What software can generate this ?

[Ad Hoc]

Would be interesting to see a few comments from Ad Hoc on this, hope he chimes in!

Just noticed this thread

Yup. But I have come across yacht designers that do this - have more than 50% cutout as noted. It is ostensibly a throwback to the days when such yacht structures were never surveyed in detail. The yacht designer didn’t know better or simply subcontracted out the structural side to the yard to sort out. Yards always make it easier for themselves!! This is next to impossible to do these days with Class and now ISO to cover the left overs, as rightly pointed out that Class does not accept cutouts which are greater than 50% of the web depth.

Welding the flange of the long.t Tee directly the web of the transverse is never a good idea, unless there is significant strength reinforcement…i.e the stress in the member is low that one can get away with such, usually owing to much thicker scantlings. But considering more than 50% of the frame has already been taken away..that option has long bolted!

In the joint itself there is a biaxial stress raiser at the welded connection, here is a typical detail, taken from Proboat Magazine article showing this.



As for primary etc structures. Primary supports secondary and secondary supports tertiary. The tertiary in this case, is the plating.

As a general note, there does appear to be a lot wrong with what is shown. But It is hard to say 100% because it may be just that the other structure as yet to be added. But having breaks where they are does not help and it also appears the sequencing of ******** & fabrication is also not ideal either!

 

[TANSL]

I'm not saying that shear stress is non existant. Direct effective shear stress in theat transverse from the panel distributed load is going to be what ? ............ lets say worst case around a 1/5 the bending stress. .

Could be 1/5 or 5/1, who knows?
I think we're not talking about the shear plates but which is produced in the frames. The cut-outs which reduce the cross section of the frame, may result if the shear stress is a problem.
As I am sure you know, buckling occurs when an item works in compression. Therefore say that :

Long before you can get a shear failure of that frame the grillage will fail in buckling. The higher stressed additive ( shear plus bending stress) parts of the structure will be the skin and the flange which is factored into the scantling anyway. Much of the web is effectively structurally redundant.

is incorrect.

If you are interested in this maybe a new thread and I can post some FEA grillage analysis for you when time allows.

I think this thread is appropriate to discuss the cuts that reduce the section of the frames. The method of finite element analysis is just one method of calculation. Adds nothing to the concept of the strength of materials and elasticity. Now if you want to "post some FEA analysis grillage" I´m afraid not to be able to avoid it but, as you sould konow, that's not going to explain anything in this discussion.


Aside: you will agree with me that the current calculation programs allow anyone to make what appear to be very sophisticated and exact calculations, having not clear some concepts.

 

[MikeJohns]

........The method of finite element analysis is just one method of calculation. Adds nothing to the concept of the strength of materials and elasticity..................

FEA is a very good tool for all those things, the next best is destructive model testing. Buckling of grillage panels has been well validated now for FEA. Stress and shear are simple direct calcs within the elastic response.

Aside: you will agree with me that the current calculation programs allow anyone to make what appear to be very sophisticated and exact calculations, having not clear some concepts.

As Ad Hoc just pointed out and as I said earlier the danger is in the longitudinal flange weld to the transverse and it would need to be bracketed unless the structure were over designed. Even the simple linear FEA modellers can show people a lot about stress concentrations in these sorts of scenarios. At least enough to understand where weld fatigue is going to become an issue. But all this is used to be covered by class rules. It's more of a concern now with ISO and "self approval".

 

[TANSL]

MikeJohns, you can do with your money what you like.
I have never despised the finite element analysis. All I'm saying is that, in this discussion, does nothing.
I fully agree with AdHoc. But that binding of a reinforcement to the frame by a plate has nothing to do with the shear forces to which a frame is exposed due to the loads acting on it.

We must be more careful with the concepts, my friend.

 

[MikeJohns]

MikeJohns, you can do with your money what you like.
I have never despised the finite element analysis. All I'm saying is that, in this discussion, does nothing.
I fully agree with AdHoc. But that binding of a reinforcement to the frame by a plate has nothing to do with the shear forces to which a frame is exposed due to the loads acting on it.

We must be more careful with the concepts, my friend.

I'll try again....I'm not doing a good job expressing myself am I.

I think you'll find is that the max depth to 50% takes shear into consideration and that the web is larger in X area than it needs to be to resist shear because it's primary function is producing a decent section modulus for bending and it needs to be thick enough not to buckle in bending.

Also once shear yield is reached the effect of the total grillage is more complex than a shear failure of a simple I beam.

