Lightning FDM

16 Nov 2018 11:29 #40917 by StuartC
It does need a tweak to work fully with the latest yasim versions.

Please Log in or Create an account to join the conversation.

16 Nov 2018 12:36 #40919 by Algernon
It's very difficult to know whether a pilot would feel a YASim - or for that matter ANY FDM - was a realistic representation. But we do know that YASim isn't likely to be super good at fast jets, especially those with hardly any wing and which can do crazy high mach numbers. I personally feel the current FDM doesn't quite generate enough lift at the right airspeeds, but I think it's the best we could do and it matches the performance envelope fairly well.

Stuart is right, it definitely needs a tweak or two - it's throwing up an error regarding elevator trim, and even if it wasn't, I'm sure it could benefit from the improvements made to YASim lately.

Please Log in or Create an account to join the conversation.

16 Nov 2018 15:11 - 16 Nov 2018 15:12 #40921 by ScottBouch
Hiya,

I was asking mainly out of curiosity having no knowledge of how the FDM / YAsim works.

Was wondering what data can be picked out of the operating data manuals and used? examples attached.

I have the operating data for the: F1, F1A, F2, T4, and F6 (Algy has copies).

The T5 cockpit shape gave it the slipperiness of the T4 (area rule - a nod toward the Buccaneer coke bottle shape), but with the more powerful engines of the F6 (30201 Avons) - due to these factors it actually was the quickest of all the Lightning Mks.





Cheers,
Attachments:

Please Log in or Create an account to join the conversation.

16 Nov 2018 16:49 #40923 by enrogue
Yasim isn't really designed for supersonic FDMs - and any supersonic wave drag effects have to be done manually via a variable drag item + scripting (I've done this with a couple of subsonic/transonic aircraft linearly)

It's one of the many things yasim needs...
The following user(s) said Thank You: ScottBouch

Please Log in or Create an account to join the conversation.

16 Nov 2018 19:44 #40927 by ScottBouch

Please Log in or Create an account to join the conversation.

16 Nov 2018 19:58 #40928 by StuartC
FG is capable of industry level flight models, not just home simming.

Please Log in or Create an account to join the conversation.

16 Nov 2018 22:17 #40929 by timi
www.x-plane.com/desktop/meet_x-plane/
“X-Plane contains subsonic and supersonic flight dynamics, allowing users to predict the flight characteristics of the slowest aircraft or the fastest.“

Please Log in or Create an account to join the conversation.

17 Nov 2018 02:53 #40930 by Algernon

timi wrote: www.x-plane.com/desktop/meet_x-plane/
“X-Plane contains subsonic and supersonic flight dynamics, allowing users to predict the flight characteristics of the slowest aircraft or the fastest.“


That's what X-Plane themselves say - it's not what I've heard from our members who've tried it out.

Stuart is right that FG is capable of industry-standard simming - however, I'm fairly certain YASim is not it. To my knowledge, all the "serious" projects use something else. JSBSim is generally pretty highly spoken of, but requires quite a lot of aerodynamic data, data that in most cases, we don't have.

Please Log in or Create an account to join the conversation.

17 Nov 2018 11:17 #40931 by ScottBouch
Did i send you the operational data on the lightnings Algy? I sent you everything T5, but may have misseed this as its for other mks.

Its hundereds of pages of aerodynamic data, that may be of use to JSBsim? But i guess that's a mammoth project in itself?

Please Log in or Create an account to join the conversation.

17 Nov 2018 12:07 #40933 by enrogue
Yasim is awesome for getting an FDM up with very little data - it calculates the inertia tensor, centre of gravity, henning recently added cog range based on wing plan, and does a decent job at the points with the important data (cruise & approach)

But it's missing quite a few of the non linear quirks of aerodynamics & aircraft:

- I'm sure it's not doing damping due to rotation about an axis correctly, which affects stability
- It doesn't know anything about the transonic regime & the effects on lift & drag, but you can kind of model it with a scripted spoiler (ideally you would need 2, but the fdm only allows 1 per wing element)
- again due to no transonic effects, there is no movement of centre of pressure, so no trim change - no idea how to fix this
- the jet engine model is extremely simple & doesn't do the supersonic regime (hence the high n values when supersonic) - this affects fuel usage & thrust when supersonic (you can work around it with atsfc & exhaust speed values)
- the gear system doesn't have any weight attached to it so the change in inertia has to be done with weight points
- the gear has drag values attached to the compression value, so the change in trim with gear down is usually wrong

hopefully some of these can be fixed (poke henning as he has a better understanding of the code), but also have a look at VSPAero to output JSB fdm data - it seems to be the tool of choice for detailed FDM work when you have limited data at the moment

Please Log in or Create an account to join the conversation.

17 Nov 2018 12:43 #40934 by timi
I wouldn’t have the expertise nor access to the source code for X-Plane to make a good judgement. But I guess the likes of Richard could just by observing the behavior.

Please Log in or Create an account to join the conversation.

22 Nov 2018 12:25 #40998 by Richard
As previously said yasim is limited in what you can achieve; it's built around the concept of landing and cruise and figures out the rest, so you get reasonable flight characteristics that meet performance at two points in the flight regime. This is well suited to but airliners however with a effort it's been proven that you can model fighters (ref; FGUK Tornado, buckaroo's work on the Mirage2000).

For FG there are two other options, JSBSim and UIUC[1]; however I'm not going to cover UIUC as it is covered elsewhere. Both JSBSim and UIUC can provide a linear or non-linear aerodynamic model.

