AAIB BULLETIN

SPECIAL S2/89

ACCIDENT TO BOEING 737-400 G-0BME AT KEGWORTH, LEICESTERSHIRE ON 8 JANUARY 1989

This bulletin contains a statement of the facts which have been determined up to the time of issue. It is published under Regulation 6 of the Civil Aviation (Investigation of Accidents) Regulations 1983 to inform the aviation industry and the public of the general circumstances of the accident. It must necessarily be regarded as tentative, and subject to alteration or correction if additional evidence becomes available. The bulletin is not an accident report - either final or interim.

On the 8 January the Chief Inspector of Accidents appointed Mr E J Trimble, a Principal Inspector of Accidents, to carry out an "Inspector's Investigation" in accordance with the regulations. Much work remains to be done before his report can be compiled. Following this he, as the Inspector, must invite and consider representations on the draft report from the parties involved (in accordance with Regulation 11) before it is completed for submission to the Secretary of State for Transport. Unless one of these parties asks for a Review Board (under Regulation 12) the report will then be published.

Thus nothing in this bulletin is to be taken as a final statement of the facts and circumstances of the accident, nor would it be right to draw any conclusions from it as to the cause(s) of the accident.

This bulletin can be reproduced without specific permission providing that the source is acknowledged.

Aircraft type and Registration	Boeing 737-400, G-OBME
No & Type of Engines	2 CFM 56-3C high by-pass turbo-fan engines
Year of Manufacture	1988
Date and Time (UTC)	8 January 1989 at 2025 hrs
Location	Kegworth, near East Midlands Airport, Leicestershire
Type of Flight	Scheduled passenger
Persons on Board	Crew 8 - Passengers - 117 + 1 infant
Injuries	Crew 7 (Serious) - Passengers 47 (fatal)
	Passengers 66 + 1 infant (serious)
	Crew 1 (minor) - Passengers - 4 (minor)
Nature of Damage	Aircraft destroyed
Commander's Licence	Airline Transport Pilot's Licence
Commander's Age	43 years
Commander's total Flying Experience	13180 hrs (of which 765 were on Boeing 737-300/400 types)
Information Source	AAIB Inspector's Investigation under the Civil Aviation (Investigation of Accidents) Regulations 1983)

History of the flight

The aircraft was engaged on a double shuttle between London Heathrow and Belfast. It landed at Heathrow at 1845 hrs after completing the first shuttle flight and took off again for Belfast at 1952 hrs with the first officer handling the aircraft. After take-off, the aircraft climbed initially to flight level (FL) 60, where it levelled off above a layer of stratocumulus cloud for two minutes, before receiving clearance to climb to FL 120. Soon afterwards, at 1958 hrs, clearance was passed for the aircraft to continue its climb to its cruising FL of 350 on a direct track to the Trent VOR (Very high frequency Omni-Range beacon).

At 2005.05 hrs, as the aircraft was approaching FL 290, the flight crew experienced moderate to severe vibration, a burning smell and smoke. The commander immediately took control of the aircraft, disengaging both the auto-pilot and the automatic throttle. The two pilots then diagnosed the symptoms of vibration and smoke as indicating a problem in the right engine and 20 seconds after the onset of the vibration, the commander instructed the first officer to throttle back the right engine. The commander later stated that the action of closing the right throttle reduced the smell and signs of smoke and that he remembered no continuation of the vibration after the right throttle was closed.

Immediately after throttling back the right engine, the first officer advised London Air Traffic Control (ATC) that the aircraft was at FL 300 and that they had an emergency situation which looked like an engine fire. When this message had been passed, the commander ordered the first officer to shut-down the engine; the flight crew were then engaged in ATC radio transmissions, stating their intention to divert to Castle Donington (East Midlands Airport). During this period a female cabin attendant used the cabin address system to advise the passengers to fasten their seat belts. The right engine was shut down 2 minutes and 7 seconds after the vibration began. By that time power had been reduced on the left engine, which continued to operate at comparatively low power. After the accident, the commander stated that during the remainder of the descent the indications from the engine instruments were such as to confirm that the emergency had been successfully concluded and that the left engine was operating normally. The recorded engine parameters associated with this stage of the flight are included in the section on "FDR and CVR evidence".

