A version of this article first appeared in our free newsletter, to subscribe click here
I have been helping out a new client with setting up a composite analysis methodology for a prototype program.
The final client – the OEM – has not provided a methodology and has left it up to the sub-contractors to select their own analysis methods. This is not how we recommend to set up a composite program but this is how the OEM wants it for now.
We prefer the laminate strain analysis approach – there is a derivation of this method here:
However, there are some very simple analysis methods for composites that can be very successful in sizing and predicting the behaviour of aircraft structure.
There are European part 23 aircraft development and certification programs that use simple stress allowables for composite structure. There are different allowable stress values for Glass, Carbon, UD or Fabric. These values come out of the German glider industry and have proven to be reliable – within limits.
Clearly a simple laminate stress approach can work but it has to be within limits. The simpler the stress analysis method the greater the limitation on design features and conventions.
Similar to this a few years ago a north american client had developed an in-house stress based method for sizing their composite structures. As they generally only used quasi isotropic carbon cloth, they came up with a simple stress allowable for their laminate structures using the von mises stress tensor. More information on the von-mises stress envelope here: https://www.abbottaerospace.com/aa-sb-001/3-introduction/3-4-stress-analysis/3-4-2-combined-stresses/
Their derived laminate stress failure value was very close to the values used by the German Glider industry.
I ran a comparison for the von-mises stress envelope for their failure stress for quasi-isotropic laminate to the strain envelope that is typical for hot wet carbon laminate that we would use for analysis and this is what I found:
I would rate that as ‘not bad’ or even ‘not bad at all!”
So for quasi-isotropic laminate you can use a simple laminate stress allowable with a von-mises failure envelope. You have to be careful about biaxial compression but that is a relatively rare situation to encounter.
This company, by trial and error and possibly a bit of luck, had come up with an analysis method and failure criteria that was a very close match to our preferred certification analysis method. This was limited to quasi-isotropic fabric laminate only.
Interestingly if you compare to a traditional ply-by-ply failure method we also get close to a zero margin if we set up the laminate to a net stress of the same value.
You can use our standard spreadsheets to examine these different failure modes and envelopes:
So, our new client has a range of options for potential analysis methods
Finite element models using this isotropic von-mises stress method are simple to create – you can use plate elements and equivalent isotropic materials – interpreting the results is easy as well – you can use the von mises stress tensor and a simple single check for strength. Similarly for hand analysis – you can reduce the overall analysis method to something similar to metallic structure analysis.
Of course….the problems start when you introduce core – although you can extend the method and regard the structure as a cored structure with isotopic sheets, using classical analysis methods for core panels.
And what to do about buckling……we have not run a comparison between a full laminate model in NASTRAN and the equivalent isotropic model for buckling solutions. I should look at that next.