Re: Flight Models for newbee's

> From: "Correu PelDavid"
> Hi, 
> 
> I'm studying theory control, and I'd like to play a bit with helicopter's capabilities.
> I'd like to work on their model.
> Flightgear works with YAsim, JSBsim and LaRCsim.
> I've take a look at the LaRCsim manual. It says it uses "a set of equations of motion for a rigid-body aircraft in atmospheric and low-earth orbital flight".
> In fact, using the equations of a 6D rigid-body is where I would have started if I were to choose without more info.
> 
> I wonder how the other models work.
> For instance, I know the X-plane engine considers the aerodynamic planes. A totally different point of view from LaRCsim, but still accurate and working.
> 
> Where could I get some info on YAsim and JSBsim? 
> Could you draw me a shot of the main ideas?
> 

It would be helpful if there was some documentation readily available in the source or base package that gave a nutshell description of each flight dynamics model library.  This sort of   question gets asked now and then.

JSBSim is fairly well documented.  I'm not sure if Andy Ross is reading and will answer but I'll give a brief description of what I understand about YASim.

The parts of the code that I've worked on (doing very minor fixes) seem to be using a collection of well known (as in I could find some of them on the Internet) formulas for calculating things like lift, drag, thrust, prop rpm, etc. based on weight/mass and geometric data.  Some of the calculations use very general data points about performance capabilities of the aircraft (e.g. stall speeds, cruise speeds) that are readily available for just about any model airplane.

You can get an idea of the scope of these calculations by examining some of the configuration files for the Aircraft that are currently using the YASim FDM.  My unskilled assesment is that the YASim can reasonably simulate a wider range of Aircraft operations better than JSBSim but a lot depends on how much data you give JSBSim.  In the end a JSBSim flight model is potentially much more faithful to the original aircraft given detailed enough data.

A very general distinction between using the two FDMS, in my view would be this:   If you want to make a very accurate simulation of a single aircraft for which you have access to a fair amount of information on (especially wind tunnel tests) then JSBSim is definately the way to go.   If you want to put together a reasonable simulation fairly quickly and/or the data you have is limited to general performance and geometric measurements, YASim is definately the way to go.  The laws of diminishing returns seem to kick in a little quicker refining YASim model configurations, but you can get something quite reasonable with less.  JSBSim models can be refined and tweaked and with greater attention and time can be made quite accurate.  If neither of those senarios matches exactly what you want to do, then do a little more research and see what you can get for data, and how you wish to approach 
 your project.

At the moment it seems that YASim is further ahead in terms of Helicopter simulation.  And I would even suggest that if one were interested in examining control (as opposed to flight dynamics) with maybe a variety of helicopter configurations, that YASim might be better suited anyway.

As you know from previous discussion,  the helicopter simulation isn't all that accurate.  In my opinion, the biggest problem (from a control perspective) is the lack of ground effect modeling.  The effect of the collective control (rate of accent/decent) should be significantly influenced by distance from the ground during landing, takeoff (or any extended hovering close to the ground).

On the issues of transational lift modeling, I must confess it has been a while since I've experimented with the helicopter simulation so I cannot remember exactly what was wrong.  The lift effect is basically working.  Also I think the tendency for the aircraft to rotate goes away as you transition to ETL as it should.  I do not think asymmetrical lift (the side where the blades are moving forward produces more lift) is modeled.  Just to check for basic operation, you might try some of the obvious things like pitching the aircraft up and down to see if it climbs/falls and decelerates/accelerates in response.  The accuracy of these effects might require a little improvement but if translational lift is working you should at least see something reasonable happening.

In my opinion the other issues, like the lack of autorotation, pale in importance compared to the omission of ground effect modeling.  This is a critical element for fixed wing aircraft modeling,  and it is just as or even more so for helicopters because the behavior of that cushion of air is so central to their operation.


Best,

Jim


-- 
Jim Wilson
Kelco Industries
PO Box 160
Milbridge, ME 04658
207-546-7989