of the movement area that is unfit
for use.
As part of an aerodromeâs safety management system, plans and
procedures for winter operations should be reviewed as necessary
and in a timely manner.â
The NOTAL also stated:
âIn practical terms.., using CFME in conditions beyond the
technical capabilities of the equipment and then making those
potentially inaccurate readings available to aircraft operators
or flight crew (via air traffic services) is not permitted.â
For ASU staff at BIA, the relevant operating instructions were contained
within the ASU Departmental Instruction 04/07, titled
âThe assessment and
reporting of runway surface conditionsâ
. This document was written to reflect
the requirements set out in NOTAL 9/2006.
1.9.6
Operational advice and information to flight crews
The CAA publishes FODCOMs (Flight Operations Department
Communications) on a variety of topics. In addition to being published on
the CAA website, they are sent to senior managers at companies holding
Air Operators Certificates. They are not sent to private operators or foreign
operators flying into the UK. Pilots are required to operate their aircraft
according to the instructions and advice contained in their operations
manuals. However, the inclusion of FODCOM advice into these manuals is
at the discretion of the aircraft operator and not mandatory.
30
FODCOM 19/2006, entitled âWinter operationsâ was published on 30 October
2006. Its purpose was:
âto review and refresh some of the procedures and best practice that
operators should adopt during winter operations.â
and it included information and advice about operations on âslippery when wetâ
runways.
The FODCOM stated:
âBraking action is assumed to be poor on a wet runway that is
notified as one that may be slippery when wet. Operators should
ascertain from aerodrome operators the location and dimension of
the part of the runway that has fallen below the minimum friction,
âslippery when wetâ trigger level, in order that they can assess
whether aeroplane performance is affected.
There is no reliable correlation available between the readings of
Continuous Friction Measuring Equipment (CFME) on a runway
contaminated with water, slush and snow, and aircraft braking
performance. Performance calculations must not be based on such
readings. They will not be made available at licensed aerodromes
in the UK.â
The operators of the aircraft involved in these events had not incorporated
the advice in this FODCOM regarding runways notified as
âmay be slippery
when wetâ
into their operations manuals or made their flight crews aware of
its contents. The CAA did not audit operators to establish that these processes
had taken place.
1.9.7
RFFS Communication
The flight crews of G-BWDA and G-EMBO experienced some difficulties in
communicating with the Aerodrome Fire and Rescue Service on 121.6 MHz.
RFFS staff also reported difficulties in receiving the first notification of aircraft
emergencies from ATC. The RFFS report stated:
âportable radio comms were abysmal because of constant equipment
failure due to defective batteries and radios,â
31
and included two recommendations:
âconnection of Fire Station public address system to direct link to
ATC VCR (Visual Control Room) to allow controller to pass turnout
information direct to the station [public address system]â and
âreplacement of all radios and batteries with new equipment.â
The RFFS reported that these recommendations had widespread implications.
Plans were already in place for a new fire station to be commissioned before
the end of 2010 and this station was to have facilities for direct ATC access to
the public address system within the station. A decision had been made not to
replace the RFFS radio equipment in isolation but to equip the entire airport
with new communications systems.
1.10
Aerodrome information
1.10.1
General
Bristol International Airport (ICAO code EGGD) has one runway designated
09-27. It was specified by the UK AIP
7
as having a length of 2,011 m and a
width of 46 m. Due to a displaced threshold, the Landing Distance Available
(LDA) on Runway 27 was 1,876 m. The Runway 27 displaced threshold
elevation was 601 feet amsl.
1.10.2
Runway resurfacing
1.10.2.1 Requirements for runway surfaces
Guidance on the desirable physical characteristics of runways is set out in
CAP 168
Licensing of Aerodromes
8
and this is based on the international
requirements in ICAO Annex 14. The guidance differs depending upon the
runwayâs code number. Runways with a takeoff distance available of more
than 1,800 m, such as Runway 09-27 at BIA, are code 4. Among the guidance
relating to a code 4 runway is that the transverse runway gradient should not
exceed 1.5%, but the transverse profile may be cambered or sloping. Among
the guidance relating to the longitudinal profile of a code 4 runway is that the
radius of curvature of any curved surfaces should be no less than 30,000 m.
7
UK Aeronautical Information Publication, reference AD 2-EGGD-1-4 (23 Dec 04).
8
CAP 168 Licensing of Aerodromes, Seventh Edition 8 May 2006.
32
The guidance concerning new or resurfaced runways include providing a hard
durable surface that will not generate loose materials or contaminants, provide
good surface water drainage and provide a surface friction level at or above
the Design Objective Level (DOL) defined in CAP 683
9
. The runway surface
friction guidance in CAP 683 applies to all paved runways used for public
transport operations and all paved runways exceeding 1,200 m. The DOL
friction value is 0.72 or greater when measured with a Mu-meter and 0.80
or greater when measured with a GripTester. The Minimum Friction Level
(MFL) is 0.50 with a Mu-meter and 0.55 with a GripTester.
1.10.2.2 Runway resurfacing work at BIA
The runway at BIA had previously been re-surfaced in 1990 and at that
time the runwayâs profile was not completely compliant with CAP 168. The
CAA had conceded at the time that achieving full CAP 168 compliance
within one re-surfacing operation was not practicable on economic grounds.
Therefore, in 2006, a resurfacing project was begun that would make the
runway fully CAP 168 compliant in cross-section and improve compliance
of its longitudinal profile by about 10%. The resurfacing work began on
1 November 2006 and was completed by 22 March 2007. Each night the
runway was closed at 2300 hrs for work to begin, and it re-opened at 0615 hrs
for normal operations. The work was carried out at night and during the
winter time to minimise disruption to night-time charter flights.
The technical specification used for the asphalt materials was the UK
Defence Estates Specification 013 âMarshall Asphalt for Airfieldsâ (published
August 2005). In addition to these requirements the bitumen binder had to
be 70/100 pen grade where the Marshall Stability requirement was 10 kN;
and the coarse aggregate had to have a minimum Polished Stone Value (PSV)
of 60. Furthermore, temporary ramps between asphalt layers had to have a
maximum longitudinal gradient of Âą1% and a maximum transverse gradient
of Âą 2%, with spacing between successive ramps of not less than 150 m.
