|
|
|
Accident reviews on this page:
Engine Failure Related Accidents
Accidents after engine failure continue to happen. A list with more than 500 reported accidents with multi-engine airplanes since Jan. 1996 can be downloaded here; more than 4,000 people lost their lives during these unnecessary accidents. Unnecessary? Yes, most of them definitely, because if pilots would have been made aware of the limitations that apply for controlling their airplanes while an engine is inoperative and the power setting of the remaining engine is high, these accidents would not have happened.
Regrettably, flight manuals of most multi-engine airplanes and courseware for a multi-engine rating present definitions of the minimum control speed VMC(A) and engine emergency procedures that are not in agreement anymore with airplane design methods and flight-test techniques.
AvioConsult reviewed more than 400 accident investigation reports and concluded that none of them presented the real cause of the accidents and hence did not recommend the appropriate improvements to really improve aviation safety, see below. It became clear that neither the mishap pilots nor the accident investigators and in some cases not even the writers of airplane flight manuals and courseware knew about the real value and the limitations of air minimum control speed VMCA (and of takeoff safety speed V2) for the controllability and performance of their airplane following an engine failure or while an engine was inoperative. A number of training manuals were also reviewed and are also briefly discussed.
The sole purpose of publishing the comments
below is to help prevent accidents after engine failure from happening
again. Please feel free to read or download and learn from these
comments. Please do not hesitate to ask any questions remaining.
All of the theory that is needed to understand the comments presented below
can be found in the Paper Airplane Control and
Accident Investigation after Engine Failure by AvioConsult.
Please refer to the Downloads page.
Formal FAA and EASA documents, Test Pilot School course books and references
to Aeronautical Universities can be found in the References table on the
Downloads page and on the Links page.
On the Downloads page, free papers explaining the controllability and performance while one of the engines of a multi-engine airplane is inoperative are presented, as well as a link to an interesting on-line One Engine Inoperative Trainer of the University of North Dakota. Click here for direct access to this trainer.
|
Please read the copyright and liability disclaimers
here first. |
Reviews of accident investigation reports
A limited number of investigation reports were reviewed by AvioConsult. Please refer to the links in the list in the left column, or browse further down.
If you would like an accident to be reviewed, please don't hesitate to ask.
If you would like to learn how to control an airplane after engine failure, or how to analyze engine failure related accidents using Flight Data Recorder data, click here.
Textron Aviation B-300, KADS, 30 June 2019
After liftoff, power of the left engine decreased. The resulting yaw was not adequately counteracted and the airplane continued to roll to the left and impacted an airport hangar. All 10 crew and pax were fatally injured. Inadequate rudder to maintain straight flight and not maintaining a favorable bank angle of 5° into the good engine increased VMCA to a lot higher value than the POH-published VMCA; the large resulting sideslip caused much drag preventing the airspeed from increasing to V2. The actual VMCA increased above the calibrated airspeed, after which control was lost.
The NTSB reported that the pilot initially commanded a left rudder input, which was the conclusion of calculating the sideslip with only one of three lat-dir simultaneous equations of motion, which is not a good engineering practice, because other sources of sideslip were left out.
It is hard to believe that a pilot would apply rudder into the yaw due to asymmetrical thrust at low altitude, thus enhancing the yaw rate.
AvioConsult was asked to write a post on the crash on the BeechTalk/ Crash Talk forum, subject: Plane Down in KADS...(page 77, 21 July 2019). Click here or download the pdf version below.
In anticipation of more questions, AvioConsult reviewed NTSB studies in the accident docket as they appeared and wrote notes to the NTSB Lead Investigator to help improve the final report, because engineering errors were made in calculating the sideslip. In addition, the engine emergency procedure in the B-300 POH was not quite complete. Regrettably, these notes were not used. The pilot was blamed, while his training and the FAA review of the POH were obviously inadequate. In addition, the NTSB report was inadequate as well (as most, if not all engine failure accident reports).
Download the annotated studies and the notes sent to the NTSB in one document.
Another post was written and posted on the BeechTalk forum on 24 Nov. 2022, page 111.
Download a pdf with both posts on the BeechTalk Forum | Top
Incident with Boeing 747-400 ERF, Maastricht Aachen Airport, 17 nov. 2017
When takeoff thrust was being set, the thrust of engine #4 suddenly reduced to 20%, most probably due to a compressor stall. The airspeed was about 30 kt. The captain, as pilot flying, used nose wheel steering, differential braking and even reverse thrust trying to prevent the airplane from vacating the runway. He of course, was not successful. The throttles were not closed at once and the airplane was not brought to an immediate stop when diverting from the runway centerline. The errors by the crew raised doubts about their training.
