judging takeoff run acceleration

[lynn789]

the near disaster with the emirates airbus at melbourne, australia, with a much too low weight entered into the system, raises the question, how is the acceleration during the take off run measured. Is it judged by the push in the back from the pilots seat only

is there a G meter?

it seems that until the last possble minute, none of the pilots realised the take off acceleration was much too slow

also in something of an apparent panic they pulled back too far on the controls, causing a heavy tailstrike

[barit1]

The answer is - there is none.

Unless - someone takes a stopwatch to time the run from brake release to a target IAS. B-52 crews did this, per my lunch mate. Of course, they had a surplus of crewmembers during TO roll.

It would be pretty easy to implement software to monitor this; the accelerometers in the IRS provide all the raw data you need.

[galaxy flyer]

barit1

Actually, the BUFF had the standard 2 pilots during take-off roll. Also, just about all USAF planes do an acceleration check when doing a standing take-off, even some fighters. The C-5, typically, ran to 110 KIAS in about 38 seconds, plus or minus, on a near gross weight take-off. The engineers ran the data and the pilots monitored the take-off acceleration. On brake release, the co-pilot started the clock, if at the expiration of the calculated time, speed was not less the "line" speed minus 3 knots, nothing was said. "Abort", if you didn't have the speed. In 4,000 hours, only saw one take-off come close and it was obvious that it would be close as we passed 80 knots, I was in the right seat. We made it, but right at 3 knot tolerance.

GF

[catiamonkey]

Emirates wasn't the first to mess up that way. Cases in Canada and the UK come to mind on both Boeing and Airbus. Systems like these have been proposed and made it to the prototype stage. One of the first systems was on the NASA 737 demonstrator, which invented the A320 for Airbus.

Typically called a Takeoff Performance Monitor (TOPM or TPMS). The problem with these systems is that the false warning rate, and subsequent high speed aborts, outweighs the benefits.

The real problem in Emirates and several other cases seems to be miscalculations or simple misentries. You can solve these by cross checking, uplinking performance data, and simple sanity checks on entered speeds, which both Boeing and Airbus haven't implemented. Often, TOPM still relies on aircraft GW and runway length, so you wouldn't be solving the problem.

A lesser factor is whether the engines are producing the commanded thrust, which is solved by the ACUTE system on the A380. There, percent thrust is calculated from a number of sensors and a model, and is used as your primary thrust setting measure. The landing performance monitor is somewhat easier to implement as well, which is showing up in the BTV system on the 380.

[capt. solipsist]

All I do is I make sure I reach 80kts before the last 1,000' marker. If not I reject. :ok:

[galaxy flyer]

Actually, Capt, 100 knots at 3000 feet works rather well for accelerating or slowing. Look at a take-off on a runway with distance remaining markers to check it out.

GF

[barit1]

Typically called a Takeoff Performance Monitor (TOPM or TPMS). The problem with these systems is that the false warning rate, and subsequent high speed aborts, outweighs the benefits.

I fully appreciate the false-reject problem.

However, the objective should be to identify substandard performance EARLY in the roll, to avoid high speed rejects.

[capt. solipsist]

Hey GF,

My bad, I wasn't specific enough. Yup, that's what I mean, the last 1,000' foot marker on the TDZ, which is actually the 3,000' marker. Thanks for the clarification. :}

[catiamonkey]

However, the objective should be to identify substandard performance EARLY in the roll, to avoid high speed rejects.

That's going to create a ton of low speed rejects. Let's say you're going at 40 kts and a 10 kt headwind comes along. Now, according to the model, airspeed is 25% off. The plane shows 55% more drag than I expect. Is it because of a dragging brake? Wing problem? Wind? Not to mention the dynamic change in engine performance. The NASA system commanded a reject at 15% discrepancy, so you'd be stopped by now.

The best method early in the roll is engine parameters, which is what you do today. And then couple the engine parameters to a model in order to detect problems.

[Tmbstory]

As a pilot and as a passenger, I have always counted the seconds to reach VR as a passenger, or as a pilot by a stop watch (if the machine had a capable system).

