So the computer in the Germanwings disaster was listening to a human and if left alone would not hav e crashed the plane. Perhaps we need MORE technology and less reliance on potential human error.
On Sun, Mar 29, 2015 at 10:42 PM, G Mann <g2ma...@gmail.com> wrote: > Rich, > > Nice bit of aviation history. While I salute your well earned hard work in > all the FBW software, as a pilot I still hesitate to give command of the > airplane over to a machine that runs on tiny bits of electrical signal.. > There is simply to much that can fail or render a false signal which > affects control of the airplane. There have been many instances where > pilots then failed to recognize what was wrong and took incorrect action to > correct the sensor/computer problem, and people died. > > While I'm sure it's a lovely system, we now have a syndrome among pilots > where they are so busy flying the computer they forget to fly the airplane, > their ultimate responsibility. > > The term "Pilot in Command" means exactly that, in the most draconian > sense. I'm old school and I'm draconian. No excuses. > > In the recent case of GermanAir crash. The PIC left the cockpit because > Pilot Relief Tubes have been removed, and reliance on a computers ability > to fly the airplane regardless of having a low time, less skilled Second > Officer in control, which the PIC apparently did not trust to make the > landing, based on CVR info. > > The old aviation wisdom that says "all airplane crashes start on the > ground", in my opinion, apply in this recent case. Far in advance of a > deeply off balance Co-Pilot being given control of the aircraft, decisions > were made about how the airplane could and would be controlled other than > by the crew. The PIC released the aircraft because he was, we now know, > "dying to take a pee", and trusted the computer to fly the airplane, > apparently, while not completely trusting the man left in charge.. > > If you work backwards to discover the logic which drove such a decision > making process, I believe I see a logic breakdown, influenced by over > dependence on computer control. > > Of course, I come from a group of old school pilots who would pee their > pants rather than release control of the aircraft, and some did. > > On Sun, Mar 29, 2015 at 3:15 PM, Rich Thomas < > richthomas79td...@constructivity.net> wrote: > > > So back when I was in kawledge I had summer internships at NASA Langley. > > One summer I was in the Flight Research Division, having worked the > > previous summer in the Flight Simulation Division, wherein the aircraft > > being simulated I was then working on in the real world. The primary > > aircraft was a 737 called the Terminal Configured Vehicle, which had a > > replica flight deck in the back cabin, from which the airplane could be > > flown "by wire" with the safety pilots up front being able to manually > > override the pilots in the back. We did all kinds of tweaking of the > > software to allow the plane to take off, fly, and land, all either > > automatic or by the "wire" inputs to the flight computer. I "flew" the > > simulator quite a bit while we were tweaking things, we would do a > > simulation of something then my colleagues would twiddle with the > software > > (running on CDC Cyber computers, which were about the size of my 30 cuft > > refrigerator) and then we would see what happened. If stuff worked OK > then > > it would be uploaded to the airplane, probably on tapes or something, I > > never saw that process. > > > > As far as I know this airplane and the software was the precursor to > > pretty much all the FBW stuff in all the planes today, so it is kinda fun > > to think back on that and whatever small role I might have had in that > > "progress." > > > > One day while I was in the Flight Research group someone came to me and > > said there was some problem in the autoland algorithms, pull the flight > > data and go over to my previous group and see what we could figure out. > > The problem was that when the plane was landing on one particular runway, > > it would pound in hard. The safety pilots would think it was descending > > too rapidly and then try to correct and the result was a hard landing. > Of > > course the pilots, whose man-equipage needed its own seat, were never to > > blame for this. > > > > So I printed off all the data, made plots, looked at all the control > > loops, we put it all in the simulator and "flew" the same landings with > no > > problems. I spent a lot of time trying to figure out the problem. One > day > > I was driving to lunch on the other side of the airfield and noticed > there > > was this big drainage ditch off the end of that runway, and the runway > was > > 10 or 12 ft above the ditch. HMMMM I am thinking, remembering some radar > > altimeter data. (Langley sits on the marsh on a backwater of the > > Chesapeake, and is about 2ft above sea level, pretty much like my place > is > > now) > > > > So I get back and look at all that, and sure enough you could see a big > > "altitude" jump right before landing, where the altimeter was sensing > that > > ditch and commanding the aircraft to go down a bit to compensate. But in > > the simulator, once it crossed the threshold and got over the runway, it > > would pitch up a bit to maintain proper descent to landing. At that > point > > it was maybe 30ft above the runway, the ditch made it think it was like > 40+ > > ft. So