The Dutch TSB reported that "there are no certification requirements for situations in which a critical engine becomes inoperative during takeoff below VMCG". Well, there are, as (nearly) all multi-engine rated pilots know: the certification requirement is that the minimum control speed on the ground (VMCG) must be determined and decision speed V1 made available. If any engine fails during takeoff when VMCG or V1 has not yet been reached, the takeoff must be aborted expeditiously. The report by the Dutch TSB is really embarrasing, therefore, as an apology to readers worldwide, AvioConsult annotated the original report and makes it available for download.
Download the original annotated report
| Top
A limited review in the Dutch language is available here.
DC-10, Faro Portugal, 21 Dec. 1992
At the request of 30+ victims and next of kin of the deceased of this fatal accident, AvioConsult critically reviewed the formal accident investigation report and the contribution of the Dutch Aviation Safety Board as accredited representative.
This resulted in the involvement as an expert witness in two court cases that ended 8 Jan 2020 with the verdict that the cause of the accident, that took the lives of 57, was not windshear, as was 'determined' by the Dutch Board. The court passed sentence that the crew did not follow procedures and made several errors. The initiated go-around just before touchdown failed because the engines rpm had decreased to flight idle as the throttles were already closed and held back at 150 ft against the fully functioning autothrottle system.
The detailed analysis of the last 80 seconds of flight, written by AvioConsult using flight data recorder and other data, and experimental flight-test expertise, is available for download.
Download
The Last 80 Seconds of Flight MP495 (2017).
The first report, written in 2012 with very limited means and data, includes the dubious role of the Dutch Aviation Safety Board and is also available for download.
Download Analysis DC-10 Faro accident, version 2012, revised 2017.
A few more downloads:
The Portuguese accident report (7 MB).
The reports of the National Aerospace Laboratory (NLR) and Review (12 MB).
The AIR Inc. en Gillespie reports and Review (7 MB).
Review of the Draft (Interim) Report V17 of Court-experts (2016) (5,6 MB).
Final Report of Court-experts with annotated Comments (2017) (3,5 MB).
Review of the Final Report of Court-experts (2017) (1,5 MB).
These reports prove that experienced airline pilots may be very well educated, trained and experienced to operate airplanes, but should — without proper engineering education — never be assigned as (independent or sole) accident investigator. Analyzing FDR data requires airplane systems knowledge as well, as became clear in the report The Last 80 Seconds ... referenced above. Windshear was not the cause of this accident, but numerous consecutive pilot errors were.
Other data are available on the Dutch downloads page | Top
Britten Norman BN-2, Bonaire, 22 Oct. 2009
A BN-2 ditched very close to the destination island Bonaire (Caribean) following
an engine failure shortly after takeoff from Curaçao. The passengers could escape and were rescued;
the pilot regrettably died in the accident. The accident investigation was conducted by the
Dutch Safety Board.
It seems that the pilot was never made
aware of the technique to achieve maximum performance (minimum drag) while
one engine is inoperative, most probably because neither his training
manuals, nor the BN-2 flight manual did provide this
life-saving information. The accident investigators were obviously not
aware of this either. Refer to the Downloads
page for papers on this subject.
The accident investigation report with many comments, written by AvioConsult using experimental flight-test expertise, is available for download:
Download the Safety Board Report with comments of AvioConsult | Top
Lockheed C-130H, Netherlands, 15 July 1996
A C-130H Hercules crashed at Eindhoven Airbase in the Netherlands
following the failure of engines #1 and #2 due to bird ingestion
during a go-around that was initiated just prior to touchdown. Engine #3 was
was shutdown by the crew, either before the approach or just
before #1 and #2 failed.
The go-around was initiated at an airspeed below 97 kt, which
is far below the flight manual required go-around airspeed of VMCA2 (134 kt in
this case). The flight manual requires acceleration to VMCA2
(VMCA for n - 2) before selecting full go-around power to be
able to maintain control and in anticipation of another engine to fail on the same wing. If two engines fail on the same
wing and the power setting on the other engines is high, the airplane will be uncontrollable at speeds below VMCA2.