In the cabin, the passengers and the cabin attendants had heard an unusual noise accompanied by moderate to severe vibration. Some passengers had also been aware of what they described as smoke, but none were able to describe its colour or density. They described the smell of burning as "rubber", "oil" and "hot metal". Many had seen signs of fire from the left engine, which they described variously as "fire", "torching" or "sparks". Several of the cabin attendants described the noise as a low, repetitive "thudding", and one described how the vibration had been severe enough to shake the walls of the forward galley. Soon after the right engine had been shut-down, and in response to a cabin "chine" from the commander, the flight service manager (FSM) came to the flight deck. The commander asked him if they had had smoke in the cabin. He replied that they had. Later, after another statement from the FSM that the passengers were becoming concerned, the commander broadcast on the cabin address system that there was trouble with the right engine which had produced some smoke in the cabin, that the engine was not shut-down and that they could expect to land at East Midlands Airport in about 10 minutes. Passengers stated that the smell of smoke had dissipated by the time the commander made this announcement.

The right engine had been shut-down approximately 5 nm north-west of East Midlands Airport. Having cleared the aircraft to turn right and descend to FL 100, London ATC passed control to Manchester ATC, who passed headings for the aircraft to descend to the north west of East Midlands Airport, before vectoring it to the east of the airport to begin its approach to runway 27. ATC control of the aircraft was then transferred to Castle Donington Approach.

The approach then continued until the aircraft was on the localiser of the instrument landing system (ILS) for runway 27, with flaps lowered to 5°. At 2,000 ft the landing gear was lowered and, as the outer marker was passed at 4.3 nm from touchdown, 15° of flap was selected. One minute later, at 2023.50 hrs, when the aircraft was 2.4 nm from touchdown and at a height of 900 ft above ground level (agl), the left engine lost power with compressor speed reducing rapidly and high vibration levels. The commander told the first officer to relight (ie restart) the right engine. 17 seconds after the power loss, the fire warning system operated on the left engine. No power became available from the right engine before the aircraft struck the ground at 2024.43 hrs, 36 seconds after the fire warning.

The initial ground impact was in a nose-high attitude on level ground just to the east of M1 motorway. The aircraft then passed through trees and suffered its second, and major, impact on the western (ie northbound) carriageway of the M1 and the lower part of the western embankment: this second impact occurred some 70 metres after the initial impact and 10 metres lower. The fuselage was extensively disrupted and the aircraft came to rest entirely on the wooded western embankment, approximately 900 metres from the threshold of runway 27 and displaced 50 metres to the right of the centreline of the approach lights.

Ground witnesses who saw the final approach of the aircraft saw clear evidence of fire associated with the left engine. The intake area of the engine was filled with yellow/orange fire and flames were observed streaming aft of the nacelle, pulsating in unison with "thumping noises" which emanated from this engine. Metallic "rattling" noises were also heard and flaming debris was observed falling from the region of the burning engine.

Wreckage examination

Examination of the fuselage showed that two major structural failures had occurred in the impact, one slightly forward of the wing leading-edge and one aft of the wing trailing-edge. These had respectively resulted in the fuselage nose section becoming detached from the centre section and the tail section buckling over, and to the right of, the centre section. The forward fuselage had therefore sustained a high degree of disruption in the passenger cabin, with floor structure, seats and furnishings becoming detached; similar damage had occurred around the aft fuselage failure. The floor structure and seating were much less disrupted in the centre (ie over-wing) section of the cabin and in the inverted tail section.

The Flight Data Recorder (FDR) and Cockpit Voice Recorder (CVR) were removed during the early morning of Monday 9 January and taken to AAIB Farnborough for readout and analysis.

The No 1 (left) engine showed evidence of fire damage, including that arising from ground fire. The left wing appeared to have contained its fuel, whereas the right wing had suffered sufficient impact damage to cause leakage of its fuel contents, which had run back down the embankment slope, on to the motorway.

The No 2 (right) engine showed no evidence of fire.

Fragments of engine fan blades from No 1 engine were found under the approach path, up to 31/2 km east of the crash site.

The airframe wreckage was progressively removed from the site and transported to AAIB Farnborough during the period from 10-14 January and the engines were transported directly to the manufacturer, CFMI at Villaroche in France on 13-14 January where they were the subject of detailed strip down examination, under AAIB supervision.

