Hi!
I'm doing some product development / Experiments where i'd like to stiffen up our current designs with semi-rigid eaves instead of the currently pinned version.
The idea is to get the structure to behave more similarily to a 3-pinned frame - although we'd keep the ridge beam.. so i suppose the truth would be closer to something like this:
however - the design i'm currently thinking off is probably closer to that of a knee brace rather than a complete eaves haunch. I suspect the moment diagram would in this case look something like this:
the first draft of a sketch would be something like this:
Where I've inserted a hollow section in the aluminum frame - then bolted on 2 bent plates - these plates are then used to bolt the actual knee/moment resisting eaves connection.
See below for individual parts:
Now to the question - how on earth would i go about modelling this to get a somewhat accurate representation of how the forces act on the various bolts / knee brace / triangular frame..
Do i remove the knee brace (truss member) and just add hinges that are semi-rigid (rotational spring?) instead and calculate an aproximate stiffness of my design? What would be a good approach here to not over / under value the strength of this design?
Thankful for any tips & tricks!
Warm Regards,
Tobias M
UPDATE:
To avoid buckling of the truss / axial rod - i fiddled with the design a bit.
Should result in a buckling length almost half of the prior design.. i suspect this design would also make it act a little more like the "haunch + knee brace" combination i was going for - as opposed to the former design who was pretty much a simple knee brace.
Hi Tobias, 
Due to the slender knee brace and the stability issues, I would recommend not using a equivalent spring stiffness. This would need to be determined first anyway. You also need to make sure that the knee brace does not become decisive for the overall behavior.
In principle, you can simplify the model as a frame structure with multiple bars (one member per component) and couple the bars at the connection points using rigid members.
Alternatively – for more accuracy – you can use a surface model:
- All components must be modeled as surfaces.
- A hybrid approach using both members and surfaces is also possible.
In particular, in your last image, the knee brace with the bracing can no longer be represented as a single member. Whether your optimized knee brace will provide much improvement is questionable, as it can also buckle out of plane, i.e., around the weak axis.
I therefore recommend using a surface model first. If this works well, you can later consider adapting it to a bar model.
The procedure for creating a surface model is shown in this video:
Surface Modeling in RFEM
Hi Gerhard!
First of all - thanks for the response.
You're of course correct in regards to buckling out of plane. Thanks for the heads up. I did not reflect over the fact.
Would the hybrid approach be reasonable to execute by using RSection for the aluminum cross sections - perhaps even adding the inserted hollow section. I know there's some video on this published.. could also do a line release i suppose to add the hollow section within the aluminum member.
And then modelling the bent plate + knee brace (or whatever we can call this 
) as surface members?
The bent plate i believe should be pretty straight forward - because i can create a cross section in Rsection and extrude those as members and convert to surfaces. but for the more complex "knee brace" i'm not quite sure how to go about it?
I'm not very accustomed to the software yet, at least not the more in depth possibilities - but how would one go about modelling a unique shape such as the one in my last image? is that even a possibility?.. i realise that regular members can be converted to surface members by right clicking and choosing the "convert to.." option. But in my case it's not like i can use a cross section or follow a "normal" modelling path? Can i achieve a custom part such as the one i've shown in dlubal?
Lastly, bolts gets represented by rigid members? I guess?
Once again - thanks for the response - I apologize for the additional questions. As i mentioned - i've got a lot to learn still..
Yes, members can easily be converted into surfaces. The knee brace can be created using standard CAD functions. 
Here are two options:
- Import the geometry directly from your CAD program.
- Create the geometry manually or using a DXF file.
If you send me a DXF file of the knee brace, I can make a short video showing how to create the surfaces in RFEM.
I'd love a video if you could provide that!
I made a very basic dxf file from a triangle - hopefully that will be good enough for what you had in mind.
could you perhaps give an example of you "attatching" the surface/part to a beam + rafter in your video? in order to showcase how the part reacts when "bolted" at various locations?
It doesn't have to be a complex beam or so - i regular hollow cross section will do just fine - i'm mostly curious about the way you'd tell the software that "this part is bolted to another part at these locations" correctly.
once more - thank you so much for all of your helpful assisstance. i really appreciate it!!
Sketch Video Dlubal.dxf (79.6 KB)
Hi, I have uploaded a video file here:
Video File
Of course there are many more options to connect the bolts. These are just the basics.
this was amazing - so many things i had no idea that you could do in RFEM.
I'll definately attempt to utilize this more in my modelling practise - thank you so much for your assistance.
And i apologize for the late response..
Keep rocking!