In this case, the airplane lost the thrust of three engines (# 1, 2
and 3). Only #4 was operating at max. thrust following the
bird ingestion. In this very special case, the actual minimum control
speed, however, was VMCA1 (one asymetrical outboard engine). As the airspeed was a
little below VMCA1, the airplane started to slowly roll and
slide away in a direction away from the operating engine, until it impacted the ground.
The pilots of the airplane and the accident investigators did not know about the real value of the minimum control speeds of the airplane, not about the factors that influence VMCA, not how the magnitude of VMCA can be 'controlled' by the pilots and not what VMCA really means for the controllability and safety of flight before and after engine failure, despite the fact that Lockheed provided good VMCA data and explanations, including control limitations, in the airplane flight manual as well as in a very good publication 'C-130 low speed flying qualities' that is available to all C-130 pilots.
The Flight and Performance Manuals of the C-130H airplane present numerous warnings, cautions and notes on propulsion system malfunctions, explain the reduced controllability after engine failure and present recommended flight techniques and the consequences if these are not adhered to. Lockheed did a good job, but the warnings, cautions and notes are regrettably not understood by most pilots.
Cause of the accident
The cause of the accident was that the pilots
were not (made) aware of the go-around speed warnings in the Flight Manual.
They initiated a go-around at a speed 37 knots(!) below the Flight
Manual required go-around speed when an engine fails or is inoperative.
The Dutch Aviation Safety Board refused to
re-investigate the accident following the publication of a book
on the accident, at the request of the survivors and relatives of the
victims (2010).
Download the analysis of the accident (Dutch language) here | Top
Cessna 404, Australia, 11 Aug. 2003
Shortly after liftoff, while still over the runway, the right engine
failed. The pilot retracted the landing gear, selected flaps up
and feathered the right propeller and then, at very low altitude,
turned left, into the operating engine, to return for landing.
One of the conclusions in the report was: The aircraft was
manoeuvred, including turns and banks, at low altitude resulting in a decrease
in airspeed below that required to maximise one-engine inoperative performance.
While on the left hand downwind
leg, the airspeed decayed, the altitude could not be
maintained and the airplane descended and impacted with the
ground. Weight and cg were within limits.
Cause of the accident
The pilot did not maintain 5 degrees of bank after engine failure
as required in step 6 in the engine failure procedure in the flight
manual, but initiated a turn at too low an altitude (100 ft AGL) while the airspeed was too low. Control was
not lost immediately, because the turn was in the direction of the operating
engine, which is the favorable direction for keeping actual VMCA
low, i.e. below the published VMCA.
The accident investigation report did not include the effect of bank angle on VMCA and the necessity for a 5 degree bank angle into the good engine as a life-saving factor that influences both the controllability and the one engine inoperative climb performance. Pilot and accident investigators were obviously not familiar with the effect of bank angle on both VMCA and the airplane performance, may be because it was never taught to them.
A supplemental analysis to the formal accident investigation report, written by AvioConsult,
using experimental flight-test expertise, is available for download:
Download this analysis | Top
Beech C90, Australia, 27 Nov. 2001
Just prior to, or at about the time the aircraft became airborne, the left engine failed. After liftoff, the aircraft remained airborne for about 20 seconds. The aircraft was rolling through about 90 degrees left bank, it struck power lines about 10 m above ground level and about 560 m beyond the end of the runway. It then continued to roll left and impacted the ground inverted in a steep nose-low attitude.
Cause of the accident
The accident was caused because of an
incomplete and deficient engine emergency procedure. In
addition, the real value and meaning of VMCA was neither
clear to the pilot, nor to the accident investigators.
AvioConsult wrote a letter to the CEO of Raytheon Aircraft Company on 8 August 2006 expressing concerns about the definitions and the engine emergency procedures in the operator manual of the Beech King Air C90 and to present recommendations to improve. Ratheon never responded.
A supplemental analysis to the formal accident investigation report, written by AvioConsult,
using experimental flight-test expertise, is available for download:
Download this analysis | Top
Boeing 737-200, Algeria, 6 March 2003
Just after passing V1, an engine failed. Almost immediately after
liftoff, control of the airplane was lost and the airplane crashed,
killing all but one on-board.
AvioConsult reviewed the accident investigation report and concluded
that the engine emergency procedures in the Flight Crew Training
Manual were inappropriate because the recommended control inputs were
not leading to the 10% airspeed safety margin over VMCA that aviation regulations require
when an engine fails during takeoff.
