I think this is a really important topic that is drifting a bit…
Let’s consider the following points
(1) Humans by their very nature are unreliable creatures, they can not be relied on to repetitively complete a task without occasional errors.
(2) Systems are inherently reliable by their nature, they can continuously repeat a task with absolute reliability.
The factory can inject energy into both human training or improved systems. I am stating that we might now be at the point where the marginal return from training is at it’s limit – we’ve been improving pilot training for 100 years and we still crash. However, system implementations like the ones noted above have a material impact on statistics that training is unable to achieve. Therefore, this thread asks what further system improvements could reduce the accident rate. I proposed my idea as dampner to make the landing record like that of the DA42, which Noe correctly points out is very difficult to bounce. No bounce = no accident.
The great challenge is that in order to innovate, society needs to understand the limits of human ability. The average GA pilot in Europe flies <20 years per year and is much more proficient typing on a keyboard than flying. Furthermore, I am continuously surprised by pilots who feel they don’t need training and train for the minimum 1 hour every 2 years… far easier to fix the system than ground all these untrained pilots.
I don’t really see why this discussion focusses on Cirrus only. The problems are the same for all GA planes.
Both accidents you mention do not appear to have been Cirrus specific problems, the one which landed on the shute had an engine failure and the other one is anyone’s guess so far but it does not appear Cirrus specific.
pistonfever wrote:
We need to take responsibility away from the pilot somewhat and make the system safer. Without doing this, we will be the last generation to accept the inherent risks associated with being PIC.
That is the way airliners do things since the introduction of the A320 or at least that is when it started to become prominent. The trend towards automatisation and “taking away responsibility” has however imho proven that it opens a new can of worms, the most prominent of which now has been the 737Max problem, where the airplane took on the full responsibility for 2 fatal crashes (and possibly some more even before the Max) and left the crew powerless to counteract the responses of the airplane.
The view that people should be taken off the controls of just about anything is not restricted to airplanes but goes everywhere today, they are talking of pilotless airliners, self driving cars, not mere “assistant” programs but without any controls, they are automatizing workplaces in the interest of safety (but of course the real reason is that automats don’t want pensions and sick leave and don’t usually go on strike) but the problems don’t get less, they increase.
To me, the mere suggestion to take the responsibility away from a PIC who then actually is no longer PIC but merely a passenger of automatisation has been exhaustingly proven wrong by the technology of the last 3-4 decades. Even Airbus is moving away from this following the amock of AF447, but they should have known much earlier, after loosing 2 A320’s to the control particularities and some very close shaves on the 330’ties which, had they happened near the ground, would have ended up exactly like the Max accidents.
There is nothing wrong with assistand programs and certain protections as long as they can be overruled and managed. This will have a positive impact on safety, as we can see already with current terrain awareness, collision avoidance and other such systems which are available. There is room for improvement in automatic flight as well if autopilots finally get IAS based pitch modes and basic protections for their own problems.
IMHO, the main problem with GA safety is a very different one: Pilot proficiency.
GA Pilots fly between 12 to 50 hours a year, MOST are flying less, even down to the bare minimum of 12 hours in 2 years to keep their license valid. That is NO proficiency at all. Most do not fly in the winter months, so GA activity is restricted to maybe 2-3 months per year with them sitting idle the rest of the time. Consequently, every year in spring the amount of accidents start to rise, having a peak in summer when people have vaccations and try to do their flying in the very limited time they have. Add to that time pressure and resulting “get there itis” enhanced by restrictive and financially punitive systems of clubs and workplaces and you end up with a scenario where the whole system becomes inherently unsafe.
The most urgent problem is to get people to either fly more or to stop flying altogether. 12 hours in 2 years is nowhere close enough. If we want to keep some sort of proficiency, you have to fly at least every month, preferrably once a week or more. I’ve seen studies which say if the requirement would go up to 5 hours per quarter year and a number of landings in the same period, with the obligation to take a requalification if you don’t do this. Other numbers I’ve seen are 100 hours per year with at least 5 approaches per quarter for IFR ops.
Clearly, this would have a massive impact and my estimate would be that such requirements would be the end for most pilots, so such demands should be done with great reflection. The general question would have to be if GA piloting is something you can really do as a past time or if certain pre-requisits are needed to do it. Personally I have to admit that I see it less and less feasible to fly responsibly along a family and work due to lack of time. And I know a lot of other people who tell me the same thing when working and family eat up to 100 hours of time per week.
On the airplane side, I think Cirrus already have taken the logical step with the CAPS and I would estimate that within a few years this will become a requirement for new airplanes. There is nothing comparable to the increase of safety a full size rescue system provides. Unfortunately this was brutally shown not too long ago yet again when a brand new Acclaim crashed following engine problems with fire after landing which has massively injured it’s owner pilot. A BRS system would have saved the day there.
The comparison with motorcycle driving is not surprising for several reasons, the most of it being that in large part of Europe, motorcycles are only used during summer. Consequently accidents mount in spring and reach a peak in summer. Also there, proficiency is a huge issue after 4-5 months not driving and the inherent danger of motorcycles with it’s occupants being largely unprotected in case of collision or falling over adds to the often deadly results. The prohibition of driving cycles and even bicycles these days without a helmet has also helped reduce fatalities, I guess we shall see more of the same.