An onsite laboratory was used to monitor compliance with the materials
specifications.
The runway surface at BIA is made of Marshall Asphalt which consists
mainly of stone material bonded together with bitumen. The top layer is
called surface course (also known as wearing course) and the layer beneath
this is called base course (also known as binder course). Both layers are
made of Marshall Asphalt, but the base course has a more indented texture
9
CAP 683 The Assessment of Runway Surface Friction for Maintenance Purposes, Third Edition 14 May 2004.
33
and larger aggregate size compared to the surface course. A regulating
course can also be used when reshaping of the runwayâs profile is required.
A regulating course is usually made up of a âbase courseâ mixture which
can be laid and compacted in thicker layers than a âsurface courseâ material
because it contains a larger aggregate size. A surface course is then laid on
top of the base course.
Water does not drain through the surface of Marshall Asphalt, so in order
to meet friction requirements in the wet, transverse grooves are made in
the surface (see Figure 4). The grooves are typically 3 to 4 mm wide and
4 mm deep with 25 mm spacing, and combined with a transverse slope,
these grooves allow water to drain towards the sides of the runway.
Figure 4
Section of original weathered grooved Marshal Asphalt runway surface.
Grooves are 3 mm wide, 4 mm deep with 25 mm spacing
(photo courtesy J. Barling)
The approach to the resurfacing works was to start with the reshaping and
base course layers before beginning to lay the surface course starting from
the 09 end of the runway. Due to bad weather, laying of the surface course
was delayed and by 16 November 2006 most of the base course had been
laid without any surface course. Approximately 60 m of surface course was
then laid during a typical nightâs work, starting at the 09 threshold end and
34
progressing eastwards. The new surface course was then left for 72 hours
before grooving to allow time for evaporation of volatile components and
for cooling so that a degree of surface hardening could take place. This
minimised any damage to the new surface course during the saw-cutting
process. Typically no more than 100 m of un-grooved surface course was
exposed at any one time.
The state of the runway resurfacing works between 29 December 2006 and
3 January 2007 is shown in Figure 5. The white sections in this diagram
represent the original weathered runway surface consisting of grooved
Marshall Asphalt (see sample in Figure 4). The green sections consist of new
base course and regulating course. The purple sections consist of new surface
course that has not yet been grooved. The blue section consists of new surface
course that has been grooved. The green and purple sections, taken together,
represent the surface area of the runway that was un-grooved.
Figure 5
State of runway surface between 29 December 2006 and 3 January 2007
The difference in water drainage capability of the grooved surface compared to
the un-grooved surface is visible in Figure 6, which is a snapshot taken from the
RFFS video of the incident scene shortly after G-BWDA departed the runway
on 29 December 2006.
35
1.10.2.3 Runway rectification work following incidents on 29 December 2006 and
3 January 2007
Following the incidents on 29 December 2006 and 3 January 2007, the airport
operator decided to improve the braking characteristics of the un-grooved
base course under wet conditions. Closing the runway until the base course
had been covered with surface course and then grooved was considered
uneconomical. It was considered that grooving the base course might provide
a temporary solution. It is not normal practice in the UK to groove base course
because its more indented texture and larger aggregate size makes damage
to the groove shoulders more likely following multiple landings. Therefore,
on 7 January 2007 a trial area of base course 10 m long was grooved and
subjected to a day of landings. An inspection the following day revealed that
the grooved base course was holding up well and probably would not present
a FOD (Foreign Object Damage) risk in the short term. Figure 7 shows the
difference between the temporary grooved base course in the foreground and
the un-grooved base course behind it.
Figure 6
Image taken shortly after G-BWDA departed the runway, looking eastwards
towards the 27 threshold
36
The runway was subsequently closed between 1400 hrs on 7 January and
1000 hrs on 8 January 2007 to permit grooving of the exposed base course.
Normal resurfacing operations resumed on 11 January 2007 which included
laying surface course on top of the grooved base course, after machining away
the grooves.
The airport operator carried out frequent monitoring of the temporary grooved
base course and by 10 January 2007 the grooves had remained generally intact
with little sweeping necessary. However, over the 12-day period that the grooved
base course was exposed, a small quantity of aggregate was lost from the groove
ridges.
Figure 7
Temporary grooved base course in foreground and un-grooved base course
behind it. Grooves in the base course are 4 mm wide,
4 mm deep with 25 mm spacing
(photo courtesy J. Barling)
37
1.10.2.4 Independent runway surface inspection on 7 January 2007
The AAIB employed an experienced runway surface consultant to provide an
independent evaluation of the runway surface condition. He examined the
runway on 7 January 2007 before the temporary grooving was started. In his
opinion, the newly laid base course was well laid, â
fairly tight knitâ
, with only
small areas of segregation of the mix and no evidence of irregularities in the
profile. The surface at the time of inspection was damp to wet and no â
ponding
â
was present. He also reported that the original weathered grooved Marshall
Asphalt:
âsurface had good macrotexture, having lost most of the fine material
at the surface through wear and weathering, exposing coarse
aggregate fractions of the mixed material.â
Only a cursory visual examination of the new ungrooved Marshall Asphalt
surface course was made as this had not played a part in the G-BWDA and
G-EMBO runway excursion incidents. He reported that this surface appeared
typical of a well laid new Marshall Asphalt surface course, with a tight surface
with little macrotexture.
1.11
Flight Recorders
Recorded data was successfully recovered from each aircraft by the operators,
and sent to the AAIB for analysis. Despite a 2-hour Cockpit Voice Recorder
(CVR) being fitted to G-EMBO, the voice data was overwritten before the CVR
was impounded.
The condition of the runway surface affected the three aircraft in different ways.
The G-XLAC flight crews reported concerns of runway overruns whereas the
G-BWDA and G-EMBO crews reported lateral control issues.
1.11.1
Longitudinal Effects on G-XLAC
A number of recorded parameters were available to characterise the longitudinal
landing performance of G-XLAC. One which gives the broadest view of the
rate at which the aircraft was slowing down is the longitudinal acceleration. It
takes into account all forms of aircraft drag force: aerodynamics, reverse thrust
and braking. The total longitudinal retardation force is a product of the aircraft
mass and longitudinal acceleration.
