Menu

Human factors assumptions, part 1

As I write this, the Lion Air and Ethiopian Air Boeing 737 MAX accidents are still being investigated. While we know that the MCAS system is going to get the major share of the blame, there is also a push to change the way pilots are trained. One of the topics that has come up is one that was addressed in the movie Sully, the story of the USAirways dead-stick landing in the Hudson River, and that is some of the assumptions that go into aircraft and systems design.

Engineers—both hardware and software—creating a new design need to make some basic assumptions about pilot reaction time, knowledge, and experience. Reaction time delay is one of the most difficult things to predict. Modern aircraft are so dependable and so reliable that it’s easy to take them for granted. And that’s the problem: When something does go wrong, it’s critical that the time lag of a response be given adequate consideration. As Sully showed, when the crews in the sim knew exactly what was going to happen and when, and were allowed to respond immediately, they had no trouble getting the crippled A320 back to departure airport, especially when allowed to practice several times.

In reality, though, such events almost never go so smoothly—after all, who ever anticipates losing both engines to a flock of geese? Imagine dealing with the shock of some kind of a collision, followed by a marked change in the normal noise pattern of flight, and then the audible chimes and lights and other indications of an anomaly. Then, once that has begun to set in, the brain has to convince itself that what it is seeing or hearing is real.

Media reports indicate  this happened with the Lion Air and Ethiopian Air crews. Additionally, it involved a system that the crew of the Lion Air flight was totally unaware of, and the crew of the Ethiopian Air flight was only marginally aware of. The noise of the stick shaker—which is extremely loud and distracting in the 737 by design—combined with the realization that the airplane was descending and trimming itself nose down must have been overwhelming. In both incidents, there was surely a realization at some point that the crew was unable to overcome the airloads in order to reverse the trim.

It’s one thing for designers to try to anticipate crew responses during the early phases of flight. But they also need to look at human factors from several angles, including crews that might be in the middle of a longer flight on the back side of the clock, such as a red-eye or a transcon. The effects of fatigue on sensory response need to be accounted for, which is another reason that some warnings are designed to be loud and attention-getting.

The type of fatigue matters too. Is the crew tired because it’s the last leg of a six-leg day, or is it because they are flying in the middle of the night? Crew experience also needs to taken into account. An experienced, well-trained crew is going to have a better response under virtually any circumstance, and there is reason to believe that at least one of the pilots involved in the Ethiopian Air crash may have been extremely low on the experience meter. Throw in a similar situation with fatigue or personal stress, and such an individual could easily be overwhelmed. It might impossible to account for every possibility, but realistic common denominators need to be established.

Manufacturers do what they can to test their theories and assumptions in the simulators, but there are limits to the effectiveness. Every pilot knows that during a sim flight, something will go wrong. They may not know what, or when, or where or how, but they are primed for a surprise, so even the surprise isn’t a total surprise. Further, when you know that you’re in a box, you know that you’re eventually walking away. That means that the effect of full-blown fear and panic is almost impossible to test for or measure.

There has already been much discussion about human factors assumptions moving forward as result of these accidents, and it’s a discussion that will go on for some time. Checklists and procedures are already being retested, rewritten, and studied. Pilots have complained for years that inexperienced cockpit inhabitants—usually first officers—are unable to cope with a sensory onslaught of often conflicting information. These accidents seem to bring some evidentiary data to that argument, though we must wait for the final reports to be written.

What we do know is that 346 people were killed in very preventable accidents, and the laws are written in blood. Changes will be coming.—Chip Wright

2 Comments

  1. Great article, Chip. As an Accident Reconstructionist, I am frequently asked to explain Perception/Reaction Time. In motor vehicle accidents, it can vary based on the situation a driver is faced with. While waiting for a light to turn green, it can be as quick as 1/2 second. When faced with a situation totally unexpected at night, 2.5 seconds is more acceptable for the average driver. If age or physical wellness is a factor, it can be much longer. These times were determined from testing. You correctly pointed out that the reaction times in testing must be as real-world as possible. Obviously, this is easier in vehicle testing where we can simply notify the subject drivers that we just want them to test drive a car. We can then throw a situation at them and time their reaction. Even if the driver knows in advance that something will happen, his awareness will be heightened and reaction time shortened. Perhaps in aviation, a larger factor of safety should be considered because real-world testing is not feasible. Until we can eliminate the human factor entirely, Perception/Reaction times must be considered.

  2. Chip. As a retired Flight Engineer on the P3 Orian there is a lot to be said for the 3 person Flt Station. (being politically correct) there is many a time the expertise of the Flt. Eng. Saved the day. The fact that both the navy and the the airlines have reduced Flt Station personnel Has not made flying safer. Here we have cases that automation is not always the answer.

Leave a Reply

Your email address will not be published.

*

*