Human factors and pilot error in Air France Flight 447

Air France Flight 447 Accident Analysis Through an Aviation Psychology Perspective

This paper addresses Air France Flight 447 accident analysis with an aviation psychology perspective. The accident is one of the biggest in the history of aviation. It is still being referenced today. This analysis' goal is providing a multidisciplinary platform for exchanging ideas and findings made in psychological and human factor research on what might have contributed to the disaster. It covers the roles played by human errors, organizational factors and individual factors in the accident. Theoretical as well as model based accident explanations are also given. Furthermore, current trends and requirements of automation concerning crew resource management, advanced automation and selection criteria for air flight safety are also considered in the paper.

Air France Flight AF447 Accident -- What Happened?

Air France Flight AF447 disappeared over the Atlantic during a stormy weather with 228 people on board consisting of passengers and crew. It was the morning of June 1, 2009 and the flight was headed to Paris from Rio de Janeiro (Stone, Keller, Kratzke & Strumpfer, 2013). There were 12 crew members and 216 passengers. The set departure time was 22h 00 and the crew was given clearance to start the engines at about 22h 10. The plane took off at 22h 29. Pilot Not Flying (PNF) was the captain and Pilot Flying (PF) was one of the copilots and junior officer (FO1 (BEA, 2012).

Four hours following their departure, the plane got into a turbulence zone whose intensity grew. The aircraft was to make contact with the Senegal ATC at 0020 UTC but no contact was registered and even in the waypoint that followed. Confusion abounds on the disappearance of the plane and Air France made an announcement about its disappearance on 1 June and a search and investigation was ordered. The sea search for bodies and parts span close to three years and BEA released the final accident report in 2012. The report indicated that temperatures outside the plane were -40 C. And the pilot tubes' design limits could not meet the conditions and they froze up thereby occasioning the failure of the Air Speed Indicators. Because most aircraft input systems are formed by ASI, the failures were successive and the initial warning of failure was the disconnection of the Auto Pilot (AP). The FO1 took control of the aircraft. He however could not correctly recognize the state of the situation and he tried climbing as he thought they might have been losing altitude. Most of the systems failed in succession giving several alarms at once and the plane was in an aerodynamic stall. Such a stall means the lift being produced by the aircraft isn't adequate and so the aircraft plunged into the ocean. No clues or traces were left behind.

Analyzing the Accident through Aviation Psychology

Someone may wonder why the airplane accident should be analyzed through psychological science. Human error accounts for between 60% and 8% of aviation accidents. Scientists are therefore turning to cognitive psychology -- working memory, error detection, as well as shifting attention, for instance - to help them understand the limitations of human performance in flight. The demand for reliability by society, particularly in aviation is unlimited. Volume of flights is growing steadily. Maintenance of the current rates of accidents will mean more fatal accidents in the next decade. It is important to understand what might have contributed to the errors that could have been made by skilled controllers, pilots, mechanics or dispatchers and so come up with more watertight controls to reduce the probability of such accidents in the future (Hoffman, 2015).

Why there is Need for an Inter-disciplinary Approach

The knowledge already garnered concerning the psychology of human beings can be applied in resolving the human factors that could have contributed to the accident just as they have been applied in other engineering fields. Coming up with the best way to move forward in aviation requires a mufti-faceted approach with players from several disciplines. This is the goal of this paper (Jorna, 2011).

Role of Organizational Factors on Aircrew Operations, Operational Effectiveness and Flight Safety

All aviation organizations are involved in one way or another. The organizations range from manufacturing organizations, corporate flight departments, airline operations departments to air traffic control. Contributing organization factors include corporate culture, management, organizational culture, recruitment, training and management. The organizations that have "utter probity" are those organizations that secure the best tools and equipment, use the equipment with intelligence and maintain them carefully. Further, they are honest and show responsibility to the aviation profession. The organizations embodying this principle can be considered organizations of "high integrity." The concepts in question are high performance and reliability. A high reliability organization focuses on having the least amount of accidents. Such organizations tolerance to errors is significantly low and consequences for making errors grave. A high performance organization focus is effectiveness. Instead of taking the multi-faceted approach taken by the high reliability organization, they make use of a critical measure. High-integrity organizations might possess common features including:

1.

