Draft findings from a Federal Aviation Administration study into pilots' relationship with airliner flightdecks are expected to argue that flightcrew are not properly trained for modern cockpits, and that there is a need for radical change in their recurrent training.
The FAA report, due to be published later this year, is expected to establish a connection between accidents and inadequate training, identifying specific areas in which there is a need for change to pilot training, airlines' standard operating procedures, and even the design of interfaces between pilots and automated systems.
 |
|---|
© CAEFlightdecks have changed radically over the past 40 years but training has not adapted to reflect that |
For some years, there has been wide industry agreement that the regulations defining airline pilot recurrent training requirements need a review. After all, flightdecks have changed radically over the past 40 years, as has the way in which pilots manage normal flight operations - but training has not changed to reflect this.
FAA human factors expert Dr Kathy Abbott is leading a team carrying out a study provisionally entitled Operational use of flightpath management systems. Acknowledging the capability of today's sophisticated flight management systems, the study examines how successful pilots are at using them and the effect they have on pilot performance overall.
Abbott's mission is to "review the operational use of, and training for, on-board systems for flightpath management". She presented her initial findings from the ongoing study at the Flight Safety Foundation international aviation safety seminar held on 3-5 November 2010 in Milan.
At the same event, four aviation industry heavyweights from Airbus, Boeing and the US Airline Pilots Association presented on subjects that acknowledged a worrying decline in line pilots' basic flying skills. There were three presentations on stalling and stall recovery and an entire presentation on the art of the go-around. Both subjects were topical because of recent stalling and go-around events in which pilots demonstrated insufficient knowledge and skill to handle them safely.
Stall recovery was addressed by Boeing's Dave Carbaugh, Airbus's Claude Lelaie and ALPA's top human factors expert Capt David McKenney. Then Air France corporate safety manager Bertrand de Courville dealt with the art of safe go-arounds.
The ability to carry out safe stall recovery and go-arounds is fundamental to basic pilot competence, so the need to cover them in such depth at one of the world's main forums for presenting safety policy suggests that airline recurrent training is not addressing the basics.
And now Abbott reveals that training is also failing to impart skills for managing advanced automation, suggesting that training at many airlines is deficient on all counts when it comes to ensuring that pilots gain and retain the skills needed for the job.
The evidence gathered by Abbott suggests, among other things, that pilots concentrate on programming the automation at the expense of monitoring the flightpath. But training policy cannot be changed unless there is hard evidence that change is required, and Abbott's team may be able to provide data that establishes the relationship between today's pilot training and actual pilot performance on the line.
In 1996 the FAA published a landmark study called The interfaces between flightcrews and modern flightdeck systems. Since then, however, not only has automation advanced, but pilots have been progressively conditioned by its existence. The need for the study that Abbott is now leading was recognised by the US Performance-Based Operations Aviation Rulemaking Committee and the Commercial Aviation Safety Team, and it is to those organisations and the FAA that she will submit the final draft.
The provisional findings Abbott presented in Milan came from analysis of accident and incident data and line operation safety audits (LOSA) in the period 2001-09. The study took account of data from 734 events from the Aviation Safety Reporting System, and aggregated LOSA data from 9,165 flights. Abbott adds the caveat that she is presenting raw information at this point, and there is more work yet to do to understand it fully.
The information so far gained suggests that flightcrew have never been properly trained for operating highly automated aircraft and that there are no checklists for many of the problems they have to deal with from time to time, leaving the pilots to manage using ingenuity and airmanship.
ACCIDENTS AND SERIOUS INCIDENTS
Inadequate crew knowledge of automated systems was a factor in more than 40% of accidents and 30% of serious incidents between 2001 and 2009, Abbott found, while recurring handling problems include lack of recognition of autopilot or autothrottle disconnect; inadequate monitoring of flight and navigation instruments, and associated failure to maintain aircraft energy and/or speed; incorrect upset recovery technique; and inappropriate control inputs, and dual sidestick inputs.
Regarding flight-management-system use, Abbott discovered that pilots frequently focused on programming the FMS to the detriment of monitoring the flightpath. And pilots have to deal with many failures for which there are no checklists, and for which there is no training of any kind, such as failures or malfunctions of air-data computers; computer or software failures; electrical failures; and uncommanded autopilot disconnects or pitch-up incidents for which the reason is unknown. "Failure assessment is difficult, failure recovery is difficult, and the failure modes were not anticipated by the designers," Abbott explains.
Despite the sometimes fickle nature of the automation, pilots frequently abdicate too much responsibility to automated systems, she observes. Among the reasons for this is a perceived lack of trust in pilots' manual flying performance by the airline.
It may also reflect policies that encourage use of automated systems rather than manual operations, or a lack of training, experience or judgment that means "pilots may not be prepared to handle non-routine situations".
Abbott has uncovered evidence of particular vulnerabilities in automated systems and their man/machine interfaces, such as mode confusion and a pilot tendency to use information from automated systems instead of raw data.
Other problems she has identified are that much of the information supplied to pilots is itself automated and that there is no consistency among operators in their policies for the use of automated systems or the information they provide.