IMHO, if we want to make flying safer we must get people to fly more to stay in training and to keep up the regime of bi-annual check flights and revalidations for IR as they are done now. To make people fly more, we must reduce cost and senseless restrictions at airports, so flying is easier accomplished after work or around work. Punitive systems by clubs and rental agencies should be prohibited, which however will impact them greatly and most probably will kill rental in favour of owning, which however would not be a bad thing.
And finally, GA airplanes can be made safer with a BRS, so I’d vote to make this compulsory for new planes, even if this might kill my favorite make forever. But we don’t drive cars without seatbelts around either nowadays…
Mooney_Driver wrote:
The most urgent problem is to get people to either fly more or to stop flying altogether. 12 hours in 2 years is nowhere close enough. If we want to keep some sort of proficiency, you have to fly at least every month, preferrably once a week or more.
Mooney_Driver – you make some really excellent points. However, we are not going to be able to change the human part of the above. People are limited by time, money and weather to flying more. We must accept the fact folk will just fly for 12 hours in 2 years – that’s simply always going to happen. Therefore, we need to make the environment, the human-machine interface, the skill requirements all much easier so human error is forgiven and not causing a fatality.
I refer to Cirrus as they are innovating currently, other manufactures aren’t interested. Diamond don’t crash somehow. But why are people seemingly adverse to making the plane easier to land? It sounds as if pilots WANT to need more skills to land it and WANT the repercussions of fatal accidents when humans error.
Diamond don’t crash because most are school ones, so duos most of time with a FI at least, and most of them are really expensive flyier, although they are not so luxurious, so not really in the fashion. So people that fly them are usually more serious, not flying them like a saturday expensive hobby playing with electronic gadgets in the cockpit. But actually diamond crash sometimes, and DA40 are not the best to land, very wobbly in the ground effect.
I think if GA was cheaper, people would fly more and crash less. Cirrus, or diamond, or TB… We can tell about airbus because they don’t pay taxes… They already flying a lot.
pistonfever wrote:
Therefore, we need to make the environment, the human-machine interface, the skill requirements all much easier so human error is forgiven and not causing a fatality.
That sounds very different however than taking away the responsibility and therefore giving the command to the airplane rather than the pilot.
And many attempts to make things “easier” in recent times have resulted in the opposite, at least when the protections fail or misbehave due to sensor failures or similar. Then you get a lot of malfunctions which are much more difficult to deal with than a simple sensor failure or instrument bug.In those failure modes, automatisation often does the opposite of the intentions, it makes managing the situation much more difficult up to the point where it is unmanagable.
pistonfever wrote:
I refer to Cirrus as they are innovating currently, other manufactures aren’t interested.
Cirrus were the first to put a parashute into their certified airplane. That is their big innovation and that is what de facto put everyone else out of business.
Obviously some of the new AP’s do protections as well, but whether that is a Cirrus innovation I don’t know. However, they also transfer to other makes. DFC90 with it’s protections and similar AP’s are not limited to Cirrus. Also the new S-Tec offering will have protections. Most current manufacturers have similar or equal cockpits and AP’s, the Ovations and Acclaims also use the latest Garmin cockpit with it’s AP. All that sets them apart in this is the shute and it is the most important.
Fine tuning safety systems does not improve the overall safety unless pilot exposure and skills remain the same (moral hazard that increase risk profile for rusty pilots) or the underlying physics are sound (safety system that mask profound design issues)
Things go wild when you put a complex safety nets when the underlying physical margins and pilot skills are eroded…
greg_mp wrote:
I think if GA was cheaper, people would fly more and crash less. Cirrus, or diamond, or TB…
Exactly.
Imagine what would happen on our roads if car drives only drive a couple of times a year. Those who do usually end up in accidents more often than experienced drivers.
CAT is a different story altogether. Those pilots fly 100 hours a month.
Yes, AP envelope protection across the manufactures was a wonderful innovation. I suspect after the 737 MAX accident, we may see dual sensors in the next generation. It is always getting better. No more stalling in VS mode or at the bottom of a descent.
Mooney_Driver wrote:
Those pilots fly 100 hours a month.
In which 5 hours handflying. That is not so different than us…
pistonfever wrote:
Yes, AP envelope protection across the manufactures was a wonderful innovation. I suspect after the 737 MAX accident, we may see dual sensors in the next generation. It is always getting better. No more stalling in VS mode or at the bottom of a descent.
It was and is but you appear to forget the price we paid for it. Two A320’ties crashed due to their protections doing the right thing at the wrong time, the A330 had two accidents (with people hurt) enroute in a similar scenario as the max accidents but with different logic behind it, but still they were massive. AF447 lost control because the crew were no longer capable of basic flying skills, dumbed down by the system. These accidents were huge wake up calls even then, the Max saga of course brought this a tad further.
For our GA planes I am opposed to this kind of automatisation which overrules the pilot. Assistance yes, overruling no. Not even on airliners or rather particularly not there.