Decisions are made based on the best available information.

2.

Processes that lead to decision making are open to scrutiny

3.

Personnel are put in an environment which promotes sound decision making and promotes critical thinking.

4.

Effort is made to adequately train personnel.

5.

Only individuals that are suited for carrying out of various tasks are allowed to do so.

6.

There is encouragement of imagination and ingenuity that help in the achievement of the goals of the organization (Wise, Hopkin & Garland, 2010).

Analysis of Organizational Factors in AF447 Accident

In the analysis of the role organizational factors played in the accident, some conclusions were made. An analysis of the major causing factors in major incidents and accidents, it has been discovered that human factors (synergy, decision making and situational awareness) had a role to play in 80% of the events (much higher than the role of organizational, technical and environmental factors), even where such factors can't be wished away as contributory. Factors like a lack of adequate training were noted in the operations of the flight crew and the decisions they make can directly affect the safety of the flight. Such weakness does not allow an organization to have an accurate picture of their performance as regards safety and so come up with viable solutions within the period (BEA, 2012).

Training: Even though the pilots had been adequately trained in Airspeed procedures, they couldn't recall them while in stressful situations. Training in stall recovery was existent only in lower flight levels. They relied heavily on automation and seldom practiced hand-flying. This was a major problem that made them incapable of correcting the situation fast. Also, CRM training they were given was inadequate as the pilots were unable to communicate the emergency. With better coordination and communication, there is probability that the flight could have been saved (Dhavala, 20130.

Design: the plane gave several warnings in succession and did not give a clue on what the problem could have been. Misleading stall warnings made the pilots take the wrong course of action when they were actually just doing fine. There was limited information concerning the plane's alternate mode and so the pilots assumed that the plane couldn't be stalled (Dhavala, 2013).

Operational Limitations: Air France had been alerted to the problem of malfunctioning tubes by other members of crew but not enough repair had been done. The flight was allowed to operate without being modified in any way (Dhavala, 2013).

Role of Individual Factors

Following a review of air line accidents that took place between the years 1967-1988, it was discovered that individual factors like sleep loss, chronic fatigue as well as desynchronosis were the human factors that contributed the most to the events. In several of the cases, they were as a result of bad work scheduling that made the crew work for long hours and that did not allow sufficient rest or sleep (Wise, Hopkin & Garland, 2010). Flying is a demanding, unforgiving and precise endeavor. Impairment to the ability of the pilot to perform at peak is possible cause for disaster. For instance, alcohol use is a stressor that should be done away with completely (Federal Aviation Administration, n.d.).

Past the findings of the possible explanations of recorded behaviors, this exercise took into account the level of generality or specificity of the behavior recorded; whether they were specific to the crew, or it was common among the airline staff, or if it can be found in the entire aviation sector. Human factors always indicate a culture present in an organization or field of work (BEA, 2012).

Some sources mention that the pilot had said that he had only had one hour of sleep. If this is true then the Captain was sleep deficient as one needs a minimum of eight hours for peak performance. He had logged 10,988 flying hours and was the most experienced of the crew. 6, 258 of those were as the Captain. The FO2 had logged 6,547 while the FO1 had logged 2,936 flying hours. This is an indication of the pilots' proficiency. They had also undertaken alternate law flying training over the course of their career. The Captain was a veteran in the route and he had a duty of monitoring systems and taking care of communication. It was communicated to the crew that they might fly into a convective zone linked to Inter- Tropical Convective Zone (ITCZ). An ITCZ as a weather pattern is very harsh. Many flights however go through such places without any incidents. Close to the INTOL point they made contact with Atlantico Oceanic Control Center. Their attempt of contacting DAKAR Oceanic didn't go through. On reaching the SALPU waypoint, they experienced a level of turbulence that was considered by the Captain as normal.