When using the recommended control inputs (for keeping the wings
level), the actual minimum control
speed that the airplane experiences right after liftoff, will be much
higher than the minimum control speed that was determined during
flight-testing and that is used to calculate both VR and V2.Actual
VMCA was most probably even a little higher than V2
at the moment of liftoff; because the airplane continued a slow roll
into the dead engine.
Cause of the accident
The accident was caused by inappropriate engine emergency procedures. The
pilots are not to be held responsible.
AvioConsult recommended Boeing in July 2005 to improve the procedures, but Boeing responded that 'there was no compelling reason to change the procedures'.
An analysis of the Engine Failure Takeoff Procedures in the Boeing 737-200/300/400 Flight Crew Training Manual
is available for download:
Download
this analysis | Top
Piper PA-31, New Zealand, 17 Dec 2002
Shortly after takeoff, the left engine
quit operating for unknown reasons. The pilot feathered the
propeller and returned to the airport for landing.
During the final turn, control was lost and the airplane crashed.
The analysis of this accident is included in the paper Airplane Control and Accident Investigation after Engine Failure, please refer to the Downloads page. | Top
Piper PA-31P, Australia, 15 June 2010
While passing 7000 ft, the right engine
failed and was shutdown. The airplane returned to the departure airport, but
crash landed a few miles short of the airport on a road.
Download the original report with comments added
Piper PA-44-180, Netherlands, 14 August 2002
A Piper PA-44-180 Seminole, owned and operated by Martinair Flight School in The Netherlands, crashed in a lake, killing an instructor and two students during a demonstration of flight with an inoperative engine. The Dutch Transport Safety Board thoroughly investigated the accident and concluded that following the intentional shut down of the left engine, the fuel valve of the right engine was inadvertently closed rather than the valve of the left engine, after which the right engine quit as well and an emergency landing became unavoidable, according to the report. The report also concludes that the airspeed decreased below the stall speed, after which control of the airplane was lost at an altitude from which recovery was not possible.
But to the opinion of AvioConsult, this was not the cause of the accident. Neither the pilots nor the investigators did know about the limitations of a minimum control speed. A VMCA, although never determined, also exists due to asymmetrical drag (yawing moment) caused by one feathered propeller and the other not feathered.
The analysis (in Dutch language) with comments on the formal accident
investigation report is available for download:
Download this analysis |
Top
Documentation review PA-44-180. The formal accident investigation report did not report on the training documentation used by the flight school. AvioConsult therefore asked, and received permission, to review the airplane and training documents. Many errors and deficiencies were found.
A limited analysis of the PA-44 documents, as used by the flight
school (and also by other PA-44 owners) in the English language, is
available for download:
Download this analysis | Top
This analysis will also be useful to operators of other multi-engine airplane types.
AvioConsult wrote a letter to the CEO of The New Piper Aircraft, Inc.
on 9 August 2006 expressing concerns about the imperfections in the
Pilot's Information Manuals of Piper Aircraft and to present
recommendations for improvement.
The company never responded.
Mitsubishi MU-2B, USA, 10 Dec. 2004
Shortly after takeoff, the left engine failed. The pilot returned for landing via a left-hand circuit; the left propeller was feathered. The airplane did overshoot the final approach of runway 35R and was cleared to the next runway 28. The landing lights were then seen turning down toward the terrain. The airplane crashed; the two souls onboard were fatally injured.
The analysis of this accident is included in the paper Airplane Control and Accident Investigation after Engine Failure, please refer to the Downloads page.
Boeing 747-258F, Netherlands, 4 October 1992
Shortly after takeoff, both engines on the right wing (#3 and #4) separated from the
wing due to a fuse pin failure in pylon #3. The pilots decided to return to
the airport immediately and initiated a right hand turn. The asymmetrical loss of weight of engines and pylons could obviously be compensated for by the ailerons, loss of lift of the damaged right wing and a high enough airspeed. During the second right hand turn, required to position at a longer distance from the requested landing runway, the throttles werd advanced to (near) full thrust while the bank angle was 25°, upon which lateral control of the airplane was lost and the airplane crashed in a residential area in Amsterdam. During the turn, the actual VMCA2, the minimum control speed if two engines are inoperative, had obviously increased above the Indicated Airspeed (260 kt).
In civil aviation regulations (part 25.149), only VMCA exists (for one engine inoperative). VMCA2 was deleted from the civil regulations, but still exists in military regulations, and definitely also in-flight, and causes catastrophes if not observed, like in this accident.
