Jesse: IT question

[Greg Bockelman]

How hard is it to develop a performance calculator type application that will spit out takeoff and landing data as well as climb performance based on inputted data like temperature, pressure, etcetera?

Just curious. W&B is a simple spread sheet type application but I am not sure how complicated other performance calculations are.

 

[SkyHog]

If you know the formulas, it is very easy. The inputs get stored as variables, and the output pulls it all together using the formula.

Oversimplified, but it isn't hard.

 

[jesse]

If you know the formulas, it is very easy. The inputs get stored as variables, and the output pulls it all together using the formula.

Oversimplified, but it isn't hard.

What Nick says -- just depends how complicated the formulas are and if you know them to begin with.

From there, the complexity depends on where you want it to run, how nice you want it to look, all of the features you want, and whether or not there's a bunch of red-tape (certification) crap you need to get done for it to be used in whatever capacity is required.

 

[SkyHog]

What Nick says -- just depends how complicated the formulas are and if you know them to begin with.

From there, the complexity depends on where you want it to run, how nice you want it to look, all of the features you want, and whether or not there's a bunch of red-tape (certification) crap you need to get done for it to be used in whatever capacity is required.

Somehow I missed the "Jesse" part of the subject. My apologies for stealing your thunder.

 

[Sac Arrow]

"The formulas" would be the hard part. The various performance tables would have to be turned in to graphs, which in turn would be made in to best fit polynomial equations.

 

[Everskyward]

"The formulas" would be the hard part. The various performance tables would have to be turned in to graphs, which in turn would be made in to best fit polynomial equations.

That's what I was wondering. I'm not sure if manufacturers come up with mathematical formulas or if what shows up on performance graphs are smoothed test flight results.

 

[SkyHog]

That's what I was wondering. I'm not sure if manufacturers come up with mathematical formulas or if what shows up on performance graphs are smoothed test flight results.

Well, technically, you can derive the formula from the math needed to plot the data on the graph. One hard way to come up with the performance graphs is to plot the best and worst case scenarios to find the envelope.

Then, when plotting other data, it either falls within the envelope or outside of it. Finding best and worst case is part of the Part 23 certification process, I believe.

 

[murphey]

I had a PalmPilot app for this 10 years ago. Unfortunately, the company never migrated it to iOS. But there are apps out there that do exactly this, plus you get to input weather (wind direction & speed) as well as runway condition, e.g. hard surface, grass, ice, water, etc. I'll go see if I can find the details.

For more discussion and various calculators, check out

http://www.luizmonteiro.com/Index.aspx

 

[jsstevens]

Depending on how big the data tables are, they could be stored and use simple interpolation for calculations. The real formulas would be better (at least more elegant) but the interpolation should work just fine.

John

 

[N801BH]

Well, technically, you can derive the formula from the math needed to plot the data on the graph. One hard way to come up with the performance graphs is to plot the best and worst case scenarios to find the envelope.

Then, when plotting other data, it either falls within the envelope or outside of it. Finding best and worst case is part of the Part 23 certification process, I believe.

I "think" Gulfstream tried that in New Mexico a couple of years ago while gathering flight data for the new G-650.....

Didn't work out so well..

 

[nauga]

I'm not sure if manufacturers come up with mathematical formulas or if what shows up on performance graphs are smoothed test flight results.

It's a combination of both. Most any decent airplane design or performance textbook will have the basic equations for any type of airplane performance most of us would be interested in. The tricky part, as some have mentioned, is filling in the blanks for all of the aerodynamic, powerplant, mass properties, etc... information that's required for an accurate result. Most manufacturers of any size will develop a 'mathematical model' of an airplane and use that along with the everyday atmospheric properties and stuff in the performance equations to estimate performance, then validate or 'anchor' the model with flight test data. Once the model has been validated, or proven accurate, performance at most any condition can be calculated without having to explicitly test it. It's typically the validated model that's used for POH/OpMan graphs and tables, not strictly flight test data.

In the case of the Gulfstream accident N801BH refers to, the test point that lead to the crash was to collect data to establish an 'endpoint', or an extreme of the flight envelope, specifically singe-engine continued takeoff performance and limits. Gulfstream's submittal to the investigation points to a model that was not sufficiently validated, exercised, or analyzed (or some combination of all 3) prior to this test point. They state as much but not explicitly using these words, and the company clearly and candidly accepted responsibility for the accident. They were not the first and they unfortunately will not be the last. There but for the grace of God and a strong sense of self-preservation go I.

Nauga,
who lost a friend that day

 

[Ted]

There's a reason why the job description is "test pilot." We desperately want these people to do these tests so we don't do them ourselves.

 

[N801BH]

There's a reason why the job description is "test pilot." We desperately want these people to do these tests so we don't do them ourselves.