On dimming the lights, some bright flashes could be observed from the screen. The Captain remained calm as he was familiar with such a situation but the PF must have been disturbed as he was less experienced. He did not speak out his fears, though. (Dhavala, 2013). The PF might have been anxious. But French culture is high on power-distance and speaking to superiors is frowned upon. So the PF could not voice his concerns even when the pilot took a break and left him to handle the situation. This might have put too much stress on him and might have been a contributor to the accident (Dhavala, 2013).

Part Played by Human Errors

Lee Collins points out that the first thing that a pilot is taught is looking at the airspeed indicator. It is the cockpit's most critical instrument. Airspeed is very important. A lack of airspeed spells doom on the flight.

The AP had disconnected and the pilots had to take manual control of the flight. The FO1 (PF) was in control. The other FO was fine with this as it was expected that the FO1 was capable of managing the situation. The PFD recorded a dramatic drop in air speed and it read 60 knots down from 275 knots. The change was picked by the Integrated Stand-by Instrument System (ISIS) and there was a change to alternate law protections by the flight. When all the basic instruments had failed, shock and confusion surrounded the cockpit. Such a failure is rare (10/flight hour). The pilots could have been shocked by the failures that took place suddenly. No visual guides were available and they were flying at night in a thunderstorm (Dhavala, 2013).

Situational Awareness

The PF situational model based on Endlsley's three stage model give an explanation on the PF situation. He had encountered sudden CAS indicator changes in the middle of a thunderstorm. His comprehension of this was a loss of speed and so the aircraft's altitude and so concluded that worse situations would come if they remained at the same altitude. A decisions was reached to increase altitude and rise above the clouds. The stall warning could have been occasioned by fluctuating speeds and not a real stall (Dhavala, 2013).

Other Failures -- Theoretical Explanations, Decision Making Failure and Cognitive Information Processing

The crew called the Captain having recognized that they needed help. The captain was not able to make it to the cockpit for almost two minutes following the call, maybe because of the aircraft's wild motion. Since the two at the cockpit may not have been aware of this, their stress levels increased. They failed in communication as well as team work and failed to sufficiently discuss what was happening. The prevailing situation was not a loss of speed but a malfunctioning of the indicators. The actions the PF took set the aircraft on a continuous climb. The PNF realizes that the air speeds they have are unreliable as well as alternate law protections and makes a vague suggestion that the thrust should be controlled (BEA, 2012). The directions were not clearly gotten by the PF who at the time was struggling to maintain the stability of the aircraft.

The flight detectors indicated to the pilot to pull up based on erroneous calculations and while the pilots knew of the unreliability of the instruments, they followed the suggestions. Humans are prone to "automation bias" and are likely to trust signals given by instruments even when there is evidence pointing to their possible inaccuracy. The PF, for instance, had learned to pitch up the plane in case they wanted to gain altitude. The pilot paid attention to the instrument suggestion and acted accordingly. Such is a case of rule-based decision making (Dhavala, 2013).

The stall warning had been triggered and it came continuously. An initial warning was considered false. They attributed it possibly to inaccurate air speed. The A330 Instructor Handbook appreciated this and warns of the same and acknowledges that it might be disturbing. An error called 'escalation of commitment' might have contributed to the decision of the pilot to continue climbing. The pilot had reached a decision to climb and even though various signals were being given by the aircraft that would have advised him against that decision, he stuck with his decision to climb because he had high commitment to the decision. All the signals being given by the aircraft at this point were contradictory to the commitment to climb. Further, the PF failed to acknowledge the presence of the aircraft in alternate law. In normal law, a stall is not possible even with full side stick control. Knowing this, they must have depended on this fact and so perceived the stall warnings as not true (Dhavala, 2013).

On entering the cockpit, the Captain was greeted by two distraught pilots and a continuous stall warning. He may have been startled by what was going on as the situation that he had left as relatively stable had badly deteriorated. Further, the two were just telling him that they had lost control and had not the slightest idea of what was happening. The stall warning must have been caused by the speed of the aircraft reducing to 60 knots and the computers made the assumption that they were making a descent (BEA, 2012).