Page 1 of 2

YSFlight Fuel Burn Test

Posted: Thu Aug 27, 2015 12:14 am
by decaff_42
Hello all,

As I mentioned a couple days ago, and in my request for info on the .yfs file parameters, I am conducting tests to determine just how YSFlight burns fuel at different throttle settings and also how throttle settings vary with speed at different altitude. Additionally I want to see if changing the initial fuel weight will have any effect on how the airplanes perform.

I am initially doing this with my 787-8 airplanes, which share a common dat file (in terms of performance).

I plan on testing at the following altitudes:
1k, 5k, 10k, 15k, 20k, 25k, 30k, 35k, 40k

At each altitude I will be engaging the straight & level autopilot at the following speeds, if I can actually fly that slow and / or fast. Note that the speeds are in KTAS, as KIAS changes with altitude.
200, 225, 250, 275, 300, 350, 400, 450, 500, etc 50 KTAS increments until Maximum speed.

To test the weight effect I am using 75% and then 25% starting fuel weights.

Each test is a 0.3% fuel burn, which is timed in order to understand how fast the fuel is being burned.

I have completed the 75% fuel, 1k, 5k, and 10k test points and I thought I would share the data.
throttle vs speed.png
throttle vs Fuel Burn Rate.png
The fuel burn, aside from some weird bumps, appears to be fairly steady with throttle position, but I can fly faster with less throttle the higher I fly.

Granted this is only preliminary data, but It is interesting. I will have to do far more testing (~200 tests) to just get data at all the possible combinations. I would like to get at least 3 trials at teach test condition, but that may not be feasible to do myself.

EDIT: I Should mention that I have completed 28 test points with these three altitude bands.

YSFlight Fuel Burn Test

Posted: Thu Aug 27, 2015 12:44 am
by Lor599
This is indeed interesting:
but I can fly faster with less throttle the higher I fly.
Very useful to know. Especially for extremely long flights like...Kodiak Island to Guam on the Guam to SoCal map :lol:
Nice graphs too!
Why is there a spike though...on the 1k for first graph, and the 10k on the second? As if the speed spiked a huge amount?

YSFlight Fuel Burn Test

Posted: Thu Aug 27, 2015 4:27 am
by Gunny
I have been doing my own fuel burn tests as well but no where as scientifically as you decaff.I take the fuel consumption numbers from the manufacturers specification and reduce them to decimals lb.For instance if the fuel consumption is 105 lb/ph I set the dat fuel burn to 0.105lb and I get a reasonably good fuel rate for prop engines.As for the heavy's I will leave that to you my friend.But as strictly anecdotal information weight does effect fuel rate and handling. the lighter the airplane the faster the climb rate as well as turn rate.Just some observations I have noted.

YSFlight Fuel Burn Test

Posted: Thu Aug 27, 2015 6:56 am
by Copperhead
Proof of the 65,000ft barrier.

YSFlight Fuel Burn Test

Posted: Thu Aug 27, 2015 12:41 pm
by Dragon_Mech
i've know for some years now that the higher you fly, the faster you go with less throttle. the magic altitude seems to be 40,000ft. anything higher than 41,000ft and your speed starts to drop and you start loosing manuverabilty. but i wounder if setting the crusing altitude in the .dat file higher will affect this?

YSFlight Fuel Burn Test

Posted: Thu Aug 27, 2015 2:58 pm
by Swift
Copperhead wrote:Proof of the 65,000ft barrier.
yeah, even the Talon MOD had trouble flying above that altitude, turning was a nightmare even tho I was testing it at mach 6-ish.

EDIT: I got bored and watched stealth a dozen times and remade the talon's FDM.

YSFlight Fuel Burn Test

Posted: Fri Aug 28, 2015 5:30 pm
by decaff_42
The data I show is really, really basic and not really a good representation of what is needed for a true statistical study. The graph only represents ONE flight at that condition. Ideally I'd like 3 or more, but time is a premium when each altitude takes about 1 hour to test.