I have noticed designing various cutouts is that the shear is not always as significant as you might think and it will also depend on the location of the cutout relative to the span as well (due to distributed load and end fixity ).
I suspect we are all in agreement anyway. And unless it's a class scantling then yes every case needs checking on it's merits and it's good to be aware of the fundamental principles, you are right.

 

[TANSL]

MikeJohns, I think that indeed both speak of similar things (but we do not say the same). What happens is that all the vessels I have designed must comply with the rules of a classification society and therefore probably suffer from some professional deformation. Deformation that the only thing that gets is to make me more rigorous in my calculations.
Cheers.

 

[Ad Hoc]

I think you'll find is that the max depth to 50% takes shear into consideration and that the web is larger in X area than it needs to be to resist shear because it's primary function is producing a decent section modulus for bending and it needs to be thick enough not to buckle in bending.

Not on smaller boats Mike. On smaller boats it is easy to get the section modulus to satisfy class for bending stress. But, once you establish the stringer modulus, it is highly likely that the cut out depth shall exceed the max 50% depth rule. Thickening up the web, is counter productive for 2 main reasons.

1) You're adding a ashed load of weight
2) Class still wont accept it as it exceeds the 50% rule, no matter how thick.

You also need to be mindful of the shear connection of stringer to the trans. frame web too...this can also dictate the final frame depth.

On larger vessels, agreed, it is, in general, not an issue, for the reasons you cite. But on smaller vessels, it does dominate.

 

[MikeJohns]

For interest I'll clarify my understanding:

The neutral axis has the highest shear stress and with a cuttout you get a lopsided parobolic shear distribution in the web, whether that's an issue depends on the web thickness and how the effective nuetral axis shifts relative to the cuttout. All that relates to the whole grillage response and the resulting strain in the local web.
But the shear decreases quickly from the fixed ends The shear stress quickly reduces towards mid span for a distributed load, so it depends just where the cutouts are too.

 

[TANSL]

For interest I'll clarify my understanding:

The neutral axis has the highest shear stress and with a cuttout you get a lopsided parobolic shear distribution in the web, whether that's an issue depends on the web thickness and how the effective nuetral axis shifts relative to the cuttout. All that relates to the whole grillage response and the resulting strain in the local web.
But the shear decreases quickly from the fixed ends The shear stress quickly reduces towards mid span for a distributed load, so it depends just where the cutouts are too.

I think that's not accurate but I'm not going to argue.
All I want to say is that if not easy to know the shear distribution is not known the bending moments distribution. So I would not dare to say that the shear stress "quickly reduces towards mid span"

 

[MikeJohns]

That's why I was referring earlier to a more detailed grillage analysis determining effective shear from the resulting total strain.
The transverse of course as a classic beam under a distibuted load as we design for worst case has a shear of zero mid span and only half of the max shear at 1/4 span. That's what I mean by quickly decreases.

 

[TANSL]

MikeJohns, I can not agree with what you say. I will present some thoughts and, if you want, think about them and wi'll discuss them:

• there is not a grillage study but a beam (the frame) subjected to bending (no compression).
• The FEA can be applied but, for the problem we are studying, the cutouts, should be applied to a particular frame, not a grillage.
• the shear stress on each section of the beam does not depend on the position of the neutral axis of the cross section.
• Given a certain value of shear stress, if you decrease the cross section must bear it, the beam can deform more than allowed.
• The stress due to the bending itself depends on the position of the neutral axis.
• Buckling beams occurs only in compression.
• We're talking a beam subjected to hydrostatic pressure. At the upper part of the beam (not in its highest point) the pressure may be zero. Therefore shear can quickly diminish to along the length of the beam, and also rapidly increase again.

 

[MikeJohns]

MikeJohns, I can not agree with what you say. I will present some thoughts and, if you want, think about them and wi'll discuss them:

• there is not a grillage study but a beam (the frame) ............

A single frame has no stress transfer no load redistribution, no local web stiffening from members crossing.

A web usually buckles through torsional elasticity and the stiffness of the transverse web here is altered by the attachment of the supported framing. That's one of the few pluses of that method.

You say we cannot know the shear stress distribution, I think it's fair to take a worst case design head as class uses, normal operational stress being well below.

Certainly we are always interested in what occurs after local failure from either simple yield or buckling. And the precursors leading to buckling. But within the elastic response within acceptable fatigue I suggested that we'll see buckling before shear yield even with the cuttouts and that you cannot immediately condemn the method as being deficient in shear strength, certainly without analysis.

 

[TANSL]

I'm sorry to say MikeJohns but, with all due respect, I think everything you say is absolutely incorrect. But that is just my opinion.
For my part, I end the discussion here. Was interesting exchange of views.
Cheers.
Ignacio