38
1.11.1.1 Deceleration vs. Runway Position
The touchdown positions for both G-XLAC landings have been estimated using
the recorded localiser and glideslope deviations. Knowing the groundspeed at
the touchdown point and integrating the longitudinal acceleration twice, allowed
an estimate of aircraft position on the runway to be calculated. Figure 8 shows
the estimated positions of G-XLACâs landings, with the corresponding recorded
longitudinal acceleration
10
.
Landings were performed on Runway 27 so, for ease of plotting, the data is
presented showing the landings from left to right. The areas highlighted in
green are the areas of ungrooved base course.
The two predominant areas of ungrooved base course start at approximately
550 m and 800 m from the Runway 27 threshold. The larger of the two green
areas marked âthe patchâ in Figure 8, shows a notable drop in deceleration for
each landing.
Analysis of the G-XLAC event of 29 December 2006 has provided the most
useful results of the two incidents, as maximum braking was commanded
throughout the transition over âthe patchâ.
1.11.1.2 Deceleration Components
Data from the G-XLAC events of both 29 December 2006 and 3 January 2007
was forwarded to the manufacturer who analysed the landing performance.
This, along with use of aircraft models of thrust and aerodynamics, allowed the
contributions to the aircraft deceleration from the aerodynamic, reverse thrust
and landing gear effects, to be separated.
Figure 9 shows a combined plot of the decelerations for the G-XLAC landing on
the 29 December 2006. The â
GEAR CONTRIBUTION
â in this graph constitutes the net
deceleration component from the landing gear which includes rolling resistances
and braking forces. In this aircraft configuration, the main contributor to the
â
GEAR CONTRIBUTION
â was from the braking. The aerodynamic contribution in
Figure 9 decreased as expected, as the airspeed decreased.
10
The position relative to the runway is an estimated position and is subject to a number of inaccuracies.
These inaccuracies can arise from sources including accelerometer drift, estimated touchdown position,
wind speed and direction. Some small adjustments were made to align the drop in deceleration of each
aircraft by adjusting the touchdown point.
39
Figure 8
G-XLAC longitudinal acceleration vs. runway position
for incidents on 29 December 2006 and 3 January 2007
âthe patchâ
40
Just after touchdown, the largest single contributor to the aircraft deceleration
was from the landing gear. At this point, the aircraft deceleration peaked at
-0.44g. After around 2 seconds, a notable reduction in deceleration occurs
down to around -0.3g due to a decrease in the landing gear contribution. The
deceleration drop was largely restored by the increase in reverse thrust at around
the same time (dashed blue line). With full brake pressure commanded, the
antiskid system operated and reduced the brake pressure, thus reducing the
overall deceleration.
The deceleration components for the G-XLAC landing on 3 January 2007 are
shown in Appendix C, Figure C1. Both G-XLAC landings show a similar
characteristic in that, just after touchdown, the deceleration contribution from
the landing gear was significant.
Figure 9
G-XLAC component deceleration contributions
29 December 2007
(Reproduced courtesy of Boeing)
41
1.11.1.3 Braking Coefficient
The braking coefficient is defined as the ratio of the decelerating force
from the braking system, relative to the normal load applied to the tyres.
This coefficient is a term which includes effects due to the runway surface,
contaminants and the aircraft braking system (antiskid efficiency
11
, brake
wear, tyre wear etc). It is a better measure of the effectiveness of an
aircraftâs braking system, in given runway conditions, than the longitudinal
deceleration. This is not the same as the tyre-runway friction coefficient
because it includes contributions from the aircraftâs braking system.
A braking system operates either under torque-limited or friction-limited
conditions. A torque-limited situation is one where the amount of braking
force which can be applied through the tyre is limited by the amount of
applied brake torque. In this case the tyre-runway friction can react the
applied brake torque, so antiskid is inactive. The braking coefficient in a
torque-limited case is a function of the level of brake pressure applied.
A friction-limited situation exists where the amount of braking force which
can be applied is limited by the friction between the tyre and the runway. In
this case, the antiskid is regulating the brake pressure to ensure the wheel
continues to rotate at an optimum speed, to provide the best available grip.
As long as the system is friction-limited, the braking coefficient represents
the maximum braking coefficient for the runway surface.
For the G-XLAC landing on 29 December 2006, runway conditions were
reported as
âdamp, wet, dampâ
. This runway surface was friction-limited on
the ungrooved area as confirmed by Figure 9 which shows the landing gear
deceleration component decreasing, despite full brake pressure being applied.
In this case, the maximum braking coefficient can be calculated using the landing
gear deceleration contribution and the normal load applied to each landing gear
(essentially the aircraft weight minus lift).
The manufacturer of G-XLAC provided the braking coefficient data for this
landing which is shown in Figure 10. The lower the braking coefficient, the
more slippery the surface (minimum 0, maximum 1).
Figure 10 shows the braking coefficient increasing after touchdown, up to a
peak of 0.33. After transitioning on to the ungrooved surface, the braking
coefficient dropped to a minimum of 0.11. Transition back on to the grooved
surface lead to an increase up to a maximum of 0.36.
11
Antiskid efficiency is defined as the effectiveness of the antiskid system to modulate the brake pressure
and subsequently braked wheel speed, to give optimum grip.
42
âthe patchâ
Figure 10
G-XLAC Braking coefficient from 29 December 2006
43
1.11.2
Lateral Effects
G-BWDA and G-EMBO both departed the left side of Runway 27 on
29 December 2006. With a crosswind from the left, aircraft have a tendency
to yaw to the left when on the ground due to the âweather cockingâ effect of
the vertical stabiliser. At higher speeds on the ground, this is counteracted
using the aerodynamic effects of the rudder. As speed reduces and the rudder
becomes less effective, there is a higher reliance on the nosewheel steering
and differential braking for directional control. Relying on the landing gear
for directional control therefore relies on the available grip between the tyres
and the runway surface.
1.11.2.1 G-BWDA
Data for G-BWDA was provided by the operatorâs Flight Data Monitoring
(FDM) programme. This data was recorded from the same data concentrator as
the Flight Data Recorder (FDR).
