# Testing boat structure with FEA

Discussion in 'Boat Design' started by EddieGreen, Feb 13, 2021.

Tags:
1. Joined: Feb 2021
Posts: 16
Likes: 2, Points: 3
Location: Nice

### EddieGreenJunior Member

Yes... and many thanks for the reminder !

Could you please elaborate then, for instance can you explain what would be the restraints applied in the example shown above by Andrei Rochian (wooden boat structure video)

Also let's take another very simple example, e.g. a hull with only 10 frames and 10 stringers at equidistance, i am OK with loads (say only the rigging forces and hydrostatic pressure to begin with, static analysis), can you give me some ideas
on what type of restraints and where are they applied ? Not talking about a local analysis with restraints for instance on the edges, would like to see the whole boat with its simplied structure here.

Can you explain why ? (a summary is OK !)

I understand there is some simplification (i like it as a starting point though !) and the risk of GiGo by using a software without knowing precisely what we are doing/need QA, however by not having to deal with meshing anymore, where mistakes can easily happened (as SnS for FEA, now ORCA for Rhino is also going along that route for CFD), it looks like an interesting starting point to quickly visualize some key areas.

2. Joined: Feb 2021
Posts: 16
Likes: 2, Points: 3
Location: Nice

### EddieGreenJunior Member

Yes, quite agreeable !
Let's take a half cylinder "hull" then with only some frames and longitudinal stringers inside to reinforce the shell, any hints regarding restraints (type/location) ?
Thanks

3. Joined: Oct 2008
Posts: 6,808
Likes: 764, Points: 113, Legacy Rep: 2488
Location: Japan

Well, technically the generic term is Boundary conditions.

You need to understand the mathematics of how the programme works, and by that I mean the matrix theory. And this is where it starts to be complex. Inside the programme the element matrices [K] are assembled from the model you have created using the element types you have used. However when the matrix [K] is singular, this is a problem, since its inverse does not exist – see matrix theory.

But what does this mean? The physical real-world meaning of this in the case of solid mechanics problems is that the structure (your model), is free to undergo unlimited rigid body motion (move about) unless some support constraints are imposed to keep the body or structure in equilibrium under the loads.

Hence, some boundary or support conditions must be applied in order for the matrix to be solved. There are two types of boundary conditions: displacement and force boundary conditions; sometimes-referred to as mechanical and geometric boundary conditions.

For most FEMs the displacement is used to describe the support or constraints on the model, and hence the prescribed displacement values are often zero. If the structure has insufficient constraint to prevent rigid body motion, then an infinite number of solutions to the model (equations) exist; each having the same deformation, but a different location in space.

The simplest type of boundary condition is when the dependent variable in the solution is known to be zero at various points in the region (and hence nodes in the finite element mesh). In simple terms, the model is “held” in some way. When this occurs, the equation components associated with these degrees of freedom are not required in the solution and information is given to the assembly routine which prevents these components from ever being assembled into the final system. Thus, only the non-zero freedom values are solved, not the endless infinite variations.

So, all you really need is 1, 2 or 3 nodes, where you have a boundary condition that prevents movement in the x-y-z axis. Ideally you locate this node(s) far away from the area in which you are investigating. Since a concentration of boundary conditions creates a singularity. In simple terms, stress = force/area…and the area in this case if it is a single node, approaches zero, therefore the stress approaches infinity! So you decide where is best to place the boundary condition to “hold” the model, to prevent it from ‘flying’ away, and the location at which you hold the model, is not going to influence the region you are interested in, and can dismiss any high stress values, or, add additional boundary conditions to smooth out the singularity.