I'm developing an excel tool that will let me get a better average thrust from the .yfs files I'm saving from each flight. I hope that will eliminate a bunch of errors.

My best guess for the bumps are (1) error on my part with my stopwatch (2) increasing Angle of attack which requires more fuel burn to maintain the lower speeds. The latest trial I did at 15k saw and AoA of about 9-10 degrees for the 200 KTAS point. That will be skewing later data when I am able to post it.


Just finished the 30k test points. The trend is pretty pronounced now. The fuel burn rate seems very linearly related to the throttle position. The bumps seen previously were more likely than not user error.
throttle vs speed -30k.png
My excel tool makes getting the throttle position really easy. Just copy paste and press a button. I plan on adding an import function and getting some other data like Angle of Attack and velocity.

YSFlight Fuel Burn Test

Posted: Sat Aug 29, 2015 2:20 am
by waspe414
It'll be interesting to see how that trend regresses above 40,000ft.

YSFlight Fuel Burn Test

Posted: Sat Aug 29, 2015 2:34 pm
by decaff_42
Based on how Mach changes around 38k feet, I would say there should be diminishing returns. I have an email in to Soji about the equations for the throttle position. There are a ton of DAT variables to consider:

Fuel weight, clean weight, payload weight, fuel burn rate, Wing area, critical speed, max speed,Cruise speed, cruise altitude, cruise throttle, landing speed, landing AoA, Landing throttle... etc

I'm interested in looking at the Lift and Drag coefficients that are defined by the Dat file. Would anyone happen to know more about that? That would help a TON with figuring this all out. Once that is defined, I can take a look at modeling the drag force that the airplane will have to deal with and get throttle position.


I just completed the 35k test points and found something interesting. In the Dat file, we can define cruise conditions for altitude, speed/mach, and throttle. I specifically tested the cruise speed in my dat file and found the actual throttle to be very different from the input dat throttle. I had 65% in the DAT and 49% in my testing.

YSFlight Fuel Burn Test

Posted: Sat Aug 29, 2015 3:07 pm
by Dragon_Mech
decaff42 wrote:but I can fly faster with less throttle the higher I fly.
you do know that air(and the atmosphere) gets thinner the higher you go right? and as a result, there is less friction and drag to overcome. this is why it takes less fuel to fly at 30k than it does at 3k. it looks like soji did very well in simulating this in YSFlight.

YSFlight Fuel Burn Test

Posted: Sat Aug 29, 2015 5:12 pm
by decaff_42
Yes I know this. I have a degree in aerospace engineering. The thing I was pointing out was that the throttle position decreased with altitude while retaining the same speed. At that time I did not have enough data to conclusively say that the fuel burn rate was linearly dependent on the throttle position and I assume, thrust. The variations in the initial fuel burn curve are most likely errors I introduced, but the trend later on, with more practice and test points, shows a linear relationship.

The overall aerodynamic simulation from Soji is fairly basic. IRL the thrust profile of a turbofan engine is a decreasing curve with altitude. One thing I have not figured out is how the drag calculations are made based on the DAT file, as I cannot find a cross-sectional area. I assume there was some sort of assumption made, perhaps with the wing area for this.

Also, I have gotten up to 40k feet. Feast thine eyes on DATA!!!!
throttle vs speed -40k.png
The interesting thing I have noticed is that the left-most points on the curve seem to fall on a roughly linear line (I had to interpolate a few). This shift is what I believe to be the influence of the altitude. The general curve starts at around 15k, but the 10k points appear to also follow this trend if I slowed down enough for sub 200 KTAS trials.
Screen shot 2015-08-29 at 12.42.45 PM.png
Screen shot 2015-08-29 at 12.42.45 PM.png (15.88 KiB) Viewed 4123 times
In addition to the raw data, I am trying to figure out the CL/CD values based on the DAT file. I have the calculations in excel, but they are odd to say the least. I'm glad that the CL changes at the lower altitudes as the landing approach is 5 deg AoA, so I think that might be the C_L curve, but that introduces a new parameter that I have to consider in the general process. I think this is the right path however, because the Dat File REM line mentions CLzero, which might mean CL at 0 deg AoA, which is a really important point on any CL curve (along with CL = 0 and Max CL).