The approach to BIA was made with the autopilot disengaged, 28° of flap and
heading slightly into wind. Data shown in Appendix C, Figure C2, begins with
the aircraft just prior to touchdown. In the 10 seconds prior to touchdown, a
number of rolling manoeuvres are noted from -9.5° left, to 4.6° right and then
back to -11.9° left.
Touchdown occurred at 12:18:25 hrs at an airspeed of around 100 kt. The
power levers were slowly moved into the reverse thrust range and torque on
both engines increased over 10 seconds. Reverse thrust was maintained until
the aircraft came to a stop. Ground spoilers deployed on touchdown and the
aircraft decelerated to a steady state longitudinal acceleration of -0.26g within
four and a half seconds. No brake pressure or pedal angle parameters were
recorded so it was not possible to ascertain the level of braking applied.
The recorded localiser deviation suggests that touchdown was achieved almost
on the runway centreline. At the point of touchdown, 3.4° of left rudder was
applied with a roll angle of 1.8° to the right and the left aileron deflected 1.3°
up (maximum deflection is ¹ 14°). During the next four seconds, progressively
more and more right rudder was added to a maximum of -26.7° (maximum
travel is ¹ 27°). Heading then began to increase (right yaw), after which 15.4°
of left rudder was applied.
This led to a decrease in heading (left yaw) and with this left rudder maintained,
at a groundspeed of 76 kt, the localiser deviation shows G-BWDA starting to
44
deviate to the left of the runway centreline. Rudder position was then again
reversed to provide -25.0° of right rudder in an attempt to arrest the rate of turn.
However, with this rudder input maintained, localiser deviation continued to
increase, signifying the aircraft moving to the left of the runway centreline.
When the aircraft came to a rest, heading had decreased to 227°.
Nosewheel steering angle and tiller position were not recorded and with no
mechanical linkage between the rudder pedals and nosewheel steering, pilot
inputs cannot be ascertained.
Appendix D shows where G-BWDA stopped on the grass, just beyond the
green area of ungrooved base course. The photograph in Appendix D shows
where the runway surface transitions from ungrooved base course, to the
normal runway surface condition. This confirms that G-BWDA was located
on the ungrooved base course area as it left the runway.
Appendix C, Figure C2 also shows the longitudinal acceleration relative to time
which did not show any significant loss in deceleration while the aircraft was on
the runway. Braking system data was also not available so it is unknown what
the landing gear contribution to the deceleration was and whether the braking
was symmetric. However, if the runway condition was slippery, it would have
had some impact on the tyre adherence to the runway. The drop in deceleration
towards the end of the landing shown in Appendix C, Figure C2 was most likely
caused by the wheels contacting the grass.
1.11.2.2 G-EMBO
The aircraft landed at around 2133 hrs at a computed airspeed of around 136 kt
with 45° of flap and on a heading of 263°. The spoilers deployed immediately;
no thrust reversers were fitted.
Recorded data indicated that 3.5 seconds after touchdown, brake pressures
on wheels one and three increased
12
, which led to an increase in longitudinal
deceleration, peaking at -0.32g (see Appendix C, Figure C3). The deceleration
then decayed, coinciding with a decrease in brake pressures. The pressures did
not then rise above 360 psi (maximum is 3,000 psi) for the next six seconds.
During this six-second period, progressively more and more right rudder
pedal was applied to counteract a slow decrease in heading. The heading
continued to decrease, despite further rudder pedal inputs, and 14 seconds
12 Brake pressure is only recorded on wheels one and three.
45
after touchdown, at a groundspeed of 67 kt, full right rudder pedal deflection
was recorded, with heading still decreasing. Although left wing down roll
had been commanded from touchdown, at this point this was reversed to right
wing down roll with the control wheel deflected to 24 degrees (maximum
deflection is 41 degrees). The longitudinal deceleration then began to
increase in line with brake pressure on wheel three, just as the localiser
deviation shows the aircraft deviating to the left of the runway centreline. As
this deviation continued, the heading started to increase. Right rudder pedal
demand was reduced as the heading continued to increase and left rudder
pedal was slowly applied to reduce the rate of turn. Brake pressure on wheel
three then increased significantly up to 1,236 psi as the localiser deviation
decreased and the aircraft regained position on the runway centreline.
The low brake pressure coincided with a decrease in longitudinal deceleration
and an increase in the heading. The reason for the drop in brake pressure
was either due to the pilot reducing the brake pedal input, or the antiskid
reducing the pressure after detection of a skid. This cannot be confirmed as
the brake pedal angle and wheel speeds were not recorded, so the command to
the brakes cannot be ascertained. Brake pressure was also only recorded on
two brakes and only every second
13
.
1.12
Aircraft examinations
1.12.1
G-BWDA examination
One blade from the left propeller had sustained some impact damage, most
likely from clumps of dirt that had been thrown up during the runway
excursion. There was a lot of mud inside the gear bay and on the landing
gear; this was washed off to allow examination. Both nosewheel tyres were
found worn close to limits. Each tyre had four grooves and the remaining
tread depth was an average of 1.3 mm for the two centre grooves and 2.6 mm
for the outer grooves on both tyres. The tyres are required to be replaced
when the bottom of any groove is reached at any location; the wear on the
tyres had not reached this limit. No anomalies were noted on the four main
wheel tyres and they had tread depths remaining of between 3 mm and
6 mm. The normal nosewheel tyre pressures on this aircraft type were 64 to
66 psi and the normal main wheel tyre pressures were 114 to 119 psi. The
tyre pressures were not measured after the incident: however, they were
routinely checked every three days.
13
Antiskid systems regulate brake pressure at a rate much faster than 1 Hz.
46
1.12.2
G-EMBO examination
G-EMBO did not sustain any damage from its runway excursion, but the left
main gear brake units were subsequently replaced due to dirt ingress. The
left main landing gear tyres were also removed for inspection. They had not
sustained any damage and were only about 30% worn with 7.2 mm of tread
depth remaining. The tyre pressures on G-EMBO had been checked daily; the
last recorded values on 21 December 2006 were 80 to 86 psi for both nosewheel
tyres and 147 to 150 psi for the four main wheel tyres.
1.12.3
G-XLAC examination
G-XLAC did not undergo any special inspection as it did not depart the
runway in either incident. There were no reports from general daily checks
that there were any anomalies with the tyres or any relevant anomalies with
the aircraft. The operatorâs standard policy was to maintain the tyre pressures
at 200 +5/-0 psig and to check tyre pressures each day.