It's important to understand the difference between linear and non-linear in this context. Basically a linear model has a single factor that is used throughout the flight envelope; and a linear model will become unrealistic once the curves change direction (e.g. alpha > 10). Non linear usually uses lookup tables to provide curves that allow the forces and moments to be calculated accurately across the whole flight envelope (although sometimes only limited data is available so you may find aero models that start to go wrong at higher alphas).

JSBSim aerodynamic data (lookup tables) as it uses the coefficient buildup method. Getting this data is the hard part; but there are tools to help - however these tables are aerodynamic coefficients so it really helps to understand aerodynamics at the mathematical level. There are often reports on NTRS that provide aerodata for older US aircraft and sometimes you'll get lucky and find the aircraft you're trying to model.

To create the JSBSim model we have tools to help us; I've covered the computational methods in my article[2] - but it's also worth mentioning that Aeromatic++ is a lot better than aeromatic where data isn't available, but OpenVSP is my favoured approach. For the record OpenVSP can be considered to be CFD but you'll get much better results with CFD using something like Navier Stokes in a full cell based model however this is going to take months of CPU time; and is considerably harder to setup than a VSP model.

X-Plane with its blade element modelling does things differently and it sits somewhere between yasim/aeromatic and an OpenVSP model. I've attached research document that covers this in more detail. I've always found the X-Plane models somewhat lacking - but it has been a few years since I really looked into this. Generally the longer a model takes to generate the better it is, so as X-Plane does this dynamically it's obviously not going to give as good results as something that has taken 100hours of CPU, but X-Plane probably also has a bunch of tricks to make things appear better than they actually are.

Important things for an aero model are

1. Accurate geometry
2. Accurate mass balance data (including moments of inertia)
3. Aerodynamic data, but particularly the derivatives.

[1] m-selig.ae.illinois.edu/apasim/Aircraft-uiuc.html
[2] chateau-logic.com/content/using-vspaero-...el-jsbsim-flightgear

File Attachment:

File Name: X-Plane-F-...1-PB.pdf
File Size:1,676 KB
Attachments:
The following user(s) said Thank You: StuartC, enrogue, timi, ScottBouch, hans05

Please Log in or Create an account to join the conversation.

22 Nov 2018 14:26 #40999 by ScottBouch
Hi Richard,

Thank you for the help on understanding this topic!

I have placed the Lightning operating data on my server for you guys to download.

Would you mind taking a look at this data to see if it would be useful to modelling the aerodynamics of these aircraft though one of the methods you suggested?


F1, F1, F2, and T4 Lightning:
www.scottbouch.com/AP-101B-1001,2,and,4-...4_Operating_Data.zip

F6 Lightning:
www.scottbouch.com/AP-101B-1006-16_F6_Operating-Data.zip


Unfortunately I don't have it for the T5 or F3, but (if this is actually useful) I guess an approximation may be made form the other Mk's data?

Many thanks, Scott.

Please Log in or Create an account to join the conversation.

23 Nov 2018 15:27 - 23 Nov 2018 15:27 #41014 by ScottBouch

Please Log in or Create an account to join the conversation.

24 Nov 2018 09:14 - 24 Nov 2018 13:43 #41020 by Richard
The operating data you linked to is generally not that useful for creating an aero model but is useful for testing an aero model.

However you're in luck as I found this naca.central.cranfield.ac.uk/reports/arc/rm/3558.pdf which is a good basis for an aero model.
The following user(s) said Thank You: ScottBouch

Please Log in or Create an account to join the conversation.

24 Nov 2018 11:43 #41021 by Porcius
Could somebody PLEASE, please, please , put the 't' in the subject title - it's sooooo annoying for me ! :)

Porcy

Lightning DOES strike twice.

Watchmen Quote - "I am looking at the stars.... Thier light takes so long to reach us. All we ever see of Stars are thier old photographs"

Please Log in or Create an account to join the conversation.

24 Nov 2018 12:27 #41022 by StuartC

Porcius wrote: Could somebody PLEASE, please, please , put the 't' in the subject title - it's sooooo annoying for me ! :)

Porcy


happy now ?
The following user(s) said Thank You: Porcius

Please Log in or Create an account to join the conversation.

24 Nov 2018 13:20 #41023 by Porcius
Much better now, ta muchly.

Porcy

Lightning DOES strike twice.

Watchmen Quote - "I am looking at the stars.... Thier light takes so long to reach us. All we ever see of Stars are thier old photographs"

Please Log in or Create an account to join the conversation.

26 Nov 2018 09:30 #41047 by ScottBouch

Richard wrote: I found this naca.central.cranfield.ac.uk/reports/arc/rm/3558.pdf which is a good basis for an aero model.


Hi Richard,

That is fantastic! Great find - I had no idea such data was publicly available, thank you for the education!

Cheers,

Please Log in or Create an account to join the conversation.

02 Dec 2018 14:46 #41115 by Richard
There is a lot more publicly available data out there than most people imagine; however the hard part is often finding it, and knowing what to look for; as often they don't mention the airframe (or whatever it is you're looking for) directly.

We got lucky finding that report; but I still spent an hour or so sifting through lots of irrelevant reports.

Please Log in or Create an account to join the conversation.

Time to create page: 0.224 seconds
Powered by Kunena Forum

PM Mailbox

You are not logged in.

Latest updated downloads