A question of philosophy

Some answers can be found there : http://eba.eu.com/7 Technical Matte...rts/ISO-DIS-12215-9 keel validation study.pdf

 

FEA will not help much for grounding load case of small keelboat. FE models are solved for static or quasi-static cases. While grounding of keelboat is highly dynamic process with high accelerations, angular and linear, structure own weight playing part and so on.
So FEA or not, for grounding load case huge assumptions will have to be taken, and huge safety factors with them.

 

How about FEA with anisotropic multilayer structures? Not that simple as well...

 

It is a high end technology to do it. SP claim quite a success in this area
"http://www.gurit.com/news.asp?itemid=1872&itemTitle=Developments+on+Advanced+Composite+Analysis+and+Design+Optimisation&section=000100010071000300130002&sectionTitle=Marine+Case+Studies&year=&month="

 

Looks more like experiments, not something daily use... We are developing some fireproof structure with them now.

 

Thank you again.

The EBA document is very interresting : it show the empiric aspect of the problem of grounding ("The current F4 formula ... is still highly non scientific").

I note also that R. Loscombe give a "time to full stop" of 0.5 sec., vs 0.25 sec. in the book of Larsson & Eliasson ; by I suppose this factor is very uncertain in practice ...

 

Time to stop is highly artificial simplification, in order to enable simple procedure to "extract" forces for strength calculations to be used. It is taken deliberately too short, to make accelerations bigger and design forces higher.
As nobody really know, what impact scenario and what corresponding structure deformation and hence stress scenario will really happen, it is prohibitively expensive to do detailed evaluations for myriad of possible impact-deformation-stress combinations; not to mention that the very methods and tools to do it are yet far from being perfect and widely used.

 

Experiments of this kind provide guidelines, insight in processes, etc. . All this add to experience without the need to load real thing to destruction.

 

With further reading of SP publications it looks FEA is quite common tool they use.

 

The goal of FEA is not to analyse the stress of a real grouding case, but just to analyse the stress generated by the idealized grouding forces assumed by the rule, and the show the stress in the structure is within the limits of the rule.

 

I believe crash tests can give more info on stress/damages than FEA. But for boats they are too expensive; I know only few such research where full-size boats were crash tested.

 

I understand it perfectly well, thank you.
Here we have a contradiction of the kind "take eyeball measurements of what you see, guess the rest, and calculate the results to 6th decimal; believe the result to 4th decimal".
Main rule of any evaluation, calculation, analysis... is: The result cannot be or assumed to be more exact as the initial data or analysis procedure.
And for this very reason I would not consider FEA of grounding load case as something final, unquestionable. Or skimp on safety factors for this particular load case only because the structure was Finite Element Analyzed.
While in the same time, I am a great advocate for using FEA wherever possible, BUT with important condition: there has to be good reasons to believe that initial information we feed in to analysis is accurate to ~10%. Only then, it is worthwhile and possible to run detailed analysis, apply safety factors for loads, and allow permissible stress = 95% of certified yield stress of steel.

 

You perfectly right on this.


But there is a side problem. I cannot say I just made an eyeball compliance with the rule, because the rule itself is just a wild guess. The first court issue will kill me on this.

The goal of FEA for me is not to analyse the physical problem, but just to analyse the compliance to the rule.

BTW, iso 12215-9 appendages (keels and grounding loads) explicitely states :"This standard recognizes that some structural components within the scope of this are often best analysed using 3-D numerical procedures (e.g. box keels, girder and floor grillage frames).
...
The term '3D numerical procedures' is intended to indicate any structural assessment method which is not limited to simple geometries. In most cases the term will correspond to finite element analysis"

 

Compliance to the rule is made in three steps:
1) Rule Loads
2) Rule Procedure and Formulae to calculate resulting stress
3) Rule Stress Limits to be not exceeded


FEA is actually a tool to asses deformations/internal stresses in some structure in much more detailed and less assumption-loaded way as classic (and tried, tested, well understood and simple to apply) formulae, commonly used today.
So it is actually as close to analyzing physics as it gets for now. In fact, precise analysis of internal structure physics is the reason, why it is used at all. While, with FEA, compliance to the rule is provided with two points of three:

1) Rule Loads
2) FEA for stress analysis instead of Rule Procedure and Formulae
3) Rule Stress Limits (sometimes quite a bit higher if FEA is used)