1.13
Medical and pathological information
Not applicable.
1.14
Fire
There was no fire.
1.15
Survival aspects
Not applicable.
1.16
Tests and research
None.
1.17
Organisational and management information
The operator of G-EMBO was acquired by another established operator
after the incident and the new operator does not plan to operate Embraer 145
aircraft in the long term. However, action was taken to introduce procedures
for operations in wet conditions on runways which had been notified
âslippery
when wetâ
, introducing a crosswind limit of 10 kt. At another UK airport,
where similar runway works were carried out in 2007, the operator took action
to consider any lengths of runway promulgated as
âslippery when wetâ
, as
being absent from runway distances, for performance calculation purposes.
47
Following the events on 29 December 2006, one operator took action to impose
a temporary maximum crosswind of 15 kt on its Airbus A319 fleet, for all
departures from BIA.
Following the event on 3 January 2007, there was further discussion, both
formal and informal, between the airport authority and operators. On
5 January 2007, some operators took action to cease flying at the airport
altogether, and others imposed restrictions on their operations. The additional
restrictions typically included application of crosswind limits similar to those
applicable on slippery runways and performance adjustments to take account
of poor friction in the landing and rejected takeoff cases.
1.17.1
ICAO and CAA action
AAIB investigators met with CAA staff to discuss the events described in this
report. CAA staff explained that CAA policy with regard to runway friction is
derived from the provisions of Annex 14 to the ICAO Convention.
At the first meeting of the ICAO Aerodrome Operations and Services Working
Group (AOSWG), in 2005, it was agreed that the provisions in Annex 14
Volume 1 relevant to runway surface friction should be reviewed. The basis of
such review was to be the safety factors inherent in the measurement of runway
surface friction where the runway is contaminated. At the third meeting of the
group, in March 2006, the UK CAA proposed considerable amendment to the
Annex. The CAAâs view was that, except on runways covered by compacted
snow or ice, friction values should not be used as a basis for aircraft operations.
Their reasons for this viewpoint were an absence of a common method across
the world, doubts about the accuracy of such measurements, the difficulty in
reading friction measurements across to aircraft operations and concerns that
such measurements are time-critical.
The Group recommended that a task force should be established to investigate
runway surface friction issues and subsequently to develop an action plan for
submission to the ICAO Aerodromes Panel. Draft terms of reference for a
proposed Friction Task Force, along with some of the friction related issues
which the Group felt needed to be addressed at a global level, were outlined.
The CAA has a seat on this task force.
48
1.18
Additional information
Runway friction measurement
The CAA published guidance to aerodrome licensees on runway friction in two
principal documents, CAP 168,
âLicensing of Aerodromesâ
and CAP 683,
âThe
Assessment of Runway Surface Friction for Maintenance Purposesâ
.
CAP 168,
âLicensing of Aerodromesâ
CAP 168,
âLicensing of Aerodromesâ
, gives advice and direction to aerodrome
licensees regarding runway surfaces:
âThe aim should be to provide in the first instance a runway surface
that is clean and has a uniform longitudinal profile and friction
levels that will give satisfactory braking action in wet conditions.
These issues should be addressed at the time of the design of
runways, pavements or subsequent resurfacing.â
Another relevant passage deals with surface friction characteristics:
âThe surface of a new runway or a newly resurfaced runway should
be designed and constructed to enable good braking action to be
achieved by aeroplanes in wet runway conditions. When a new
runway is built or an existing runway resurfaced, the wet surface
friction characteristics shall be assessed in order to classify the
friction level.â
In a section entitled
âNew Asphalt Runwaysâ
, the publication stated:
âNew or resurfaced runways with an asphalt surface normally
do not provide adequate friction levels for aircraft operations
immediately after the new surface has been placed... In these
circumstances it is generally necessary to treat the surface by
either the application of a coarse textured slurry seal, grooving or
the addition of a porous friction course.â
The publication provided advice regarding the application of slurry and the
grooving process.
49
CAP 683,
âThe Assessment of Runway Surface Friction for Maintenance
Purposesâ
CAP 683 reflected the CAAâs interpretation of the Standards and Recommended
Practices laid down in Annex 14 to the Convention on International Civil
Aviation, in so far as they had been adopted by the UK in respect of runway
surface friction testing.
The purpose of the document was to outline the procedures for undertaking
runway surface friction assessments and to define the criteria by which friction
values were assessed on runways under specified conditions.
The criteria in the document applied to all paved runways exceeding
1,200 metres in length and all paved runways used for public transport
operations. The document detailed methods for assessment of runway friction,
using the Mu-meter and GripTester (the two types of CFME most commonly
used in the UK).
The document stated that the friction characteristics of a runway can also
âalter
significantlyâ
following maintenance activities. It stated that a runway surface
friction assessment:
âshould be conducted following any significant maintenance
activity conducted on the runway and before the runway is returned
to service.â
Further, it added:
âRunway surface friction assessments should also be conducted
following pilot reports of perceived poor braking action...â
Management of works in progress
CAP 168, â
Licensing of Aerodromes
â, details the responsibilities of aerodrome
licensees with regard to work on operational areas. It stated:
âWherever major work affecting operational areas is planned,
aerodrome licensees must be satisfied that unacceptable risks
generated by Works in Progress (WIP) have been identified and
removed, and that procedures are provided and followed which
ensure no adverse impact upon levels of safety.â
50
1.18.1
Friction Measuring Equipment
The friction measuring equipment commonly used today to measure runway
friction are continuous friction measuring devices, known as CFME. These
devices continuously measure friction as they travel along the length of a
runway. The two types of CFME accepted by the CAA for use in the UK
are the GripTester and Mu-meter. Other types of CFME can be used if their
performance can be demonstrated, to the satisfaction of the CAA, to provide
comparable results with the Mu-meter and GripTester. The Mu-meter,
manufactured by Douglas Equipment, was in use at BIA.