Pilot knowledge was found to be lacking in many areas relating to automated systems. This showed itself in line performance shortcomings, such as poor understanding of the capabilities and limitations of the flight director, autopilot, autothrottle/autothrust, flight management system and flight guidance computer systems.
POOR KNOWLEDGE
Pilots also demonstrate poor knowledge of system operating procedures, mode transitions and associated system behaviour; and unusual attitude recognition and recovery.
Abbott's terms of reference for the study stipulated that her team should make recommendations to overcome any identified shortcomings. Change is needed across the board, she implies, predicting recommendations will be forthcoming in operations, training/qualification, equipment design, organisational factors, safety data collection and analysis and future operations. There may be others, Abbott suggests.
Probable operational recommendations, she predicts, may include changes to training and airline standard operating procedures to focus more effectively on developing and assisting flightpath management techniques and skills.
Training should distinguish between guidance and control; flight crews should be encouraged to tell air traffic "unable to comply" when appropriate; and individual operators should ensure that standard operating procedures are tailored to their specific needs.
The study indicates that industry as a whole needs to review practice, regulatory guidance and requirements for training in numerous areas. These include flightpath and energy management, recovery from off-path circumstances, use of alternative modes to meet air traffic clearances/requirements, airlines' operational policies, and the management of malfunctions.
Former United Airlines human factors and command development manager Capt Michael Gillen has meanwhile conducted a study to determine whether advanced automation and "glass cockpits" are causing pilot skills, particularly in instrument flying, to degrade.
 |
|---|
© ThalesPilot knowledge was found to be lacking in many areas related to automated systems |
A group of 30 pilots from a major US carrier - split between long-haul widebody and short-haul narrowbody flightcrew - performed five manoeuvres using raw instrument data only. The group had all been flying "glass cockpit" aircraft for a number of years, 73% of them for 10 years or more. Some had plenty of experience in classic cockpits, but 43% had only two years or less flying non-glass flightdecks.
The manoeuvres performed for the test were take-off, engine failure at V1, holding pattern, instrument landing system approach and missed approach. Before the test, the pilots were all questioned on whether they believed their instrument skills have declined. Most believed they had degraded, but that they were still good enough.
Presenting the results of his study in Milan, Gillen revealed that the average group score for all the manoeuvres was below that required for an FAA pass, some manoeuvres well below. The short-haul pilots generally scored higher than the long-haul operators. The highest scores were for take-off and missed approach - but even those were still below the FAA pass level. The lowest scores were for the holding pattern.
"A majority of the manoeuvres were flown at or below the proficiency standard required for basic instrument certification grade," Gillen noted - and all were less than that required to win an instrument rating at air transport pilot licence level.
Summing up, Gillen said that the study found that "professional pilots have suffered a significant decline in their basic instrument skills and, in general, pilots are not aware of this skill degradation".
The level of concern for pilot skills among senior Airbus, Boeing and ALPA training experts, the findings that Abbott's report is soon expected to reveal, and Gillen's study all point in the same direction: airline recurrent training needs radical change.
WHEN THE GOING GETS TOUGH
When US investigators reported on the crash of a Colgan Air Bombardier Q400 at Buffalo, New York in February 2009, it became clear that there was a gap between the skills provided by the pilot training that US licensing laws require and what aviators really need when things get tough.
The Q400 crew had reacted wrongly to the onset of a stall - something most pilots believe they would never do - and the aircraft went out of control, crashing with the loss of all on board.
Airbus, Boeing and the US Air Line Pilots Association delivered three separate presentations on the subject of stall recovery technique at the Flight Safety Foundation international aviation safety seminar in Milan in November 2010, indicating the level of concern about basic piloting skills. And Airbus's presentation on safe technique for go-arounds reflected industry concern at the number of serious incidents and fatal accidents that have followed some go-arounds in the past decade.
Earlier in the year, at the EBACE business-aviation conference in Geneva, Bombardier had conducted a two-day European version of its much-respected Safety Standdown for pilots, and the subject of stall recovery was addressed at the associated Advanced Aerodynamics Workshop. The workshop revealed that the stall-recovery drill commercial airline aircrew are taught is different from the way test pilots manage stall recovery when they are certificating the aircraft that the airlines fly.
 |
|---|
The Q400 crew of the Colgan Air Bombardier Q400 that crashed in Buffalo, New York in February 2009 had reacted wrongly to the onset of a stall |
Federal Aviation Administration test pilots conduct stall recovery the classic way: push the control column forward to unload the wing and gain airspeed and, when the wing is no longer stalled, apply power until a safe airspeed is attained and the resulting loss of height may be recovered. The technique most US carriers teach, the Advanced Aerodynamics Workshop heard, is to act at the stall warning, maintaining pitch while simultaneously applying power. No-one at Geneva was certain as to when and why this technique came to be adopted in place of the classic stall recovery.
One suggestion was that it might have dated from the 1950s, because it would have worked with piston-powered aircraft, where power application would immediately provide propeller-driven slipstream over the wing, and where the opposing power and drag vectors act through much the same point.
In modern jet aircraft there is no extra slipstream over the wing when power is applied. The power vector is acting below the wing and the drag vector is slightly above it, leading to a natural pitch-up moment on power application, which can upset the stall recovery - especially if the wing has not been properly unloaded.
Source: Flight International