Try building a kit plane and stuff a 400HP motor auto engine conversion in it when the design calls for 160-180 HP......

I have a HUGE respect for test pilots now.........

Ps... So far, I have survived my creation..... I am either DAMN good... Or DAMN lucky.....
I hope I can type this same comment next year...

 

[nauga]

There's a reason why the job description is "test pilot." We desperately want these people to do these tests so we don't do them ourselves.

There were two test pilots and two flight test engineers on that flight and while I haven't seen a Gulfstream position description for either I'll bet I've got a pretty good idea of how it reads. The full final, contained in the accident docket, is worth a read for anyone remotely interested in flight test. - it is a very difficult read, both technically and emotionally. Borrowing from the final (but not the famous 'Gulfstream letter', also in the docket),

"The schedule pressure, combined with inadequately developed organizational processes for technical oversight and safety management, led to a strong focus on keeping the program moving and a reluctance to challenge key assumptions and highlight anomalous airplane behavior during tests that could slow the pace of the program. These factors likely contributed to key errors, including the development of unachievable takeoff speeds, as well as the superficial review of the two previous uncommanded roll events, which allowed the company‟s overestimation of the in-ground-effect stall AOA to remain undetected."

I'd bet a paycheck that the job descriptions for test pilot and/or FTE at Gulfstream do not mention the assumption of this type of risk. Had proper testing and analysis been conducted up to the point of this test the test would not have been conducted as it was. Test aircrew don't sign on to this kind of risk, they sign on to help minimize it. The issue is unfortunately not unique to Gulfstream; however, the G650 accident details are extensive and public domain and can be discussed openly.

Back to the OP, the type of performance data this test was intended to collect was a tiny corner of the flight envelope with failures. Performance testing of the type used to populate the POH perf charts is typically boring and uneventful if done correctly, with a few outliers for extreme cases.

Nauga,
and his W/delta time-suck

 

[cowman]

"The formulas" would be the hard part. The various performance tables would have to be turned in to graphs, which in turn would be made in to best fit polynomial equations.

I thought about creating a program to compute the most fuel efficient cruising altitude for a given flight but quickly realized that this would be the problem.

Margin of error would probably exceed any actual gain for a light single piston...

 

[Clark1961]

I thought about creating a program to compute the most fuel efficient cruising altitude for a given flight but quickly realized that this would be the problem.

Margin of error would probably exceed any actual gain for a light single piston...

It's all in the error function...

And yup, for spam cans climb to 8,000 or so unless the climb takes more than 1/3 yer flight time...

 

[nauga]

I thought about creating a program to compute the most fuel efficient cruising altitude for a given flight but quickly realized that this would be the problem.

Margin of error would probably exceed any actual gain for a light single piston...

What am I missing...do you have cruise charts/tables for your airplane already? If the data are already available over a range of conditions it's fairly straightforward to build a multi-dimensional 'model' from the tables and interpolate for conditions in between those provided in the charts. PM if you'd like to discuss specifics.

Nauga,
and his n-D lookups

 

[exncsurfer]

Hmm, I was just looking at a 172 POH, and all the tables and graphs are based on max weight. I don't see any way to estimate performance with a weight variable. All you get is max weight fixed, variable pressures, temperatures, altitudes, throttle percentages.

I'm not sure what the OP is trying to figure out, but I think variable weight might be important. Like what altitude should I fly if i'm at x weight to get best fuel economy. Or would best altitude/power setting for max weight always be best altitude/power for any weight?

 

[tmyers]

look at this one greg https://play.google.com/store/apps/details?id=com.edgemontsystems.TakeoffLanding

 

[poadeleted21]

It's a combination of both. Most any decent airplane design or performance textbook will have the basic equations for any type of airplane performance most of us would be interested in. The tricky part, as some have mentioned, is filling in the blanks for all of the aerodynamic, powerplant, mass properties, etc... information that's required for an accurate result. Most manufacturers of any size will develop a 'mathematical model' of an airplane and use that along with the everyday atmospheric properties and stuff in the performance equations to estimate performance, then validate or 'anchor' the model with flight test data. Once the model has been validated, or proven accurate, performance at most any condition can be calculated without having to explicitly test it. It's typically the validated model that's used for POH/OpMan graphs and tables, not strictly flight test data.

In the case of the Gulfstream accident N801BH refers to, the test point that lead to the crash was to collect data to establish an 'endpoint', or an extreme of the flight envelope, specifically singe-engine continued takeoff performance and limits. Gulfstream's submittal to the investigation points to a model that was not sufficiently validated, exercised, or analyzed (or some combination of all 3) prior to this test point. They state as much but not explicitly using these words, and the company clearly and candidly accepted responsibility for the accident. They were not the first and they unfortunately will not be the last. There but for the grace of God and a strong sense of self-preservation go I.

