All aircraft are rigorously tested in many areas before launch. This includes safety and emergency procedures and the ability to evacuate the aircraft quickly. While this test seems straightforward in principle, there are various concerns with the tests that have changed little in over 50 years.
Testing requirements – the 90-second test
The main requirement for testing an aircraft for evacuation compliance has remained the same since the FAA introduced it in 1967. The FAA regulates tests in the US, and the rules in Europe are very similar. This test requires a full aircraft (passengers and crew) to be completely evacuated in 90 seconds or less. The 90-second limit comes from analysis of the time it would take for a fire to engulf the cabin.
Stay informed: Sign up for our daily aviation news digest.
There are rules on passenger makeup to reflect a real cabin. There should be certain proportions of male and female participants, as well as a proportion of over 50 years old. Some have to carry dolls to represent infants.
Tests take place in an aircraft hangar, in the dark using evacuation cabin lights, but not a smoke-filled cabin. Half of the exits are blocked as well, to simulate real-life problems. Passengers should exit using the aircraft slides.
Modifying the rules over time
There have been some modifications to the rules since 1967, but nothing major. These have been mostly to limit injuries during testing. There have been cases of volunteers (or company employees) being injured during frantic evacuations. Changes have included some cases where slides do not have to be used (passengers can evacuate to a platform) and some where partial evacuations can be used instead of full ones.
An amendment has also been made to allow some tests to be replaced by analysis or modeling. Generally, this is permitted when only minor changes have been made to the aircraft, but new models will still need full evacuation tests.
Is this enough testing?
Some have argued for a long time that the 90-second tests do not go far enough. There have been several problems highlighted.
Firstly, the tests do not do enough to ‘stress’ the evacuation. Passengers are not subject to any of the effects of a real evacuation, such as injury, debris, or even smoke in the cabin. Nor do tests have to include passengers with mobility problems. And passengers are all independent. In real life, many would stop, or divert, to assist their travelling companions.
And of course, all passengers know what is about to happen! A real evacuation would not usually have this chance to be prepared, a significant factor.
Secondly, flying has changed a lot since the 1970s. Seats are smaller; passengers often bring more baggage now that fees have increased, and there are a lot more electrical items to get in the way of exiting (both by blocking routes and distracting passengers who decide to use them).
And finally, there are many different types of evacuation scenarios. You can’t predict or test all of them. But is a controlled evacuation in an empty hangar enough? You just have to look at a couple of full evacuation scenarios to appreciate this. Think of US Airways Flight 1549 evacuation on the Hudson, or British Airways Flight 38 suffering a last-minute problem and crashing on landing at Heathrow in 2008.
Combining evacuation with computer simulation
Overcoming these concerns about tests is not easy. Tests have been altered to reduce the chance of injury, and regulators or airlines are not likely to want to add elements that will increase this.
One proposed solution is to make more use of simulated tests. Although these are now in regular use, the scenarios being tested have not changed.
FlightGlobal looked at the possibilities of this. It noted several advantages, including the ability to test more than one scenario. More thorough testing could also feedback to aircraft design, allowing optimal cabin layout for safe evacuations.
FlightGlobal quotes Ed Galea, UK Civil Aviation Authority professor of mathematical modeling explaining how multiple tests could be of benefit:
“It is impossible to know whether or not the outcome is a fair reflection of the aircraft’s evacuation capability. In contrast, the multiple tests enabled by computer simulation generate a distribution of times, reflecting what would happen if the full-scale evacuation scenario could be repeated. This provides a better indication of the performance capability of the aircraft.”
Moving to a full modeling scenario would undoubtedly have challenges. How to ensure it is accurate for each aircraft type? How would you model for unexpected human behavior? But real-life tests have their problems too. Plus, the cost. Each test costs around $2 million. And it comes with the risk of injury, surely both things that airlines, and regulators, would like to avoid.
What do you think of the current evacuation test requirements? Could, and should they be improved?