If this is true, it might also explain why some airplanes (like the CHF 777 and my Boeing 717) fly nose-down (negative AoA) at really high speeds as the CL does not need to be as high in order to maintain lift as the increased speed is taking care of that. While a typical lift curve stalls a little past max CL, YSF stalls based on a Max input AoA value in the DAT file. I have not been able to break this limit while in steady-level flight. I think this might be the top of the CL 'curve'. I will have to go back and find AoA data for each of the flights, but I will want to automate that with the excel tool after I update it to stop giving me bad throttle data. (~1% error because of my own bad assumption.)

The general lift equation is:
L = 0.5 * rho * v^2 * A * CL
The general drag equation is:
D = 0.5 * rho * v^2 * A * CD

I am not certain how the drag function is calculated in YSF because the Dat file mentions a CDconstant is defined by the landing condition.

CL = Coefficient of Lift
CD = Coefficient of Drag
rho = density
v = speed
A = reference area (plan-form for lift or cross-section for drag)

Re: YSFlight Fuel Burn Test

Posted: Sat Aug 29, 2015 6:28 pm
by Bombcat
Massively impressive work here, Decaff. I look forward to seeing how this may let us improve dats in the future.

Re: YSFlight Fuel Burn Test

Posted: Sun Aug 30, 2015 7:48 pm
by decaff_42
Well, I have run into some issues with my theory about the Lift Coefficient. Namely that it doesn't match with my experimental data. I'm not sure how Soji did it, but now that my excel tool can calculate the Angle of Attack, the CL's don't match up very well at all. (2.5 vs 53)

I did one test at 25% fuel load to see if it would have an impact, and that one point was almost identical in terms of throttle position (75% @ 50.6 vs 25% @ 50.8) but the Angle of Attack was almost 1.2 degrees higher for the heavier weight. This shows me that the fuel weight has an impact that needs to be investigated. The weight difference was almost 123,000 pounds (55800 kg), which is around 25 percent lighter than the 75% state. I find it interesting that the lighter aircraft needed more throttle to maintain the same speed.

Also, if Anyone is interested in helping me gather data points, please let me know! I could use a hand with all the variables that need to be tested. I have 100 points already flown and will probably fly another 20-30 for the remaining 75% fuel loads. After that there are a ton of other points to tackle at different fuel states! Even if you can only do a few a week, it will be helpful!

Re: YSFlight Fuel Burn Test

Posted: Sun Aug 30, 2015 11:51 pm
by Lor599
I'll be willing to help name a variable, and I'll fly them.
Also before I dive in, could you give me an example of one, just for an idea of how long it will take me to gather the data, etc. I will probably be busy all of next week until Saturday, but I can still manage to take care of some anyways. <----- That depends on a lot of things, but yeah.

Re: YSFlight Fuel Burn Test

Posted: Mon Aug 31, 2015 1:08 am
by decaff_42
Here is a procedure I wrote up! I have also added a sample of the test data I collect and record manually and a collection of a few YFS files for you to see how the naming system works for those. (It is also stated in the procedure files)

The procedure has a word and PDF file which are identical, so read whichever one is easier for you to open. If you have any questions, don't hesitate to ask me! I want this to be as clear as possible!

If you could start with 50% starting fuel test points, I will be working on the 25% test points. When you are selecting the airplane, lower the initial fuel to 50%. I would recommend starting with a lower altitudes and speeds first and working up the speed range at each altitude and then when that is maxed out, move to the next altitude.

Don't be afraid to fly a few practice flights. I had about 10 practice flights before I got really good at it. For these practice flights, include one at 1k, 200KTAS, and one at 10k, Max Speed. This will get you familiarized with both ends of the spectrum. Do not be afraid of large Angles of Attack when using the auto pilot. I have flown points with 12.9 Degrees Angle of attack, although it felt higher than that.

