mikehoulder

Hi, all.
I've been concentrating on the navigation and radio instruments.
Mike

mikehoulder

A note on texture mapping techniques in AFPD

Subject to the amount of RAM memory installed on the computer being used, I have found that I can use a texture file up to 4096 x 4096 pixels, that is 48 Mb in size. I don't yet know what the upper limit is in AFPD.

In addition, copies of the same texture file that have been re-scaled by a simple factor can be installed directly in the selected AFPD aircraft folder without any other changes or, in particular, any other re-mapping of the textures.

For instance, in the first example below, the texture file had a size of 4096 x 4096 pixels. In the second example, the same file was reduced by a factor of 8, 512 x 512 pixels. All I did was copy the reduced file into the folder.

Mike

mikehoulder

Hi all, join with me drinking a half bottle of champagne!

I wanted the Lancaster to be at 1:1 scale relative to AFPD. But I had a lot of doubts that I would get there. It really is big at 31 metres wingspan. Compare it to the True Scale Antonov AN2 in the second image.
Then I've loaded the model with a whole heap of fine detail - taking the polygon count well past the normal. Would my display frame rate disappear?

Well, it's here at 1:1 scale and without any noticeable hit to my frame rate!!

But only a half bottle. It does fly just - like a very drunken duck; I haven't started on the TMD file yet. But I can't make any promises yet.

But here it is - the MIGHTY LANCASTER.
Enjoy
Mike

mikehoulder

I would welcome any criticisms or improvements to this paper.

Aerofly is a technical simulator. It can be used very productively at an immediate level – as a training aid to would-be rc pilots, as a planning and practice aid for competitive rc aerobatics or just for fun. But it goes deeper than that. Users can create their own model aircraft rather than just use those supplied with the simulator.

The creator of a simulator model has the option of making the model a single rigid structure out of all its individual components. But it is also possible to make it a dynamic structure where individual components react against each other under static and aerodynamic loads.

In either case, the model is seen as a hierarchy (upside down tree) of components with the fuselage as the top or the root of the tree. To the fuselage are joined the wings, the tailplanes, fin(s), nose or tail undercarriage unit, To the wings are joined engines, propellers and the main undercarriage units. To the undercarriage units are joined the wheels. Other components may be added such as a sliding cockpit canopy, military ordnance etc; but in general they are not structural items. The moveable control surfaces are not separate components; ailerons and flaps are part of the wing, the rudder is part of the fin etc.

Joint Strength, Stiffness and Damping
Whether the model has a rigid structure or a dynamic one, any joint between components must be defined as an additional separate item. If the joint is to be dynamic, the strength and stiffness, among other properties, of the joint must be defined.

Each such dynamic joint is modelled within the simulator, I believe, by the mathematical properties of a damped spring. This might involve the linear compression and extension of an undercarriage unit in its joint with the wing. It might involve the twisting of the wing at its joint with the fuselage with the use of the aileron or flap. So the spring action of any joint may be linear like a bouncing standard helical spring. Or it may be torsional like the spring of a windup clockwork motor.

All springs have position where they are at rest without any force being applied. If a force is applied to pull a linear spring or to twist a torsion spring and then released, the spring doesn’t immediately return to its at rest position. It first passes this position. If it was pulled or extended, it becomes compressed on release. The spring oscillates around the at rest position with the oscillations gradually dying away as in this diagram:

See image below of Damped Simple Harmonic Motion (SHM.jpg)

For each joint in the model, Aerofly needs to know the strength of the joint/spring; how much force it can tolerate before breaking. It needs to know the stiffness; the resistance to an applied force or the strength of the springy-ness. It needs to know as well a damping factor or how quickly the oscillations of the spring will decay to zero – stay at the at rest position.

But it’s a bit more complicated than that. A joint can be pulled or pushed in any of the three axes: roll, pitch and yaw. It can be twisted also in any of those three axes. Each stiffness value has to be matched by a damping factor. So a set of three stiffness/damping pairs for pulling and pushing and a set of three stiffness/damping pairs for twisting. That’s a total of 12 values. Worse still, they appear not to be independent in practice. They seem to interact between themselves.

