British Airways A321 at London on Apr 20th 2012, unreliable airspeeds

Last Update: September 29, 2013 / 08:58:35 GMT/Zulu time

Bookmark this article

Photo of British Airways G-EUXM, Airbus A321 — British Airways G-EUXM, Airbus A321 (Photo credit: Craig Sunter / Flickr / License: CC by)

Incident Facts

Date of incident
Apr 20, 2012

Classification
Report

Airline
British Airways

Flight number
BA-777

Departure
Stockholm, Sweden

Destination
London Heathrow, United Kingdom

Aircraft Registration
G-EUXM

Aircraft Type
Airbus A321

ICAO Type Designator
A321

A British Airways Airbus A321-200, registration G-EUXM performing flight BA-777 from Stockholm (Sweden) to London Heathrow,EN (UK) with 182 passengers and 7 crew, was descending towards London, thunderstorms were in the area. The aircraft joined the Lambourne hold descending to FL140 with a total air temperature of +3 degrees, the crew did not notice any icing on the airframe, however observed St. Elmo's fire. Shortly after the aircraft entered the top of clouds a white flash without any associated noise was noticed by both crew, about one second later the airspeeds on both sides fluctuated with the top and bottom of the speed scale visible, the stand by speed indicator also fluctuated however to a lesser degree. The disruption of the speed indicators lasted, according to recollection by the crew, between 10 and 120 seconds, neither pilot observed an altitude or vertical speed fluctuation. The master caution sounded repeatedly consistent with the speed fluctuations, the autopilot disconnected without associated audio and the flight controls changed to Alternate Law. The pilots worked the checklist for unreliable speed indication and turned off the flight directors. A TCAS resolution advisory appeared on the primary flight display without its associated audio, the RA commanded the aircraft to remain within +500 feet per minute and -500 feet per minute (green area), indicating any vertical speed red in excess of 500 feet per minute as red, neither pilot was sure they had noticed the RA immediately due to the lack of audio. The TCAS display showed traffic 2500 feet above them flying level while G-EUXM was in a gentle descent in compliance with the RA. With ATC permission the crew turned the aircraft away from the thunderstorm cells towards to Bovingdon hold, the instruments stopped fluctuating. The crew crosschecked speed indications with expected pitch and power tables, found the airspeed reliable again, re-engaged the autopilot, noticed a NAV-DISAGREE indication and worked the relevant checklist deciding that the no speed disagree branch was to be taken. The ECAM subsequently indicated a AOA DISAGREE message revealing that the NAV DISAGREE message had been caused by the angle of attack sensors disagreeing. The crew noticed that abnormal AOA sensor readings might cause spurios stall warnings, discussed the implications of all failures encountered and the possible various scenarios, declared PAN and decided to divert to London Stansted Airport, which was clear of adverse weather. The aircraft landed safely at Stansted Airport in Direct Law.

The AAIB released their bulletin stating, that a subsequent event was included in the investigation, see Incident: British Airways A321 at Edinburgh on Jun 16th 2012, unreliable airspeeds and concluding:

On two occasions the aircraft encountered atmospheric conditions that resulted temporarily in unreliable air data.

The first event occurred within the boundary of current icing certification standards, which only consider supercooled water droplets. The second occurred outside the proposed revised boundaries and may have involved an encounter with ice crystals. Icing certification standards are being reviewed by the manufacturer and EASA.

The hazard of such events persists. However, the safe outcome of these incidents indicates that training to deal with unreliable air data can be effective.

The AAIB reported that the captain requested maintenance to secure the flight data recorder after the flight but did not mention the cockpit voice recorder. In the time between the crew leaving the cockpit and the recorders being removed from the cockpit the cockpit voice recorder erase button had been pressed. Although the voice recording had been erased there are possibilities to recover the recording, though the recovery of the signal took longer than usual and delayed the investigation.

An examination of the aircraft could not clarify whether lightning strike marks above and below the cockpit windows were caused by a confirmed lightning strike on Apr 19th 2012, evident by the log book entries of the aircraft, or whether additional damage had been caused during the occurrence on Apr 20th. There was no damage outside the areas mentioned for Apr 19th. ADIRU3, both associated Air Data Modules and the captain's angle of attack sensor had recorded faults. ADIRU3 and its two Air Data Modules, the captain's AoA sensor and the TCAS computer were replaced.

A subsequent laboratory examination revealed no fault with any of the removed components.

The flight data recordings showed that beginning with the aircraft descending through 14,800 feet all three air data sources began to jump to unreasonable but valid or invalid data over a period of 27 seconds, at the same time there were jumps in the recorded air temperature and mach number. The #3 Air Data Computer latched a fault which remained active for the remainder of the flight, the other two air data computers did not show any faults. Following the 27 seconds period all data remained reasonable.

During the 27 seconds, although the aircraft was in a shallow constant descent, erroneous barometric data indicated a steep climb resulting in the TCAS resolution advisory, the TCAS stored a value of 3,250 fpm climb rate. The other aircraft did not receive a TCAS resolution advisory in accordance with the criteria for TCAS operation.

The AAIB reported that during both events the cockpit area microphones (CAMs) recorded "a number of periods during which large audio pulses were recorded, often resulting in a recorded waveform using the full amplitude capability of the recording. The time between pulses varied during the affected periods."

The AAIB analysed that an unrelated but similiar event, see Incident: British Airways A319 at London on Dec 17th 2010, unreliable airspeed on short final, showed such recordings of the CAMs as well, unknown to the crew with no such recordings on other CVR channels. The AAIB stated about the results of the investigation of 2010: "The investigation found that the effect on the CAM could be replicated with an electrostatic discharge applied to the connector of the CAM control panel."

The AAIB analyzed: "The April 2012 incident began shortly after a bright flash of light, generally associated with lightning. There was no noise that is often associated with lightning strikes and identifiable damage was not found on the airframe. Existing aircraft skin damage may have masked any new lightning damage. Coincidence with the bright flash does not prove causation and it was impossible to be certain that a lightning strike occurred as there are other explanations for unreliable air data indication."

The AAIB analyzed that the spurios rapid vertical speed fluctions recorded by the Digial AIDS recorder caused the TCAS resolution advisory, though the crew did not notice any such indication, possibly due to their attention focussed elsewhere or these indications not being displayed on the primary flight display. The resolution advisory was "Monitor vertical speed", the aural alert was not detected by either crew member possibly because of "inattential deafness" which is normal within human performance and is why critical alerts should be provided to more than one sense.

The AAIB analyzed: "In this incident the crew reacted appropriately to a transient unreliable airspeed situation. They maintained the aircraft within known, safe datums which allowed its systems to recover from the initiating event. The crew then made a series of decisions which reduced consequential risk: they selected a hold in VMC, diverted to an aerodrome with better weather than the planned destination and, as the weather changed, prioritised the landing task over supplemental information gathering."