Henk_v wrote: ↑03 Jan 2024, 22:17
Nonsense.
Any good designer knows what its design can handle before they submit it.
Crash tests are validation tests, not design iterations. That's done in-house.
If a design fails the test, engineers first look how it is possible their tools did not predict it. If the tools are off, you are sailing blind.
Generally, yes.
But as a design engineer i'd like to argue that its much less clear cut than that.
Confidence in the design depends to a large extent on the margin you're allowed to keep against a given load. Often its fully acceptable and sometimes even desirable to add a small but effective amount of mass, which also from a cost perspective might not even be that significant. It can be like a drop in the ocean, from a project perspective. With mass being so critical in F1 and the margin required to be as little as possible, I can see why the engineers designing the impact structures might be less than fully confident.
In addiction there is the complexity of simulating and approximating these kind of impact structures. It's not like a fully homogeneous steel rod in a tensile test machine generating a stress-strain curve. One has to appreciate the difficulty in simulating not only fracture mechanics, but with composite material with anisotropic properties.
Also from a material perspective, the less margin the higher the requirements on predictable defined and consistent material and structural properties. Properties such as friction and adhesion within the fabric and adhesive might vary slightly throughout the entire structure, or from one item to the next.
This leads of course to manufacturing. On-the-limit design will require extreme meticulousness and care in all the phases of production. And while the final product might be scanned, machined and polished repeatedly I doubt that all surfaces will be as perfect as said tensile test specimen with almost perfect diameter and run-out. The crash structure might for example have inner cavities that are hard to reach.
Finally one has to consider repeatability. How many front impact structures can the organization realistically test in-house to aid development and to come to a final specification?
And then, once that final specification is locked in, what is the probability of the end product failing of succeeding the validation test, given the variances and tolerances of the design? Welded structures, just as a reference, can be engineered entirely to welding standards such as IIW or Eurocode, using state of the art methodology. Yet those standards recognize variance and tolerance, and as such still state a probability of one out of many structures failing. Not per million but per hundred.
As far as I'm aware the FIA requires only one item to pass the validation test for homologation. Then, further produced items have to be according to the homologated specification - but aren't tested (obviously, since its a destructive test). So depending on the margin in the entire chain from design to manufacturing, there is still room for an "unlucky" scenario where lets say 1 bad structure was tested while 9 good one's weren't.
So with all this in mind F1 engineers should be either humble about the fact that there is a chance that their design fails the validation test, or risk being arrogant and in denial, or know that they left some margin which there might be good reason for.