With a rigid structured model, it is only the shape of the model that is simulated. With a dynamically structured model, the structure itself is also simulated. If I create a model for the simulator that I intend to build in real, I want to know if the wings will come off, pulling out from a dive, if the main spar is 3/8 square balsa. I’ll have to study strength of materials and do some experiments, of course, to obtain those 12 values. But at least it is possible.

In modelling a real, full scale, aircraft, apart from putting restrictions on the forces generated by the control surfaces before breakage, the amount of twisting of a wing with the use of the aileron governs the sensitivity and control response subject to the inertia of the whole aircraft. I assume that using near-correct values for these joint values will make for flight characteristics in the simulator that correspond better to the real full-scale aircraft.

Most user-created Aerofly models are similar in the size and weight of their components to at least one or two of the models supplied by Aerofly. In this case, the values of strength, stiffness of any joints, together with the weight/mass of the component, can be copied from the Aerofly examples and adjusted with small changes.

Unfortunately for me, my Lancaster will be, I hope, full-scale within the simulator. That is it will have a wingspan of 31 metres when the norm is 0.5 to 2 metres. There is a biplane model of 12 metres wingspan, an Antonov An-2. But, in general, I cannot copy values and change them a little. I have to do the calculations to arrive in the right ball-park before I start trial and error methods.

To do this I need to know or estimate with some accuracy the weights/masses and centres of gravity/mass of the individual components listed above.


Inertia and Moments of Inertia
The inertia of an object is another way of saying its mass. It gives its resistance to a change of motion in a straight line. We’re talking about Newton’s first and second laws of motion. Moments of Inertia are related to circular motion. Here we are talking about the torque force required to accelerate an object to go against its resistance in a circle. Imagine undoing the wheel nuts of a car wheel. Torque is the force you apply to the spanner with your bare hands.

Imagine an aircraft parked on the tarmac. The wings are supported at their joints with the fuselage. The force of gravity acts all along the wing and acts to twist the wing joint downwards. But to do any calculations about the force on the joint we need to find a single point in the wing where the whole mass of the wing acts together with its distance from the joint; and so find some numbers we can use. The distance of the single point from the joint about which the wing tries to rotate downwards under gravity is called the radius of gyration.

There is tons of stuff about moments of inertia and radii of gyration. But it’s all formulas. It is difficult to find anything which says what those two are. I think it’s a maths abstraction which actually works for some unknown reason. The best explanation I have found talks about the radius of gyration being the optimum distance from the axle or axis of rotation to apply torque force to achieve a required degree of rotation.

Actually the maths is really quite good. The easiest shapes to find the moment of inertia are solid rectangles and cylinders. You can approximate a complex shape like an engine nacelle with a number of different sized solid rectangles and cylinders. Each of these smaller components will have its moment of inertia and radius of gyration about its own easily known centre of mass. It’s easy to re-calculate the individual two values about the centre of mass of the nacelle itself. Once you have that, it is easy to combine all of them together to give the moment of inertia and radius of gyration of the whole nacelle about its own centre of gravity. You can continue combining and re-calculating until you have these two things for the whole aircraft

Once all the moments of inertia and radii of gyration have been combined for, say, a tailplane, Aerofly 5 requires that a solid rectangle be defined which has the same characteristics as those of the tailplane.

You can get some idea from this drawing. But in this case the cylinders corresponding to propellers and wheels need to be converted into equivalent solid rectangles:

See Moment of Inertia Equivalent Solid Rectangles Image below (MoE.jpg)

First, for the whole aircraft, this all tells you how the model should react to aerodynamic forces: how much force is needed and how quickly the force acts.

Second, for my pushing, pulling and twisting of individual joints I need to know how much force on the joint there’ll be for a given airspeed and g force when a change of motion is commanded by the control surfaces.

Again, to do this I need to know or estimate with some accuracy the weights/masses and centres of gravity/mass of the individual components listed above and shown in this diagram.