The decision to reject a take-off at high speed will always be tricky. Training helps, up to a point, but it is also vital to get rid of the preconditions of these incidents.
A rejected take-off is your Hamlet moment: to go or not to go, that is the question—just that there is no time for a soliloquy. The odds are that every air transport pilot will face this decision at least once in their career.
Figures compiled by Boeing show a rejected take-off (RTO) occurs once in about every 3000 take-offs. If you add unreported, low-speed rejections, the figure might be as high as one in 2000 take-offs. Boeing calculated a pilot who flew primarily long-haul routes would be likely to face an RTO decision once in 20 years. However, a pilot on short-haul routes making 30 take-offs per month might see an RTO every seven years.
Boeing found about 75 per cent of RTOs were initiated at speeds less than 80 kt and rarely resulted in an accident. About two per cent occurred at speeds above 120 kt, and these were associated with runway overruns and incidents.
- More than half the RTO accidents and incidents reported in the previous 30 years were initiated from a speed in excess of V1. (In simple terms, V1 is the calculated fastest speed from which an aircraft can be stopped on the runway, or the approved stopway area beyond the runway.)
- About one-third were reported as occurring on runways that were wet or contaminated with snow or ice.
- Only slightly more than a quarter of the accidents and incidents actually involved any loss of engine thrust.
- Nearly a quarter of the accidents and incidents were the result of wheel or tyre failures.
Boeing says about 80 per cent of RTO runway overruns were potentially avoidable if appropriate operational practices were followed.
Pilots should have a bias towards continuing, certainly after V1, and even shortly before it, a US operator, quoted by Airbus says:
‘In the high-speed regime, the pilot’s bias should be to continue the take-off, unless there is a compelling reason to reject.’
There is one, and only one, compelling reason to reject a take-off after V1: if the pilot judges the aircraft cannot fly. This is a tough decision, made in moments, and pilots don’t always get right—it would be extraordinary if they did.
A study of 135 rejected take-offs initiated faster than V1 found the pilots made the correct decision about one third of the time. It was the wrong decision, with hindsight, in 44.4 per cent of cases, and unclassifiable as right or wrong about 23.6 per cent of the time.
The study found that after the US Federal Aviation Administration’s take-off safety training aid was released, these figures improved from 31.3 per cent correct, 50.8 per cent wrong and 15.9 per cent unknown.
In addition, the rate of high-speed rejected take-offs fell by 25 per cent, after the training aid was introduced in 1993. The training aid emphasises that it is usually better and safer to continue take-off after V1.
In aggregate, training can improve the likelihood of high speed RTOs. But academic pontificating on what pilots should or should not have done in specific incidents can be an arid intellectual exercise. There is little value in such conclusions on individual cases, reached free from time pressure, and from behind the contemplative comfort of a desk, rather than in the tumult of a flight deck.
Another approach is to do all that can be done to remove the causes of high-speed RTOs—animal strike, engine failure, and, in one case, a control panel ambiguity.
In January 2011, a crew operating a Boeing 777 for an international airline rejected a take-off at an Asian airport, and did about $US600,000 damage to its tyres and brakes in the process. There was no official investigation, but the airline conducted an internal investigation that, frustratingly, it is not making public.
The internal investigation found the RTO on that January night had a long genesis. It began with an auto-thrust problem that meant take-off/go-around mode could not, at first, be engaged at the runway threshold. Auto-thrust had dropped out on a previous flight when the aircraft was in mild turbulence, but this problem could not be found when the module was tested later.
The captain reached up to the mode control panel and pushed a switch in a similar position to one that engaged auto-thrust on the Boeing 767 that he had flown for thousands of hours. On the 777 however, this similarly placed switch turned on the autopilot. Take-off proceeded and at 174 kt the captain called ‘rotate’. The first officer pulled back on the control column and felt resistance—from the autopilot, trying to keep the aircraft straight and level. After continuing to pull against resistance, he said words to the effect of ‘it won’t fly’. The captain immediately took over and moved the levers back to idle reverse while engaging autobrake.
At that moment the first officer was able to break through the autopilot and the nose wheel came off the ground briefly. The aircraft stopped more quickly than its published data said it would, but damaged its brakes and tyres. Nobody was hurt.
A broader investigation found there had been both precedent and warning. In 2009, Boeing had issued a multi-operator message to 777 operators regarding ‘autopilot inadvertent engagement on the ground’. This had been bundled into a technical bulletin and issued to the airline’s flight crews on a CD-ROM, containing the airline’s manuals, technical bulletins and memos. There was nothing to alert crews to new and significant information. Another airline had had a similar RTO just days earlier. Soon after, avionics maker, Rockwell Collins, issued a service letter announcing revised software that would prevent the autopilot from operating on the ground. The airline’s 777 fleet adopted the software less than eight weeks after the incident.
The pilots were, reportedly, subjected to some hostile and ill-informed criticism from certain of their peers. It is easy to imagine that you would have done better, but the record of rejected take-offs reminds us that as human beings we pilots are fallible. It’s part of being human.