Nauga,
who lost a friend that day

As a side note, the G500 flew for the first time in Monday. This article describes some of the test methods they used before electing to fly it.

http://www.forbes.com/sites/businessaviation/2015/05/19/gulfstream-g500-logs-first-flight/

 

[Sac Arrow]

I thought about creating a program to compute the most fuel efficient cruising altitude for a given flight but quickly realized that this would be the problem.

Margin of error would probably exceed any actual gain for a light single piston...

Deriving the equations isn't extremely difficult. When I was in grad school, I wrote a program (in Turbo C) to calculate pump performance from the published curves. The input variables were several points on the curve, flows, and hydraulic gradients. The curve was generated by a polynomial equation fit routine, which could replicate them accurately enough to match actual conditions.

I didn't "write" the equation generator, but I found the code and adapted it. I actually did write the matrix inversion routine (for solving simultaneous equations) in Fortran and translated it to C for the program.

 

[cowman]

Ok, I will admit right off my math skills are not my strong point so maybe this is easier than I think. I'll have to dig out the POH or find one online. My airplane is a 78 Piper Archer II... pa-28-181. All the performance data is in charts- the type where you have to follow a reference line over, then another, then find your point on the graph. Then do it 3 more times and 3 different answers because the darn plotter always slips.

I'm really too busy to start a project until later in the year but I figure if one could model all that, pull in temperature/wind data, and figure the route distance up one could find out the optimal fuel economy altitude.

The problems I see with this are
1. I'm VFR and my altitude is usually chosen due to factors other than ideal fuel use.... eg getting to a cruising altitude vs cloud level
2. I highly suspect the error margin would make this a lot of work for little advantage but maybe I'm wrong.

 

[nauga]

I'm really too busy to start a project until later in the year but I figure if one could model all that, pull in temperature/wind data, and figure the route distance up one could find out the optimal fuel economy altitude.

If you already have performance curves a function or series of functions to calculate performance at any condition within the range of the charts should be relatively straightforward. Once you have that there are many ways of finding a near-optimal solution - some are really easy depending on how you'd like to implement your tool (NPI). You would probably want to roll in climb performance too so you don't lose all of your cruise perf gains in an extended climb. Still not difficult if you have decent perf charts already. Your comments re: margin of error (aka uncertainty) are well taken, but a decent algorithm is likely to get you closer than if you just eyeball charts with a plotter. The savings may not be as much as you think even if you *do* find the optimum, however. Even if you're constrained by ATC or FARs you can still use a perf calculator to get a little more precision (and hopefully accuracy) than from charts and eyeballs on the fly.

I'll leave my offer of several posts ago standing for the time being, PM if you'd like to talk about what's involved in putting something together.

Nauga,
and his flying Jacobians

 

[flyingcheesehead]

Greg,

I'm curious - Do you have something like this for the 73? Or is that why you asked?

 

[Greg Bockelman]

Well that is a function of the load planners at UAL I believe.

It is just that the performance data is so sparse in the 170 manual that extrapolation errors are more than possible.

Besides, in this digital age, why hasn't someone developed something like this?

 

[N801BH]

Well that is a function of the load planners at UAL I believe.

It is just that the performance data is so sparse in the 170 manual that extrapolation errors are more than possible.

Besides, in this digital age, why hasn't someone developed something like this?

The answer is two fold...

First, there is not enough accurate data to formulate a graph.

Second... the 170 is a certified plane and the liability to create that and let other people use it is too risky...

IMHO....

 

[Clark1961]

Besides, in this digital age, why hasn't someone developed something like this?

Many industries live in the past and if there is no money in it then why bother? Management certainly isn't motivated to make anything easier unless they can easily see that the change will make them more money.

I'm in the middle of converting a small business over to data based management right now. I'm doing 'cause it makes life easier for me and it will allow the business to expand. It was pretty much maxed out with paper based management tools.

 

[flyingcheesehead]

It is just that the performance data is so sparse in the 170 manual that extrapolation errors are more than possible.

Besides, in this digital age, why hasn't someone developed something like this?

Like most things in aviation, probably because of the possibility of huge liability exposure with limited ROI.

In addition, it's not all that easy to do.

It sure would be a neat tool to have though! I've thought about doing something similar that would adjust V-speeds for weight, etc... But there's always a lack of time.

 

[denverpilot]

Just a question, I don't know. Isn't the perfect flight for maximum efficiency sans-winds aloft, a parabola that doesn't exceed Vne while wide open on the throttle in the descent and such that you run out of fuel within gliding distance of the runway?