I typically run the simulation using the coarse graphics, or the FAST simulation, but use what graphics settings you want.

Thank you for your help!

Edit: Forgot to include the Test map. Nothing fancy, just ocean and a whole bunch of start points.
(1.15 MiB) Downloaded 71 times

Re: YSFlight Fuel Burn Test

Posted: Fri Sep 04, 2015 10:45 pm
by decaff_42
A second post because the starting package now includes a dedicated 787-8 test bed aircraft!

I have also expanded the YFS naming scheme to also indicate the percent flaps used. I think this may prove to be useful because in theory the fully deployed flaps add a certain amount to the CL of the aircraft. If we can see differences in the AoA and throttle with different flap settings (0, 25, 50, 75, 100) using the w and s keys (default setup) then perhaps we can directly see what the CL is.

If anyone would like to help, simply download, read the instructions and get flying!
Download Here!

I have currently completed 124 test flights with a total of 7.1 flight hours. The baseline 75% fuel trials are completed and I am currently on the baseline 25% fuel trials.

Re: YSFlight Fuel Burn Test

Posted: Thu Sep 17, 2015 11:49 am
by decaff_42

Been doing testing on and off for the last 2 weeks and I decided last night to try something a little different. Start at full speed, with a full load of fuel, but set the fuel burn to 2000 pounds per second. Burn off the fuel and let the aircraft glide to a stall. This gives a nice look at the angle of attack vs speed, throttle, and all sorts of goodies. Then I went to town and did a lot of post-flight calculations to determine AoA, which is different than I previously thought as I now calculate it as the difference between the flight path angle and then aircraft pitch rotation. Using a rough fuel burn rate calculation (~2% error) I get the aircraft's weight at all points in flight. This gives me the weight.

Using the weight and the thrust force in the vertical direction, gives me how much lift the wings actually need to produce. Since I know the speed, air density (from altitude), wing area and lift force, I can calculate the aerodynamic lift (see force N in the image below). I have done two trials like this, and in both cases, the CL v AoA curve is a straight line (like my original hypothesis) and doesn't get wonky (stall) until you go above the Max input AoA as defined by the DAT file. :D The trend lines I put in with excel had a linear fit R^2 value of 0.993 or better (1.0 is a perfect fit)


Here are the two trials I did with this method of testing. Any comments or ideas for how to explain the differences in CL v AoA is appreciated!
AoA v CL & KTAS, 35k Vmax Decel.png
AoA v CL & KTAS, 35k 300KTAS Decel.png

Re: YSFlight Fuel Burn Test

Posted: Sat Sep 26, 2015 3:14 pm
by decaff_42
These slow-down tests really do help! I have confirmed the following:

(1) The lift coefficient is a straight line with respect to the angle of attack of the aircraft (Calculated from aircraft orientation and flight path, both with respect to the horizon).

(2) YSFlight uses Standard Day atmospheric conditions for density. My y-intercepts between altitudes were withing 0.023 CL, which can probably be attributed to a lot of different things, but a mere 2% error for this is nothing to sniff at.

What I still need to figure out is how the lift coefficient is calculated from the DAT file.

Re: YSFlight Fuel Burn Test

Posted: Mon Sep 28, 2015 1:34 am
by decaff_42
I have my first draft of the Fuel Burn Testing Report written up to cover the first three 'discoveries' made in this endeavor. There is a lot of text and equations for a mere 5 pages, but I hope you can read through it and understand what is going on.

If you spot any errors (grammer, spelling, etc) or find something confusing, please let me know! I'm using a LaTeX template to get the scientific report look, and my LaTeX editor only has a basic spell check, no grammar checking.

Read the report draft here!

Re: YSFlight Fuel Burn Test

Posted: Sun Feb 21, 2016 10:13 pm
by decaff_42
The analysis has been expanded. Now the fuel burn and thrust in both afterburning and non-afterburning modes of operation have been analyzed.
Decaff_42 wrote:YSFlight Aircraft Thrust and Fuel Burn data has been analyzed and the report is written! Get it hot off the presses!