01 Mar 2023
Unrealistic Simulator Scenarios & Negative Training
What is Negative Training?
This question has perpetually intrigued flying instructors and ignited countless discussions in the briefing rooms. The need for training is self-evident in aviation, but how it is conducted is the subject of some debate. As is well known, the bulk of airline training is delivered through simulators. The modern flight simulators, known by the umbrella term 'Flight Simulation Training Device (FSTD),' have become so good at replicating airplane characteristics within the programmed feature set that it is almost impossible to identify the areas in which they become unrealistic. For instructors, it is challenging to discern where the machine is trying to mask its 'infidelity' by extrapolating the programmed behavior into regions where either the actual aircraft behavior is unknown, or the training device itself has not been tested at all.
The next question naturally comes to mind; why not train on an actual aircraft? On balance, the capabilities of a simulator in synthesizing scenarios, including weather and runway conditions, at the press of a few buttons far exceeds the training capability offered by an aircraft in the normal and non-normal envelope, even if we ignore the arguments for safety and economics. Once we agree that an actual airliner will not be used, the concept of training fidelity becomes paramount to ensure that the training provided is faithfully transferable back to the actual aircraft.
Coming back to our original question- what is negative training? It may be defined as a state where the crew actions result in a safety event even though they were applied precisely to how the training was delivered. There are two distinct ways this can happen- The first is negative training, and the second is the negative transfer of learning.
If instructors either intentionally or unknowingly operate outside the limitations or the certified capabilities of the training device, it is possible to systematically teach inappropriate skills. The training assimilation will then be based on the simulator's behavior in an area for which the device was never designed in the first place. Such behavior may turn out to be fundamentally different from the way an actual aircraft would behave in flight, but that would never be realized by the instructor and the trainee during training. This is called negative training, and its consequence is imparting skills that can be ineffective or even dangerous when applied in real aircraft.
Another common way in which negative training can occur is the instructors teaching improvised techniques in simulators as well as aircraft. Such techniques may offer a quick-fix solution to some specific situation but may be completely counterproductive when applied in a different context when the scenario shares similar cues.
Negative Transfer of Learning
When previously learned techniques are inappropriate for the current aircraft, environment, or procedures, it is called negative learning transfer. When the crew faces a stressful time-critical situation, heuristics come into play, and it is all too common for them to instinctively fall back into the previously reinforced habits.
Instructors can contribute to this kind of negative transfer by passing on their previous learning from other types to the current aircraft.
An excellent example of negative training can be found in a 2019 article titled 'The Adverse Effects of Unrealistic Simulator Scenarios' published in the Safety First Magazine of Airbus- https://safetyfirst.airbus.com/the-adverse-effects-of-unrealistic-simulator-scenarios/ . The article describes how the « TOTAL PITOT BLOCKED » malfunction that is available in some simulators leads to negative training. This failure simulates a simultaneous obstruction of both the inlet and drain holes of both Pitot probes.
As a consequence, the measured total pressure remains at a constant value corresponding to the total pressure measured at the time of the complete obstruction, and the following sequence of events is triggered:
IAS increases as the aircraft climbs due to the decrease in static pressure. The captain's and first officer's speeds finally reach VMO/MMO with the triggering of the Overspeed warning. Autopilot disconnects and the High-Speed Protection activates ? The aircraft pitches up, and the flight crew tries to counteract with full forward sidestick input without success ? To recover, the flight crew is trained to switch off two ADRs ? Flight Control System reverts to alternate law and Overspeed protection is lost ? crew is then able to manage the aircraft trajectory.
Example of the effect of the “TOTAL PITOT BLOCKAGE” simulated failure on the airspeed computation when in climb
Such simultaneous dual failures with permanent and consistent dual airspeed increase (decrease) when the aircraft climbs (descends) with resulting undue activation of the flight envelope protections have never occurred in service because, in terms of probability, multiple PITOT obstructions can never occur exactly at the same time and can never have the same obstruction characteristics along the timeline. Single or non-identical PITOT obstructions will undoubtedly lead to airspeed discrepancy being detected by the flight control system, which will then reject the erroneous airspeed information and revert to alternate law automatically.
Inserting this malfunction in the simulator results in keeping the normal control law and its associated protections active while the system continues to utilize corrupt data based on incorrect airspeeds. Teaching recovery in this manner confuses the trainees and imperils their understanding of flight control system, control laws, flight envelope protections, and the operational consequences of air data failures. Training consisting of requesting flight crew to switch off two ADRs to revert to alternate law in such a scenario is, therefore, clearly NEGATIVE TRAINING because if this failure were to occur in real life, the aircraft behavior would be completely different from what was practiced in the simulator.
Lessons from the Past British Midland Airways flight 092
Moving our attention to the second aspect, the tragic crash of British Midland Airways flight 092 in 1989 comes to mind as an example of negative learning transfer. On the 8th of January that year, an almost brand-new Boeing 737-400 aircraft was on a flight from London Heathrow to Belfast. During its climb through 28,000 feet, the plane began to shake violently, accompanied by loud bangs and a burning smell in the cabin. Some passengers saw sparks and flames shooting out of the left (No1) engine. Unknown to the crew, inside the No.1 engine, part of a fan blade had detached, creating a mass imbalance in the fan rotor, causing heavy vibrations as it swayed from side to side and rubbed against the fan case. Simultaneously, the disrupted airflow led to an engine surge and compressor stall. These were textbook symptoms of engine severe damage/stall/surge, which is a serious malfunction, but nothing that should cause a plane to crash. In this particular case, most of the engine's internal components were not damaged, and it was still capable of operating at a reduced thrust level while producing sufficient hydraulic and electrical power for the associated systems of the aircraft.