The formal public accident investigation report (No. 92-11) does neither discuss the air minimum control speed for one inoperative engine (VMCA) nor for two engines inoperative (VMCA2).
Why were the pilots and the accident investigators not aware of the real value of these minimum control speeds? And why didn't they know that turning away from straight flight, while maintaining a small bank angle away from the inoperative engine(s) as required for VMCA's to be valid, increases the actual VMCA considerably and becomes almost certainly deadly when the power setting on the remaining engines is (increased to) maximum? Most probably because the engine emergency procedures and VMCA definitions in airplane flight and operating manuals are incorrect or inappropriate / deficient, or because the engine-out pilot and investigator training is inappropriate and incomplete. It seems that the manuals- and training-approving-authorities were not aware either. Test Pilot Schools though, still teach engine-out flight in theory and in-flight; experimental test pilots are thoroughly made aware of the do's and don'ts after engine failure and do not crash.
Refer to the Paper Airplane Control and Accident Investigation after Engine Failure that AvioConsult presents on the Downloads page of this website to learn about improving procedures, definitions and engine-out training.
The huge effect of bank angle on VMCA and VMCA2
was neither considered by the pilots, nor by the accident investigators. They might have never been taught that an airplane is only certified for straight flight when one or more engines are inoperative while the airspeed is low, and the thrust setting on the opposite engine(s) is (increased to) maximum. And that the actual VMCA, i.e. the VMCA that they will encounter in-flight, will increase considerable when banking. In addition, a bank angle larger than 10° into the operating engine(s) may lead to a fin stall, causing the loss of directional control, if the rudder is not reversed and the speed is too low. The effect of bank angle is calculated and illustrated in
Paper The Effect of Bank Angle and Weight on VMCA that is also available for download:
Download this paper | Top
Please also refer to the comments on the Boeing 737 FCTM above.
Saab SF-340B, Netherlands, 4 April 1994
Ten minutes after takeoff from runway
24, an oil pressure warning of the right engine (#2) made the captain
decide to return to the airport. He left the affected engine #2
idling; its propeller was not-feathered. The wind was 270/ 11 kt
when the captain accepted landing runway 06. On short final,
with an actual wind of 280/ 8 kt, the airplane was displaced
to the right. At 45 ft Radar Altitude, the captain
therefore decided to go-around using the thrust of the left engine
only; the right engine was kept idling. The airplane crashed 13 seconds later,
far to the right of the runway.
Download
this paper and learn about investigating engine failure related
accidents. | Top
The analysis of this accident is also included in the paper Airplane Control and Accident Investigation after Engine Failure, please refer to the Downloads page.
DHC-6-100 Twin Otter, USA, 29 July 2006
A De Havilland DHC-6-100 airplane carrying skydivers crashed into trees and terrain after takeoff. Witnesses at the airport reported that, shortly after the airplane lifted off from the runway, flames emitted from the airplane’s right engine.
The analysis of this accident is included in the paper #3 Airplane Control and Accident Investigation after Engine Failure, please refer to the Downloads page. | Top
EMB-120ER, Australia, 22 March 2010
During the takeoff of an EMB-120ER twin-turboprop airplane for a revalidating command instrument rating, the training and checking captain retarded the left power lever to zero torque just after liftoff. The airplane crashed, both pilots died in the accident.
The Safety Report was reviewed and many suggestions for improvement were included. This review is to learn more about airplane control after engine failure.
Download the safety report with review and learn more about investigating engine failure related accidents.
 |
The analysis of this accident is included in the YouTube video (downloads page) and in the paper Airplane Control and Accident Investigation after Engine Failure, please refer to the Downloads page. | Top |
Jetstream 4100, South Africa, 24 September 2009
During takeoff, at about the instant of rotation, 5 sec. before liftoff, the right (#2) engine failed; the takeoff was continued. Control was lost shortly after liftoff, the airplane crashed.
The thorough analysis of this accident on control after engine failure is included in the paper Airplane Control and Accident Investigation after Engine Failure, please refer to the Downloads page | Top
Cessna 310Q, UK, 15 November 2013
During final approach, a go-around was initiated. The left engine failed, after which control was lost. The airplane crashed, killing both persons on board.
The accident report was reviewed, and comments and recommendation for improvement of Airplane and Training Manuals are presented. | Top
Download the safety report with comments | Top
Email the URL of this website to a friend or colleague by clicking here.