FDR and CVR evidence

The aircraft was equipped with a Sundstrand Universal Flight Data Recorder and a Fairchild A100 Cockpit Voice Recorder. Both were replayed satisfactorily at AAIB. The FDR read-out established that, as the aircraft was approaching FL 290, an event occurred which led to the No 1 (left) engine recording its maximum indicated vibration level ("5 units"), recordings of rapidly fluctuating fan speeds and fluctuating HP core speeds, with an associated rise in exhaust gas temperature (EGT) and fluctuation in fuel flow. Approximately 20 seconds after this occurrence, the No 2 (right) engine was throttled back, at which point the No 1 engine fan and core speeds settled down, although at slightly different values, with the indicated vibration level remaining at maximum and EGT markedly higher than before - ie 840°C, compared to 780°C previously. Prior to being throttled back, the No 2 engine had been operating with steady engine indications, with a fan speed (N1) of 99%, HP compressor speed (N2) of 96%, EGT of 770°C and low vibration level ("0.5 units"). 1 minute and 47 seconds later, the No 2 engine was shut-down. It was apparent that the flight crew were aware of smoke on the flight deck and thought that they had a possible engine fire. There was no fire warning at this time.

The flight proceeded under the control of London, Manchester, and finally Castle Donington ATC. The No 1 engine appeared to respond reasonably to applied throttle demands, although the engine was at "flight idle" for a considerable time (10 minutes).

At about 900 ft above ground level on the final approach, the No 1 engine fan speed dropped rapidly and EGT increased significantly with other engine parameters unchanged, and a maximum indicated vibration level. The No 1 engine lost considerable power and some 17 seconds later the fire warning bell sounded. About 36 seconds later the aircraft impacted with the ground.

Engine strip examination

Inspection of the fan assembly of the No 1 (left) engine showed extensive damage had occurred to the titanium alloy blades, with many associated damage-induced overstressing failures. One fan blade was found to have fractured outboard of the "mid-span shrouds", due a progressive fatigue failure originating near the leading edge of this blade, adjacent the pressure face. This failure had released the outer "panel" (Approximately 4 inches) of this blade. The cause of this fatigue failure, which was the only instance of fatigue fracture found amongst the fan blades, is being pursued.

The abradable seal material which surrounds the fan assembly would have been progressively removed by the damaged fan assembly, as would the rubberised seals surrounding the low pressure compressor "booster" section, due to out-of-balance running, and would have led to associated smoke products entering the air conditioning system.

Inspection of the No 1 engine revealed evidence consistent with the anticipated effects of this engine having run under severe out-of-balance conditions due to fan damage, with some damage to the high pressure compressor arising from fan debris ingestion. No failures were found within the coremodule rotating assemblies.

Investigation of the source(s) and development of the airborne fire on the No 1 engine is still in progress. Evidence found to date is provisionally indicative of two areas of fire, one of which was located around the upper/outboard region of the exterior of the fan casing and the other which appears to have trailed from the thrust reverser duct on the left side. It is considered that the sources of both areas of fire were secondary to, and were induced by, the primary engine failure and subsequent continued operation with attendant high vibration.

Detailed investigation of the No 1 engine and its operating history will continue in an attempt to identify the cause of the fan blade fatigue and to explain fully the initiation and progression of the fire.

The No 2 (right) engine has been fully stripped and shows no evidence of pre-crash fire of failure.

Systems examination

Checks have been made of the wiring leading to the Engine Indicating System (EIS). These have confirmed that the indications of both primary and secondary engine parameters were displayed in the correct sense. The primary EIS display unit passed a full function and calibration check. The secondary EIS unit exhibited an obvious fault condition affecting the display of oil pressure, hydraulic pressure and oil quantity, but not the engine vibration indicators. The secondary EIS unit had suffered significant impact damage and the fault appears consistent with such damage. Since the Flight Data Recorder receives engine data (except vibration) from the EIS output to its display, it is implicit that such data was displayed by the EIS.

The Airborne Vibration Monitor unit (AVM), which feeds vibration signals from the engines to the EIS and FDR, was subjected to a full test schedule and, despite some minor external case damage, fully met the acceptance requirements. Both engine vibration indicators operated satisfactorily. The engine vibration indicators and FDR cannot register vibration levels higher than "5 units", since the AVM limits its output.

