Structural design

Time only allows a series of small snapshots to indicate what we did. Perhaps more detail etc later……
The prototype will be "experimental" category, but is designed using JAR22 certification criteria.

Early in the structural design Steve Dwyer developed a nice spar design spreadsheet. The inverse design ability and deflection calcs
made it easier to compare materials of various properties while tailoring the spars for all the lifting surfaces. At first we thought that the
high strength properties of some of the available carbon materials would mean stiffness issues would dominate. But due to our limited
resource for exploring and testing new materials/process we decided on wet layup with locally available fibre, so spars were in the end
sized on strength issues. Generalised material properties like those used by German Akafleigs were finally used, and the small amount
of testing we did do suggests this was wise. This approach was used for the whole structure and the result may be that the SG-1
prototype is above its target empty weight of 140kg (for Ultralight Glider category, formerly Light Glider category).
These graphs were taken from work on the final wing spar done near the end of the process.

The lifting surfaces are carbon/PVC sandwich with carbon uni caps. The fuselage is a glass shell with carbon uni 'longerons".
Because the pilot interrupts the normal line of the front tube carrythrough we have a spar bridge box idea and some part-bulkheads.
The vertical loads go through these bulkheads and the rear carrythrough connection is pinned for the tensile forces. The configuration
Allows some good possible relationships between the load carrying elements this way, but it does add complexity to the engineering..

Fea models of the flying surfaces were made early in the work and gave some ideas about performance and weight of the composite shels and spars. A manual breakdown of hardware enabled correct mass distribution in FEA vibration models for a flutter analysis using SAF code (a flutter analysis method extracted from a NASA optimization code). Introducing this fairly early in the structural design process did allow it potentially to contribute to the structural concepts and laminate sizing, and in fact the critical flutter speed did size some of the wing shell laminates. FEA was also used to look at wing section deformation due to surface pressures with wing bending, the shell laminates near the root being resized so we had an acceptable result in Xfoil.

Later most of the laminates for the structure were resized by Peter Scholz using classical methods. This prompted a final upgrade and
completion of the structural load cases. L/dy distributions with deflected controls, roll or airbrakes were done from a vortex lattice model
(VLM) but most symetric distributions were finally calculated in the spreadsheets at limit load from the local lift curve slopes. Meanwhile,
throughout the project, the detail design, internal structure, controls etc have gone through several iterations.

The analysis of flutter and divergence with the SAF program required a lot of work just to try and validate that it was working. With commonly just a field of numbers for an input file and no post processor, we had to improvise our own interface and data visualisation. Roland Broers spent most of his 3 months here helping on the FEA and flutter models. Some interesting work on the SAF code issues was done by visitor Rob MacDonald from Texas.

The two basic parts of the SAF flutter model are illustrated below. A VLM for the aero forces and a layout for input of modal displacement data. From a separate FEA vibration model one gathers modal displacements for a select number of nodes. These modal data points are defined in the SAF geometry. The drawing shows two strings of points we selected for the wing and two for the aileron, all with one degree of freedom. From these modal data points SAF interpolates the modal displacements to suit the vortex lattice model geometry. The SG-1 VLM geometry shown at the top is fairly typical.

So far the SAF models are set up for the wing and the tail block (rear of fuse fin and tailplane) separately. The "rigid body modes" were included in the vibration model but not the SAF model. We will probably try a vibration and flutter model of the complete configuration. Some of the mode shapes that were input in the SAF flutter models are viewable or downloadable below.

The three thumbnails on the right are to download some animations (AVI). The first is an ALGOR wing vibration model showing symetric 1st twisting mode. The next is an AVI from Matlab showing animated "pressures" from SAF flutter code. Roland did this to try and verify that the VLM model in SAF was correctly specified The next is an ALGOR vibration model of the tail group, first antisymetric fin bending mode. Both of the FEA models are designed to include the "rigid body modes"

Damping vs Velocity and Frequency vs Velocity data from the SAF flutter analysis of the wing symetric modes is shown below (results were processed in a spreadsheet). You can see the 1st twisting mode crossing zero damping at the critical flutter speed. Once the FEA vibration models and the SAF flutter models are set up it is fairly straightforward to reiterate for a changed laminate in the design. Also some features in SAF such as varying the stiffness for particular modes make it possible to develop some ideas quickly. When the finished glider is checked with a ground vibration test we can adjust our FEA models and rerun SAF.

To better understand the fuselage under static loads the internals were modeled in more detail and completeness with FEA. These model geometries are still a bit simplified or coarse, partly from time limitations and partly because they were started earlier with limited computer power and limited meshing methods. The thumbnails below will show some details.

Some of the fuselage load cases were applied to this model. Below are pics of the main wheel box and main fuselage bulkheads with spar bridge box. The pic of the wheel box is from the JAR22 side load case, showing shear strain and deflection (defl. scaled x50). It does not show panel thickness. There are plywood sandwich areas at the main retract gear pivot points, which were sized by the FEA deflection and some manual calcs for the bearing stress. The adjacent pic of the bulkheads shows the same load case and data. The last bulkhead pic is from a loadcase with 7.8Gs positive load on the fuselage. The sparbridge has carbon 45/0/45/0.....bearing blocks at the main bolt areas and some carbon uni to stiffen the load path above and below (again, thickness doesn't show in any of the mesh geometry). Generally strains tend to be low.
The deformations under load suggested some small geometry changes to the major bulkheads. No formal buckling analysis done yet.