1.18.1.1 Douglas Mu-meter Mk6 CFME
The primary Mu-meter used by BIA was the Mk6 Mu-meter; a Mk5 unit was
available as a backup. The Mk6 Mu-meter is a three-wheeled trailer as shown
in Figure 11. The centre wheel is used to measure distance, and the two outer
wheels are connected by a load cell to measure drag resistance. These two
outer wheels are toed-out at an angle of 7.5° so that they are partially skidding
as they are pulled along. Strain gauges in the load cell measure the force by
which the wheels are being forced apart. This force can be correlated to the
coefficient of friction
14
(mu) between the runway surface and the Mu-meterâs
tyres, and is calculated by a laptop computer connected to the Mu-meter. The
laptop is also used by the driver of the towing vehicle to set up each measuring
run and to monitor his driving speed. The target speed for the Mk6 Mu-meter
is 64 km/hr (40 mph). The Mu-meter can be used in dry, wet, compacted snow,
or icing conditions; it can also be operated in a self-wetting mode. CAP 683
states:
âA runway surface friction assessment is conducted under
controlled conditions using self-wetting CFME, to establish the
friction characteristics of a runway and to identify those areas of
a runway surface that may require attention.â
In self-wetting mode a water tank trailer is used to spray a metered amount of
water (nominally creating a surface covering 0.5 mm deep) under the wheels
in order to measure runway friction in simulated wet conditions.
14
The coefficient of friction (known as âmuâ for the Greek symbol Âľ) is a dimensionless quantity used to calculate
the force of friction (static or kinetic). The coefficient of friction is defined as the ratio of the friction force (F),
between the two surfaces in contact, to the normal force (N) between the object and surface (Âľ = F/N). The
coefficient of friction is not a âmaterial propertyâ but rather a âsystem propertyâ as it is dependent upon the physical
characteristics of two surfaces, and is also dependent upon variables such as speed and temperature.
51
The Mu-meter is calibrated by pulling it over a plywood strip that is coated
with grit bound in an epoxy resin. The surface feels rough to the touch and
is designed to generate a mu of 0.77 between the surface and the Mu-meterâs
tyres. Calibration of the Mk6 Mu-meter is generally only required if the tyres
are changed or if the unit starts to generate unusual readings.
1.18.2
Friction Measurement Data from BIA
The airport operator carried out a number of friction measurement runs
throughout the resurfacing works using the Mu-meter. No runs with self-wetting
were carried out until 10 January 2007, after the runway excursion incidents.
The operator reported that between the middle of November and the end of
December the runway was never dry enough for a Mu-meter run with self-
wetting. There was one dry period on 8 December but no staff were available
to conduct the runs.
However, many Mu-meter runs were undertaken in damp
and natural wet conditions. These types of runs give good relative friction
values across a length of runway, but the absolute values need to be treated
with caution.
Figure 11
Douglas Mu-meter Mk6 CFME used by BIA
shown here using the self-wetting equipment
(photo courtesy Douglas Equipment Limited)
52
An example result of a Mu-meter run carried out in natural wet conditions is
shown in Figure 12. This run was undertaken over the whole runway length
on 29 December 2006 at 2125 hrs, 10 minutes prior to G-EMBO departing the
side of the runway. The runway surface condition at the time was declared
as
âwet, wet, wetâ
. The friction readings varied from a minimum of 0.4 to
a maximum of 0.95. The runway surface condition is depicted along the
distance axis. There is a good correlation between low friction readings and
the ungrooved surfaces (purple and green). The grooved surfaces (blue and
white) all have higher friction readings. Marshall Asphalt runway surfaces
with good friction characteristics typically have Mu-meter values of up to 0.8
in dry conditions. The high readings above 0.9 in Figure 10 in wet conditions
are unusual. The Mu-meter was calibrated on 29 December 2006 at 1600 hrs
using the calibration board; however, Mu-meter runs undertaken both before
and after this calibration show similarly high friction readings above 0.9. One
run at 2350 hrs on the same day revealed a number of friction readings equal
to 1.0. Neither the airport authority nor the Mu-meter manufacturer could
explain this anomaly but it indicates that all the actual friction values were
probably lower than measured.
Some short Mu-meter runs of 300 m were also carried out on 29 December 2006
over the longest stretch of ungrooved base course (âthe patchâ, shown in green
in Figure 5). These were referred to as âpatch runsâ by the operator. A summary
of average Mu-meter measurements for the âpatch runsâ and the full runway
length runs on 29 December 2006 are shown in Table 1.
Table 1 lists the average mu for each run in both directions and then lists the
dual average, which is the average of the two runs. Several Mu-meter runs of
the ungrooved âpatchâ revealed friction values less than the MFL of 0.50. This
data was compiled by the airport operator and it does not include minimum
mu values or the lowest 100 m rolling average as recommended by CAP 683
for evaluating runway surfaces against the MFL. Furthermore, given that
these runs were not carried out in controlled dry conditions with self-wetting,
the overall averages need to be treated with caution. The clearest information
about the state of the runway surface was revealed in the relative differences
of friction values in Figure 10. The typical friction of the new ungrooved
portions was approximately 0.4 mu less than the original grooved sections.
On 10 January 2007 the airport operator carried out Mu-meter runs in dry
conditions with self-wetting to evaluate the temporary runway surfaces
against CAP 683 requirements. This was done after the temporary base course
surfaces were grooved, so the friction of the ungrooved base course was never
53
measured in controlled self-wetting conditions. An example friction run plot
from 10 January 2007 is shown in Figure 13. This run was carried out with the
Mu-meter displaced 4 m left of the centreline. Runs in other lateral positions
showed similar results. The lowest 100 m rolling average was not listed, but
the small figures at the top of the graph are the averages for 100 m sections.
The lowest 100 m average measured in this run was 0.67 which was above the
MFL. Of significant note is that the large peaks and troughs from the run on
29 December 2006 (Figure 12) have disappeared.
To further illustrate the change in friction characteristics after the temporary
grooves were made, the Mu-meter manufacturer provided the AAIB with some
2D colour-coded plots of measured friction for 5 January 2007 and the survey
runs on 10 January 2007. These plots are included in Appendix B.