The Engine and APU fire detection module was severely damaged by impact, such that a functional test was not possible, and will require detailed inspection. Checks of the actual detector loops on the engines concluded that those not damaged beyond meaningful test by impact were capable of providing both overheat and fire detection warnings. The Engine and APU fire suppression bottles were found fully charged. There were no indications that an attempt had been made to discharge any extinguisher.

Further investigation of the systems aspects will include an appraisal of the EIS/Flight crew interface to evaluate the effectiveness of the presentation of engine indications to pilots. This evaluation will be assisted by the RAF Institute of Aviation Medicine Flight Skills Section.

Survival factors

A full evaluation of passenger and crew testimony and injury, combined with an analysis of the pathology, is being progressed. This information will be related to the crash-induced damage to the cabin structure and seating in order that the causes of injury are identified, with a view towards consideration of what improvements may be indicated to reduce injuries and maximise survivability in future accidents.

AAIB Safety Recommendations

Shortly after the accident, on the 11 January 1989, AAIB made 2 Safety Recommendations to the CAA. These were made at a stage in the investigation when it was known that the left engine had failed in-flight, but the reason for the in-flight shut-down of the right engine had not yet been established. In addition, it was considered prudent to address the possibility of a defective engine vibration or fire/overheat warning system although no evidence of such defect(s) had been found. The following precautionary Safety Recommendations were made to the CAA:

1.	That the CAA consider increasing the frequency of existing engine inspections and engine health monitoring on Boeing 737-300 and Boeing 737-400 aircraft until the causes of the engine failure(s) are established.
2.	The CAA call for an examination of the Boeing 737-300 and Boeing 737-400 engine Fire/Overheat and Vibration monitoring circuitry for left/right engine sense.

As a result, the CAA issued letters to owners/operators Nos 905 and 906 on 11 January 1989 which required (respectively) testing of the engine overheat/fire warning and vibration monitoring systems for "correct-sense" operation; and increased frequency of certain engine "health monitoring" checks on Boeing 737-300, 737-400 and Airbus 320 aircraft.

On 10 February 1989, the AAIB advised the CAA that the left engine had suffered a fatigue-failure of a fan blade and that there was no continuing justification for the increased inspections of the "oil-wetted" components of the CFM 65-3 and -5 engines. However, in view of the fan blade failure (the cause of which has not thus far been established) and AAIB caution concerning any possible means by which the fatigue strength of such fan blades may be inadvertently compromised, the following 2 Safety Recommendations were made to the CAA:

3.	The Civil Aviation Authority, in conjunction with the engine manufacturer, consider instituting inspection procedures for the examination of the fan stage of CFM 56 engines to ensure the early detection of damage that could lead to the failure of a blade.
4.	The Civil Aviation Authority review the advice given in the Boeing 737-400 Maintenance Manual concerning the excessive generation of heat during blending operations with power grinding and blending tools.

CFMI and Boeing issued letters to operators, emphasising the daily visual check on the engine inlet and fan blades and the detailed fan blade inspection at the aircraft "B check" (approximately every 750 hours). In addition, operators were recommended to review their policies and instructions for the maintenance and repair of CFM 56 fan blades, with particular emphasis on adhering strictly to the limits and procedures detailed in the aircraft maintenance manual for fan blade repair.

On 23 February 1989, the AAIB made a further 3 Safety Recommendations to the CAA:

5.	The CAA take action to advise pilots of Boeing 737-300/400 aircraft, and of other types with engines which have similar characteristics, that when instances of engine-induced high vibration occur, they may be accompanied by associated smoke and/or smells of burning entering the flight deck and/or cabin through the air-conditioning system, due merely to blade tip contact between fan/compressor rotating assemblies and the associated abradable seals.
6.	The CAA request the Boeing Airplane Company to produce amendments to the existing aircraft flight manuals and checklists to indicate what actions should be taken when engine-induced high vibration occurs, accompanied by smoke and/or the smell of burning entering the flight deck and/or cabin.
7.	The CAA review the current attitude of pilots to the engine vibration indicators on Boeing 737-300/400 aircraft and other applicable types with turbo-fan engines, with a view towards providing flight crews with an indication of the pertinence of such vibration instruments when certain engine malfunctions or failures occur.

This page was copied from:	http://www.open.gov.uk/aaib/gobme/a6.htm
COPY! Last modification on 1999-06-15 by Michael Blume