I look at the simulator, and quite clearly the flight computer > > caught this, and the airplane was responsive enough to the computer, that > > it would land properly and smoothly, no drama. But when the pilots got > in > > the loop, they were of course much slower to react and command the > > aircraft, so it would not pitch up to compensate and would pound in, > sooner > > than the touchdown point the computer would land it. The pilots > basically > > were the problem. > > > > So a few days later we had a meeting to go over what was going on, and my > > older colleagues say, "Rich how about you go over what you found." So I > do > > that, we took the data, ran the simulations, no problems, then hey look > at > > when the pilots took over because they thought it was going to land short > > or hard, then BANG it lands short and hard." If you had let the system > do > > its job, no problems. > > > > So, the pilots then had a shitfit and start in on me, > > whothehellisthiskidwhatdoesheknowblahblah, and I see the guys smirking at > > me, they had set me up to deliver the news and catch the flak from the > guys > > with the big.... egos. > > > > So then a coupla days later one of the pilots comes charging in our space > > and walks over to me and says, "THOMAS YOU'RE COMING WITH US!" I of > course > > about evacuated, but followed him and he goes to the airplane, tells me I > > am going, and they are going to fly the profile a few times and see what > > happens. So I take a seat back behind the aft flight deck where I can > > watch what's happening, and after 3 or 4 touch and gos we come back, > > nothing is said. So we get back to the office and a couple of the other > > guys who were on the plane are all laughing at me and saying "those guys > > were pretty damn quiet, huh, looks like you were right!" So, shonuff, > the > > computer landed the plane fine, the pilots tried it and pounded it in. > > Vindication! Of course the pilots never admitted it but they treated me > a > > bit nicer for the rest of the summer. > > > > Here's a blurb about the research fleet, the bottom pictures of NASA 515 > > show the airplane. My office was in that hangar to the left in that > > picture, it was full of mice and cockroaches we would launch paper clips > at > > with rubber band slingshots. > > https://books.google.com/books?id=wyJCIhtknPcC&pg=PA91&; > > lpg=PA91&dq=NASA+TCV&source=bl&ots=fDj3VbqLZU&sig=VogQvNb- > > twMlFeWtAOAktmy4iKQ&hl=en&sa=X&ei=jVAYVfLFEoGDgwT02YGoBw& > > ved=0CEYQ6AEwBg#v=onepage&q=NASA%20TCV&f=false > > > > > > Here's a paper by one of the crazy guys I worked with > > http://www.researchgate.net/publication/23819008_ > > Verification_and_validation_of_the_NASA_Terminal_ > > Configured_Vehicle%27s_TCV_Wind_Analysis_program_using_ > > real-time_digital_simulation > > > > > > > > --R > > > > > > > > > > > > On 3/29/15 1:53 PM, G Mann wrote: > > > >> Gentle crash landing... for an Airbus. > >> > >> Since Airbus uses computer controls for landing approach and touchdown, > I > >> would want to replay the last 6 minutes prior to contact with the > runway, > >> and the full replay of all control imputs made after contact. > >> > >> It is pretty apparent from the chosen parking spot that the landing > >> envelope was exceeded. > >> > >> On Sun, Mar 29, 2015 at 8:31 AM, Andrew Strasfogel < > astrasfo...@gmail.com > >> > > >> wrote: > >> > >> It'll buff right out. > >>> > >>> On Sun, Mar 29, 2015 at 10:55 AM, Curly McLain <126die...@gmail.com> > >>> wrote: > >>> > >>> Air Canada plane 'exits' Halifax runway while landing > >>>> > >>>>> 2 passengers seriously injured. Plane severely damaged. > >>>>> > >>>>> http://www.bbc.com/news/world-us-canada-32103971 > >>>>> > >>>>> ________ > >>>>> > >>>>> Read again GMann's position on flying in airbus... > >>>> > >>>> > >>>> _______________________________________ > >>>> http://www.okiebenz.com > >>>> > >>>> To search list archives http://www.okiebenz.com/archive/ > >>>> > >>>> To Unsubscribe or change delivery options go to: > >>>> http://mail.okiebenz.com/mailman/listinfo/mercedes_okiebenz.com > >>>> > >>>> > >>>> _______________________________________ > >>> http://www.okiebenz.com > >>> > >>> To search list archives http://www.okiebenz.com/archive/ > >>> > >>> To Unsubscribe or change delivery options go to: > >>> http://mail.okiebenz.com/mailman/listinfo/mercedes_okiebenz.com > >>> > >>> > >>> _______________________________________ > >> http://www.okiebenz.com > >> > >> To search list archives http://www.okiebenz.com/archive/ > >> > >> To Unsubscribe or change delivery options go to: > >> http://mail.okiebenz.com/mailman/listinfo/mercedes_okiebenz.com > >> > >> > >> > > > > _______________________________________ > > http://www.okiebenz.com > > > > To search list archives http://www.okiebenz.com/archive/ > > > > To Unsubscribe or change delivery options go to: > > http://mail.okiebenz.com/mailman/listinfo/mercedes_okiebenz.com > > > > > _______________________________________ > http://www.okiebenz.com > > To search list archives http://www.okiebenz.com/archive/ > > To Unsubscribe or change delivery options go to: > http://mail.okiebenz.com/mailman/listinfo/mercedes_okiebenz.com > > _______________________________________ http://www.okiebenz.com To search list archives http://www.okiebenz.com/archive/ To Unsubscribe or change delivery options go to: http://mail.okiebenz.com/mailman/listinfo/mercedes_okiebenz.com