There are various ways to keep track of your acceleration,

Tmb

[AerocatS2A]

That's going to create a ton of low speed rejects. Let's say you're going at 40 kts and a 10 kt headwind comes along. Now, according to the model, airspeed is 25% off. The plane shows 55% more drag than I expect. Is it because of a dragging brake? Wing problem? Wind? Not to mention the dynamic change in engine performance. The NASA system commanded a reject at 15% discrepancy, so you'd be stopped by now.
A headwind would generate an increase in airspeed, so if the system used airspeed it would show an increase in performance. But these are technical issues that can be solved by proper design, not reasons to not bother designing it at all.

[catiamonkey]

A headwind would generate an increase in airspeed, so if the system used airspeed it would show an increase in performance. But these are technical issues that can be solved by proper design, not reasons to not bother designing it at all. You would be measuring sub-knot airspeed at 40 kts. And you're talking about acceleration. Acceleration decreases as the plane gets faster.

This was an area of active development 20 years ago, and you can search for the literature if you don't believe me. There's even a SAE AS on how well it's supposed to work. Once in a while a safety board complains about it, but you simply can't have a good model until you're close to the high speed regime. It's different measuring time to rotation or time to 110 knots, not 50.

Draw another analogy, take a car and floor it. From the time it takes to get from 0 to 30, estimate its top speed. To 15%. With a fat guy in the back, speed bumps, rain and wind.

To mathematically prove it's not going to work, take a state-space model of an accelerating airplane and calculate the observability grammian at low speeds. It's basically the calculation I showed you. It's not that nobody's tried, it's that people have tried and realized it doesn't work until high speeds, and as everybody agrees then one false alarm is going to cause a big problem.

[galaxy flyer]

The C-5 Performance Manual states that the check speed, which was variable between 80KIAS and 120 KIAS, would be as as close as possible to V1-10 knots. I think worded that right. Made for accurate checks, but was a set up for a high-speed abort for sub-standard acceleration. Other planes just used 100 knots, which seems reasonable, esp. Airbus.

GF

[Wizofoz]

On EFIS aircraft, you have the trend vector on the speed tape.

I believe Boeing test pilots need a 30kt trend at 80kts or they RTO, and I've been using that as a personal check since the MEL incident.

[captainmorgan888]

on the Falcon 900 we have longitudinal g´s on the PFD for T/O and Landing to check the accelaration/decelaration....

[Sciolistes]

737 80kts by 1000 feet seems to work as a ballpark.

[Microburst2002]

This raises a question:

speed trend vectors are purely inertial ("GS knots to be gained in 10 sec") or it is converted to "IAS knots to be gained in 10 sec"?

I think incidents like those can be detected before lining up, if the computer rejects Vr or V2 speeds non consistent with GW and thrust.

Any reference for 320s used by Airbus test pilots like that of the 737s?

[Nearly There]

I wonder what method these guys used? :E

YouTube - Close call! (http://www.youtube.com/watch?v=aWtdtuspnoM)

[Tarq57]

They used PAFL (http://www.pprune.org/questions/204288-just-after-takeoff.html#post2295083).
In this case, mostly generated by the crew, I'd imagine.

[GarageYears]

In this case, mostly generated by the crew, I'd imagine.

Well, it'd have to be, given that is a seriously large CARGO aircraft - not nearly enough PAFL to go round. Explains quite a lot, now I think about it.

- GY :}

[PantLoad]

It's always worked out.....using the distance markers painted on the runway....by the time you get to the three-thousand foot marks, you should be doing an absolute minimum 100 knots. Typical speed at this point is maybe 120 knots.

This is not scientific...probably nonsense (feel free to substitute another eight-letter word)....

As far as I know, only the military uses accel checks. Some biz jets have
accel readouts (in "g"s)....

Used this 'method' when I was sitting sideways on the 727. If we weren't making at least 100 to 105 kts at these markers, I'd respectfully mention it to the captain...respectfully suggest we 'push em up' a bit....never
had a commander object....


Fly safe,

PantLoad

[Wizofoz]

Microburst,

Excellent question.

My understanding is that it is an indication of IAS trend, but stand to be corrected!

[Walter E Kurtz]

For the A320, give or take a bit for GW and wind, 20 seconds from brakes release to 100 knots is the gate to watch.

[abc1]

Is thrust set? Over time one should develop rough thrust figures vs weight.
ASI Trend vector helps.