How can I understand "Strength" parameter?

unhandled

(Daniel Piker) #21

Thanks @osuire,

I agree that the documentation is still severely lacking. Better explanations of the underlying principles(and particularly the ways it differs from other engines people might be used to)could probably help avoid many of the misconceptions seen in this thread. I know it’s been a long time coming but I am working on this.

About the 2 versions, I thought I’d pretty much solved this now, with the exception of the remeshing and some reciprocal tools. These are taking a bit longer because I’m trying to do a closer integration of the meshing. If there are still other significant features 2 is missing from 1, I’d like to know.

Another issue is that there are still a lot of tutorials and examples out there made in 1, and beginners often find these and try and build on them. Hopefully better docs will also help with this. Also for anyone that has an example, definition, or tutorial in 1 that they’re having trouble converting for 2, I’d like to help.

The breaking change between versions was a hard choice, but really couldn’t be avoided because of the way 1 was written. At least I think I can now avoid any similar breaking changes in future versions.


(IVELIN PEYCHEV) #22

LOL

I gotta admit, feels like “me against the world”. Hopefully, at the end of this argument, Kangaroo will deserve it’s description and I’ll understand how to use the “strength” parameter to get realistic results.


#23

Hi Ivelyn,
“Strength” is not an SI unit, so that’s a warning right there.


(IVELIN PEYCHEV) #25

thanks @osuire I’m well aware of that.

Please, don’t fight Daniel’s battles. If you don’t have relevant statements, just don’t post.
This threat has anyways turned into an arena against my opinion of Kangaroo.


#26

Ivelin,

I’m not a Kangaroo fanboy at all, but I never wait hundreds of hours of struggle before I start ranting.
I follow the motto : “rant early, rant often” :slight_smile:


(Michael Pryor) #27

and task was solved in another tool

which one?


(IVELIN PEYCHEV) #28

Bongo 2.0

I was able to adjust its behavior faking the collisions. No physics just an animation, but it was good enough for the customer. The problem of the real case is we weren’t sure if the angle in the anchor pocket is small enough so that the anchor crown rotates to it’s position when hoisted. That’s why I started with Kangaroo, but anchor parts got dislocated and no collision was detected. When I started playing with “strength” parameter got even worse.


(Daniel Lee) #29

Hi Guys,

Quite relevant discussion here. I am not a Kangaroo expert, and I have been mainly dealing with engineering FEM tools (Abaqus, Ansys, Robot Structural Analysis for example). I recently came a cross Kangroo, and tried to develop it for initial stage design analysis for origami structures.
In the beginning, I must say that I was also confused by the terms such as ‘strength’ of the Hinge component and ‘stiffness’ in the SpringFromMesh component, because I immediately assumed that the terms were used in the same way as in solid mechanics or other commercial FEA tools. Later, I found out that Kangaroo uses dynamic relaxation method, which is as mentioned before, the main purpose of development is form-finding. Thus the terms were mainly used for the designers to provide arbitrary (and relative) values for creating geometrical forms, which can be verified again in FEA tools with material definitions and other constraints before the designs can be actually constructed.

Even if the values are arbitrary, it can be said that kangaroo was developed based on selected (not yet complete) laws of physics (it may be still abstract, but again it has been accepted as sufficient for many designers to serve as ‘form-finding’ tool). It would be true if one says that Kangaroo is not structural analysis tool, but it was never described so.

One point I would like to make here is that Kangaroo has still good potential to be developed for structural analysis tool if those initially confusing terms are properly addressed. For example, I recently examined the nodal displacements given by Kangaroo with a commercial FEM tool, and when the spring ‘stiffness’ value was defined based on EA/L (as Piker also mentioned initially), where the bar areas were calculated according to Filipov (2017) (also with Hrennikoff (1943), though it was for specific possion’s ratio only), the nodal displacements of both quadrangle and triangle elements under tension load were comparable to that of FE model. For lateral load, the discrepancy was larger, but I believe this can be also improved using further discretised bar elements with appropriate additional definition of bending stiffness between the bars; which I am currently working on.

Like Ivelin (and also me), many engineers may feel confused with the terms when they first use Kangaroo, however, it doesn’t take long to realise that the tool was not meant for structural analysis (it took me less than a day at least); yet in my humble opinion, its description as a physics engine will still be widely accepted also by engineers as long as it is clearly stated that it is developed for form-finding.

Sorry for long text. Just sharing my thoughts. Good evening to all. =)


#30

I agree


#31

I’d love to see this in the length parameter description. Unless @DanielPiker you feel it’s giving away too much of your intellectual property, this little bit of text can make a huge difference.