No No No… you must be in total control of the meshing and the elements types.
For example, thin shell elements each node has 6 degrees of freedom, and provide membrane/in-plane loading. Whereas Solid elements each node only has 3 degree of freedom, and no membrane/in-plane loading. Thus depending on your analysis, this can yield different results. Since this is related to their shape function, which is the key to understanding the element being used and how the mesh and hence the accuracies/errors are understood.

tlouth7 likes this.
4. Joined: Feb 2021
Posts: 16
Likes: 2, Points: 3
Location: Nice

### EddieGreenJunior Member

Yes ! this is what i am trying to better understand. SnS doesn't seem to be using points or "nodes" though. So far i have mostly applied restraints to surfaces when using SnS for exercising and training simulations.
A quote from the SnS developers :
@TANSL : any additionnal info (or simple examples) you may want to add here to help me grasp the issue? Bearing in mind we are talking about small (30 to 60 ft. grand max.) and rather slow pleasure sailboats here i.e. not an Imoca foiler where slamming is very significant

Thanks all,
Ed

5. Joined: Oct 2008
Posts: 6,808
Likes: 764, Points: 113, Legacy Rep: 2488
Location: Japan

If there are no nodes to speak of, the process of boundary conditions is being watered down and misunderstood. And in the process it is dumbing down FEA so an idiot can use it.

"......A quote from the SnS developers :
.... In general, such restraints can lead to poor stress results, especially in the neighborhood of the restraints. So this is a somewhat controversial issue......."

And boundary conditions for most cases, ARE real. A typical simply supported beam…if the supports are not real, WTF is holding the beam??

There is nothing controversial about it, it is how FEA works, and understanding structural theory and matrices to yield results that can be controlled and provide a high degree of confidence in the results.

It seems to me that SnS is attempting to make FEA available for anyone, where that is like saying we’ve automated scalpel precision in heart surgery with a programmed robot, so anyone can perform open-heart surgery!!

I would suggest reading several books on FEA and then you'll begin to understand how SnS can easily lead you to incorrect results.

6. Joined: Sep 2011
Posts: 6,448
Likes: 370, Points: 93, Legacy Rep: 300
Location: Spain

### TANSLSenior Member

I don't know what examples you want me to show you. Boundary conditions are set by the designer. I don't have any examples of finite element analysis to show, sorry.
By the way, even if a boat is small, it may have problems that make it advisable to study its longitudinal strength. If the boat, even if it is small, has foils, the stresses on the foils and on the hull must be studied. Whatever the size of the ship, the loads to which it is subjected can be of the same type although of different magnitude. It only happens that in some cases, such as small boats, it is not worth taking them into account, but the calculation procedure is the same. When a certain calculation has to be carried out, its execution can be just as complicated in a small boat as in a large boat.

7. Joined: Feb 2021
Posts: 16
Likes: 2, Points: 3
Location: Nice

### EddieGreenJunior Member

Nothing wrong with SnS here, take e.g. a beam casted into a concrete wall, restraints are easilly applied to the 4 intersection surfaces between the wall and beam, and from what i have seen outputs are consistent with material strength basic formulae...
the 4 surfaces where these physical restraints are applied look even more real to me or at least easier to apprehend than some fixed nodes ...

Where i am struggling in when these boundary conditions are less obvious as in the case of the overall (even simplied) boat structure and hull

Diverting from my initial question :
I can guess where you are coming from, but wouldn't dare saying "meshfree" technology is a piece of crap, actually there are several attempts to use it in other areas too, including CFD e.g. : Orca3D Marine CFD https://orca3d.com/pages/orca-marine-cfd
another area certainly of interest to naval architects. As long as you know the limitations, there is nothing wrong by trying to use a simplified & more efficient tool.

This means users and developers do not need to create or maintain a finite element mesh as is required by other analysis technologies. For dummies like myself this eliminates some modeling errors caused by poor meshes, missing elements or misaligned edges ... I like the idea of modeling a project only once, and not having to prepare additional geometric models only for simulation and analysis purposes (e.g. FEA or CFD analysis)

8. Joined: Sep 2001
Posts: 627
Likes: 120, Points: 53, Legacy Rep: 110
Location: Bordeaux France

### an2reirfifty boat designs

9. Joined: Sep 2001
Posts: 627
Likes: 120, Points: 53, Legacy Rep: 110
Location: Bordeaux France

### an2reirfifty boat designs

Hello Eddie Green

The animations and the image like the one below are made only to illustrate the software's capability. In real simulation the b reaking of the structure is measured function of the material's critical failure stress wich in the case of fiberglass composite is given by Eliasson at a value of 125 MPa