Basically it seems to me like its take an awful lot of flights to make the software worth much. Maybe save a tank of gas in a few years time? It's not really surprising no one is interested in the headaches of writing, debugging, and maintaining it for what amounts to $5 a copy, if that.

 

[Sam D]

This looks like an analog version for some of the requirements...

https://www.sportys.com/pilotshop/takeoff-performance-computer.html

 

[flyingcheesehead]

Just a question, I don't know. Isn't the perfect flight for maximum efficiency sans-winds aloft, a parabola that doesn't exceed Vne while wide open on the throttle in the descent and such that you run out of fuel within gliding distance of the runway?

No.

Aerodynamically, the most efficient is a particular constant angle of attack. That means you would establish a climb at max power, climb at a relatively constant speed (you would actually climb at slightly faster airspeeds as you burned off fuel in the climb, but it might not change significantly), and then you'd slow down throughout the flight as you burn fuel and lighten up to maintain that angle of attack.

Of course, in a piston-powered airplane, that's not necessarily the best course of action for overall efficiency because as you pull the throttle back you increase pumping power losses in the engine which would probably defeat at least part if not all (or more) of the aerodynamic efficiency.

The number of angels that can dance on the head of a pin is... Wait, it doesn't matter! Until we get accurate winds aloft at an altitude resolution of no more than 500 feet, all of this theoretical efficiency will probably be canceled out anyway!

I am, however, interested in doing some flight testing of climbs, glides, and cruise speeds at varying altitudes to determine parameters that, combined with winds aloft, would give the fastest and most efficient (two different things) cruising altitudes. I've read the Vz paper and I'm tending to use that for cruise climbs now, but I'm not convinced it's actually the most efficient.

Once that's done, maybe I could write an app... But man, it'd take a ton of time to do it how I'd like to do it, and it'd be awfully difficult to even get minimum wage return on that time.

 

[jesse]

No.

Aerodynamically, the most efficient is a particular constant angle of attack. That means you would establish a climb at max power, climb at a relatively constant speed (you would actually climb at slightly faster airspeeds as you burned off fuel in the climb, but it might not change significantly), and then you'd slow down throughout the flight as you burn fuel and lighten up to maintain that angle of attack.

Of course, in a piston-powered airplane, that's not necessarily the best course of action for overall efficiency because as you pull the throttle back you increase pumping power losses in the engine which would probably defeat at least part if not all (or more) of the aerodynamic efficiency.

The number of angels that can dance on the head of a pin is... Wait, it doesn't matter! Until we get accurate winds aloft at an altitude resolution of no more than 500 feet, all of this theoretical efficiency will probably be canceled out anyway!

I am, however, interested in doing some flight testing of climbs, glides, and cruise speeds at varying altitudes to determine parameters that, combined with winds aloft, would give the fastest and most efficient (two different things) cruising altitudes. I've read the Vz paper and I'm tending to use that for cruise climbs now, but I'm not convinced it's actually the most efficient.

Once that's done, maybe I could write an app... But man, it'd take a ton of time to do it how I'd like to do it, and it'd be awfully difficult to even get minimum wage return on that time.

I was pretty surprised when coming back on Southwest, Chicago to Omaha, the pilot said on the announcement that we'd be cruising at 40,000 ft. I get the high is efficient thing but on a hour flight I wouldn't have thought we'd take the 737 that high.

 

[denverpilot]

No.



Aerodynamically, the most efficient is a particular constant angle of attack. That means you would establish a climb at max power, climb at a relatively constant speed (you would actually climb at slightly faster airspeeds as you burned off fuel in the climb, but it might not change significantly), and then you'd slow down throughout the flight as you burn fuel and lighten up to maintain that angle of attack.



Of course, in a piston-powered airplane, that's not necessarily the best course of action for overall efficiency because as you pull the throttle back you increase pumping power losses in the engine which would probably defeat at least part if not all (or more) of the aerodynamic efficiency.



The number of angels that can dance on the head of a pin is... Wait, it doesn't matter! Until we get accurate winds aloft at an altitude resolution of no more than 500 feet, all of this theoretical efficiency will probably be canceled out anyway!



I am, however, interested in doing some flight testing of climbs, glides, and cruise speeds at varying altitudes to determine parameters that, combined with winds aloft, would give the fastest and most efficient (two different things) cruising altitudes. I've read the Vz paper and I'm tending to use that for cruise climbs now, but I'm not convinced it's actually the most efficient.



Once that's done, maybe I could write an app... But man, it'd take a ton of time to do it how I'd like to do it, and it'd be awfully difficult to even get minimum wage return on that time.

Thanks. I'm not aerodynamically intelligent.

Like you said, sounds like a lot of variables to account for that won't amount to much of a savings for the typical flight.