Full Runway Length Mu-meter Runs on 29/12/06
TIME
Average Mu 27-09
Average Mu 09-27
Dual Average
11:05
0.76
0.82
0.79
12:14
0.72
n/a
n/a
16:29
0.74
0.81
0.78
21:23
0.70
0.75
0.73
23:48
0.74
0.79
0.77
300 m Patch Mu-meter Runs on 29/12/06
TIME
Average Mu 27-09
Average Mu 09-27
Dual Average
6:59
0.63
0.60
0.61
8:21
0.42
0.43
0.43
11:02
0.66
0.70
0.68
12:12
0.44
0.49
0.46
14:01
0.60
0.49
0.54
14:50
0.56
0.52
0.54
16:41
0.69
0.74
0.72
21:30
0.50
0.53
0.51
23:54
0.49
0.57
0.53
Table 1
Summary of Mu-meter Runs carried out on 29 December 2006
54
Figure 12
Friction measurement run of Runway 09-27 on 29 December 2006 when the
runway state was
âwet, wet, wetâ
. The runway surface condition at the time is
depicted at the bottom
55
Figure 13
Friction measurement run of Runway 09-27 on 10 January 2007 using
self-wetting on a dry runway, displaced 4 m left of the centreline
56
1.18.3
Previous runway resurfacing works at Luton Airport
The same design and construction companies that carried out the BIA runway
resurfacing works also carried out runway resurfacing at Luton Airport earlier
in 2006.
At Luton, part of the runway resurfacing works included replacing the central
15 m, highly trafficked, portion of the runway with new base course along
approximately 75 % of the runway length. This was to be carried out before the
surface course overlay was begun. It was identified that during the resurfacing
works aircraft would be operating on a number of different surface types,
including existing grooved Marshall Asphalt, new ungrooved Marshall Asphalt
base course, new ungrooved Marshall Asphalt surface course and new grooved
Marshall Asphalt surface course.
The Luton airport operator specified in the works contract that should
the minimum friction level, measured by a 100 m rolling average, of any
temporary surface course drop below 0.55 (as measured by a GripTester),
then the contractor would be required to retexture
15
the temporary surface
to improve its friction. An exception was made during the period between
laying the surface course and the grooving operation, when the operator would
accept one continuous section not exceeding 200 m providing less than the
minimum friction level. There was no restriction on the maximum length of
ungrooved base course; this was considered to have sufficiently better friction
characteristics than ungrooved surface course.
The resurfacing work began on 1 March 2006 and was completed on
15 June 2006. Throughout this period, friction measurements were taken on
a nightly basis under self-wetting conditions. Despite the numerous different
surface conditions, all the measured friction levels were above the MFL and
the lowest recorded friction was 0.63 with the GripTester. Therefore, at no
time did the airport operator need to notify the surface as âmay be slippery
when wetâ. Based on the high friction measurement readings the contractor
received permission from the operator and the CAA to increase the maximum
ungrooved surface course length to 300 m. Therefore, between 27 April 2006
and 5 June 2006 the ungrooved length of surface course varied between 210 m
and 300 m. When wet weather was forecast, new surface course was not laid
on the main runway; instead, resurfacing was done with base course, or work
was carried out on the runway shoulders.
15
The type of retexturing was not specified but the contractor reported that they would have considered ultra high
pressure water jetting, the Klaruw retexturing process (a controlled percussive process), or temporary grooving.
57
The experience gained during the Luton resurfacing works gave the contractors
confidence that the friction levels of the long sections of ungrooved base
course at BIA would not cause an operational problem.
1.18.4
Previous runway resurfacing works at Belfast City Airport
The same runway designer was involved with runway resurfacing work at Belfast
City Airport which was carried out between November 2003 and February 2004.
During this resurfacing programme the base course and regulating course were
laid continuously along the full length of the runway before any surface course
operations were begun. Heavy Duty Macadam (HDM) was used for the base
course, with the intention of laying a surface course of Marshall Asphalt. During
the period from early November until late December, friction values of the
ungrooved HDM base course were measured at about 0.65 with a GripTester.
These high friction values were attributed to the coarse texture of HDM coupled
with the good water shedding properties of the reshaped runway.
Considering the relatively short runway length (Runway 22 has an LDA of
1,767 m) concern was raised that the ungrooved Marshall Asphalt surface might
provide an unacceptably low friction level. A consultation with airlines was
undertaken at an early stage and the main Airbus A321 operator agreed that
they were happy to continue operations as long as no stretch of runway of more
than 100 m was measured as having a friction less than the MFL. Based on this
requirement the contractor ensured that no more than 100 m of surface course
was left ungrooved at any one time. The surface course was laid during January
and February 2004 with the friction tests on the ungrooved Marshall Asphalt
surface measuring between 0.55 and 0.58 with the GripTester.
1.18.5
Risk assessment carried out by BIA prior to runway resurfacing works
BIA produced a â
Safety Case
â report on 22 September 2006 to provide evidence
and argument that the runway resurfacing project had been designed and would
be â
constructed and
brought into operational service in a safe and efficient
manner
â. It included a risk assessment outlining the potential hazards, their
severity, their likelihood of occurrence and risk reduction measures.
The hazards associated with the project were listed under four categories:
â
Batching Plantâ, âTaxiway Golfâ, âRunwayâ
and
âTaxiway Zuluâ
. Under the
âRunwayâ category the following three hazards were identified:
58
Aircraft movements will conflict with contractors
â
Vehicles, not associated with the work will conflict with
â
contractors
Debris left on the works area may present a hazard to aircraft
â
operations
The severity of the first two hazards was categorised as âaccident from collision
with plant or personnelâ, and the severity of the debris hazard was categorised as
âsignificant incident from damage to aircraftâ. The risk reduction measures for
these hazards were listed as follows:
âThe runway will be closed during construction periods. At the end
of each shift the runway will be cleaned, inspected and returned to
operational service.â
The â
Safety Case
â report did not mention friction and did not include any
hazards relating to aircraft having braking or directional control difficulties on
the temporary runway surfaces. According to the BIA Operations Director, the
friction requirements were implied in the report by references in the â
Safety
Case
â to CAP 168 and CAP 683. He said that all temporary surfaces were
required to meet the MFL of 0.5 (based on mu-meter measurements).
In addition to the â
Safety Case
â report, BIA had produced a â
Project Risk Register
â.
This included one risk/hazard relating to friction which was listed as â
Unacceptable
post construction friction readings
â. The mitigation measure for this risk was
â
Selection of Materials
â and the risk owner was the contractor.