Last edited: Feb 14, 2021
10. Joined: Aug 2002
Posts: 15,120
Likes: 898, Points: 123, Legacy Rep: 2031
Location: Milwaukee, WI

### gonzoSenior Member

Without restraints, there will be no stresses; the structure will simply move. However, for FEA to be useful, you first need to determine what forces will be applied to the structure. For example, you can't dismiss slamming loads on a hull simply because they are difficult to calculate. FEA is particularly sensitive to garbage in/garbage out. For example, it will use the Young's modulus to calculate deformation. However, unless you specify a limit, it may allow the material to infinitely elongate. The results would then be useless. Limiting the value of the inputs is the only way to make the result reflect the behavior in real life. FEA will not do the thinking for you. What I mean, is that you first need to study what a boat does and why before approaching FEA analysis.

11. Joined: Aug 2002
Posts: 15,120
Likes: 898, Points: 123, Legacy Rep: 2031
Location: Milwaukee, WI

### gonzoSenior Member

12. Joined: Jan 2007
Posts: 107
Likes: 12, Points: 18, Legacy Rep: 36
Location: europe

### fastwaveSenior Member

Hi guys,
This is a classic FEA question when it comes to complete boat models. The support and equilibrium of a boat is taken care by the buoyancy forces.
Restraining any nodes like suggested here will introduce wrong load paths and stress points. There are many tricks of the trade, eg. setting up spring elements on all the underwater nodes.
And many others. There are scientific papers discussing the matter.
It’s very different to analyzing a panel. And constraining the sheer in complete boat models is also wrong for pretty much all cases.

EddieGreen and an2reir like this.
13. Joined: Aug 2002
Posts: 15,120
Likes: 898, Points: 123, Legacy Rep: 2031
Location: Milwaukee, WI

### gonzoSenior Member

I agree, that understanding the structure and the environment it operates in is the only way to get useful results from FEA. Using springs is an interesting idea. Do the K values change depending on the location in the hull?

14. Joined: Sep 2011
Posts: 6,448
Likes: 370, Points: 93, Legacy Rep: 300
Location: Spain

### TANSLSenior Member

How few finite element analyzes have you done!. Perhaps now you begin to understand what the OP is asking.
To be clear, I have not done any either, but I know the fundamentals of the method and its tools, which, unfortunately, does not allow me to answer the OP.

15. Joined: Oct 2008
Posts: 6,808
Likes: 764, Points: 113, Legacy Rep: 2488
Location: Japan

A 'surface' in this case is merely an attempt to spread the boundary condition over a greater area to reduce high stress concentrations.
This is why you need to understand what the elements and boundary conditions do to the model. Once understood, it is a straight fwd process to ensure these do not affect the model you are analysing.

Hi Andrei,
Why would you apply boundary conditions in this location, and what load scenario is being applied in this procedure?

Im not aware of any FEA recommendations/guidance in in this ISO doc?

That is an over generalisation, as it depends upon the load scenario you wish to perform..
It is not necessary for all...

Again, that is an over generalisation, as it depends upon the load scenario you wish to perform..and whether one understands how to apply boundary conditions that do not influence the overall model.
If one understand the structural analysis that is required and has the ability to control the model and mesh, then it is a straight forward application of applying boundary conditions that are sufficiently far away from the area under investigation.

Indeed, but again, it depends upon the type of analysis you wish to do, and how you wish to do it too.
And it it not necessary just for loads below the waterline. They can be applied anywhere and generally internally if performing a general detailed frame analysis, to simulation longitudinal continuity of structure.

So why do you keep commenting on things you are not familiar with then?.. especially when as you appeared to suggest you would not:

Or are you more interested in another pissing contest with other forum members as is your MO rather than being able to answer the OPs questions?

Forum posts represent the experience, opinion, and view of individual users. Boat Design Net does not necessarily endorse nor share the view of each individual post.
When making potentially dangerous or financial decisions, always employ and consult appropriate professionals. Your circumstances or experience may be different.