However, as the investigation report would later show, the crew's actions resulted in a chain of events that eventually led to a hull loss, 47 fatalities, and numerous injuries to the occupants. Erroneously believing that the right (No.2) engine had sustained damage, the crew reduced the thrust on it to idle and subsequently shut it down. The wrong diagnosis emanated from the captain relating the source of the smoke and acrid odour to his training on the previous aircraft type flown by him. The captain's previous training and experience were on the Boeing 737-300 variant. After one year of flying the 737-300, he underwent a one-day differences course for the Boeing 737–400, which did not consist of any simulator training. After the accident, the captain stated that he had judged the No.2 engine to be at fault from his knowledge of the air conditioning system. He reasoned that the smoke and fumes were coming forward from the passenger cabin, which is supplied air from the No.2 engine, and hence the trouble lay in that engine. Whilst this reasoning was valid for the previous type he had flown, on the 737-400, some of the cabin air comes from the No.1 engine as well.
Gaps in Training
This minor change had not been covered in the differences course, and with their low time on the new -400, the crew had not had enough time to learn through experience that this assumption was incorrect. In an insidious coincidence, as the No.2 engine was throttled back, the loud bangs and shuddering ceased, reinforcing their belief that they had identified the problem correctly. At the time of the failure, the autothrottle was engaged when it detected a sharp drop in the No.1 engine fan rotation speed. When the autothrottle system detected a mismatch between the commanded thrust and the actual thrust, it was programmed to treat it as an engine surge and reduce the commanded thrust to protect the engine from damage. On this occasion, this automatic reduction in thrust on engine No.1 coincided with the crew action of throttling back on engine No. 2 and the resulting low thrust setting on the engines mitigated the engine surge. As the original symptoms disappeared, combined with the absence of any visual or aural warning on the flight deck, the stage was set for the crew to revert to their previous training and experiences for decision-making. In reality, however, nothing was wrong with the right engine, and as a result of their actions, the aircraft was now being powered only by a damaged left engine, which would later fail when they tried to increase power on the final approach.
The incorrect crew response, in this case, was a textbook case of providing training that was not transferable to the actual aircraft. The Operations Manual at that time contained no guidance on the actions to be taken if vibrations and smoke/fumes occurred together. The pilots were suddenly presented with an unforeseen combination of symptoms that was outside the domain of their training or experience.
In its report, the Air Accident Investigation Branch remarked, "It is possible to identify three aspects of the circumstances of this accident where a different pattern of training could have favorably influenced the outcome." The report also states, "The speed with which the pilots acted was contrary to their training and the instructions in the Operations Manual. Therefore, their incorrect diagnosis of the problem must be attributed to their too rapid reaction." It is well known that both negative training and transfer are most likely to manifest themselves at times of high stress, fear, and surprise. Although it is vital to quickly identify and diagnose certain emergencies, it is also a proven fact that rapid crew reaction increases the likelihood of shutting down the wrong engine. It is worthwhile to note that the pilots had accumulated the greater part of their experience on propeller-driven aircraft like Viscount, Fokker-27, and Shorts 350, where the primary emphasis was placed on the need for rapid feathering of the propeller in the event of engine failure.
In conclusion, it is hard to overstate the importance of recognizing the harmful effects of negative training and negative learning transfers. Regardless of the FSTD capabilities, the instructor remains the critical element in the success of the training. The instructor's awareness of the differences between the simulator and the aircraft is a crucial element in preventing negative training. For any training program, the positive transfer of knowledge and skills must dominate the training delivery, which is only possible when the instructor can meaningfully compensate for the shortcomings of the simulator and exploit the FSTD's strengths as a training tool while minimizing negative training transfer. All the investment in technology, systems, and procedures can come to naught if, to quote First officer David McClelland of BMA Flight 092, "What they're saying is that the people who designed it, manufactured it and carried out the specifications all got it right but the two chaps at the front got it wrong." Later, Captain Kevin Hunt would say in a BBC documentary: "We were the easy option - the cheap option if you wish. We made a mistake – we both made mistakes – but the question we would like answered is why we made those mistakes."
About the Author
As the Accountable Manager and Vice President for Jet Airways (India) Ltd, Capt PP Singh is the Key Management Person in this world-famous airline brand's resolution and revival process. He has been an aviation professional since 1984, possessing extensive training and checking experience accrued with major Indian and foreign international carriers as an instructor and examiner on large commercial jets. Over the past three decades, Capt Singh has held crucial appointments in the senior management teams of Jet Airways and Nepal Airlines in various regulatory and leadership roles. After graduating from the University of Delhi, he started his aviation career with the Indian Air Force and moved to civil aviation in 1994. He is currently responsible for the operations, training, safety, and engineering functions of Jet Airways.