1.18.6
Hydroplaning
16
There are three types of hydroplaning: viscous hydroplaning, dynamic
hydroplaning, and reverted rubber hydroplaning. All three can degrade both the
braking and directional controllability of an aircraft.
Viscous hydroplaning:
This can occur on wet runways and is a technical term
used to describe the normal slipperiness or lubricating action of the water. Viscous
hydroplaning occurs when a tyre is unable to puncture the thin residual film of
water left on a paved surface. This water lubricates the surface and reduces its
friction. This type of lubrication can be reduced by making the runway surface
rough. Viscous hydroplaning can occur at water depths of less than 0.025 mm.
16 The information on hydroplaning has been obtained from:
Aircraft Accident Investigation
by Richard H. Wood and
Robert W. Sweginnis, and from an article in
Flight Safety Australia
, September-October 2000, by Graham Bailey.
59
Dynamic hydroplaning:
This is the phenomenon that is normally referred to as
aquaplaning. It can occur when an aircraft lands fast enough on a sufficiently
wet runway. When the aircraftâs speed and water depth are sufficient, inertial
effects prevent the water from escaping from the tyre footprint area, and the tyre
is held off the pavement by the hydrodynamic force. Dynamic hydroplaning
requires a minimum water depth of 0.25 mm for worn tyres and 0.76 mm for
new tyres. Dynamic hydroplaning is also a function of tyre pressure. Studies
indicate that the minimum speed (in knots) for dynamic hydroplaning to occur
is approximately 9âp, where p is the tyre pressure in psi
17
.
Reverted rubber hydroplaning:
This situation can follow dynamic or viscous
hydroplaning and results when the aircraft wheels become locked. The locked
wheels create enough heat to vaporise the underlying water film forming a cushion
of steam that eliminates tyre to surface contact. Indications of an aircraft having
experienced reverted rubber hydroplaning, are distinctive âsteam-cleanedâ marks
on the runway surface and a patch of reverted rubber
18
on the tyre.
Hydroplaning affects both the stopping distance and directional control of an
aircraft. According to Wood and Sweginnis:
âthe loss of cornering or side-force capability when braked wheels
are operated at slip ratios greater than 25% can account for the
tendency of an aircraft to weathervane into the wind when braking
on wet runways during crosswind landings.â
1.18.6.1 Estimated dynamic hydroplaning speeds for the incident aircraft
Using the equation 9âp and the estimated tyre pressures for the incident
aircraft, the following estimated minimum dynamic hydroplaning speeds can
be calculated:
17 The equation 9âp applies to a rolling wheel. If the wheel becomes locked, then the dynamic hydroplaning speed is
reduced to 7.7âp.
18 Reverted rubber refers to rubber that has reverted to its un-cured state and become sticky and tacky.
Main gear
Nose gear
Aircraft
Pressure (psig) Speed (kt) Pressure (psig)
Speed (kt)
G-BWDA
114 - 119
96 - 98
64 - 66
72 - 73
G-EMBO
147 - 150
109 - 110
80 - 86
80 - 83
G-XLAC
200 - 205
127 - 128
n/a
n/a
Table 2
Dynamic Hydroplaning Speeds
60
1.18.7
National Transportation Safety Board (NTSB) (USA) N471WN Chicago
Midway International Airport investigation
On 8 December 2005, a Boeing 737-700 aircraft, registered N471WN, overran
the end of runway 31C at Chicago Midway International Airport. The NTSB
investigated the accident and issued a report in October 2007 (report reference
NTSB/AAR-07/06). The report included discussion of a number of safety issues,
including
ârunway surface condition assessments and braking action reports.â
The report also included an analysis of the assessment of runway surface
conditions, and this discussion of the use of aircraft-generated friction
measurements:
âThe circumstances of this accident demonstrate the need for
a method of quantifying the runway surface condition in a
more meaningful way to support airplane landing performance
calculations. The Safety Board and industry practice of analyzing
an airplaneâs actual landing performance in the aftermath of
an accident based on airplane-recorded data demonstrates that
runway surface condition and braking effectiveness information
can be extracted from recorded data.â
Two of the recommendations were:
1. Establish a minimum standard for 14 Code of Federal
Regulations Part 121 and 135 operators to use in correlating
an airplaneâs braking ability to braking action reports and
runway contaminant type and depth reports for runway surface
conditions worse than bare and dry. (A-07-63)
2. Demonstrate the technical and operational feasibility of
outfitting transport-category airplanes with equipment and
procedures required to routinely calculate, record, and convey
the airplane braking ability required and/or available to slow
or stop the airplane during the landing roll. If feasible, require
operators of transport-category airplanes to incorporate use of
such equipment and related procedures into their operations.
(A-07-64)
61
1.18.8
Previous incidents and AAIB Safety Recommendations
In 2003 the AAIB issued a report on an incident to an Embraer 135 which
overran a slush covered runway while landing at Norwich Airport. A number of
Safety Recommendations were made including:
âIt is recommended that the CAA encourage research that could
lead to the production of equipment that can accurately measure
the braking action of runways under all conditions of surface
contamination. (Safety Recommendation 2003-96)â
The CAA response was:
âThe CAA accepts this recommendation. In response to the
concerns of airlines when operating on runways of inferior friction
characteristics, the CAA has convened a working group, involving
airlines, aerodrome operators, research and development bodies
and manufacturers of runway friction measurement devices, to
address operational runway friction issues, including winter
operations. The working group recognises that research
worldwide has so far failed to provide an accurate measurement
of friction or braking action on a runway contaminated by slush
and wet snow, and that there are wider operational issues such
as the reliability of the reported measurement, that also need to
be addressed.â
In addition to the challenges and costs of developing a friction measurement
device suitable for runways contaminated by slush and wet snow,
manufacturers also have to consider whether there is sufficient market
for such a device. However, the CAA is content to continue to encourage
research that could lead to the production of equipment that can measure
accurately the braking action under all conditions of surface contaminant.
During the consultation period for issuing this report, the CAA provided
further update which confirmed that research had been carried out on a
modified friction measuring device which could measure runway friction
under contaminated runway conditions. The detail of this research was
presented to the European Aviation Safety Agency (EASA) in June 2008.