BESO for beams - Underlying Algorithm

Hello,

I am currently investigating BESO in design processes.

I am very interested in the BESO for beams component, but I can't seem to figure out how the underlying algorithm work exactly as all BESO litterature I have found refer to shell models.

Is there any principle or literature you can point me towards for to understand it better?

How does BESO for beams differ from BESO for shells on an algorithmic level?

Replies to This Discussion

Permalink Reply by Karamba3D on May 11, 2019 at 12:43pm

Hello Simon,

in the Karamba3D manual there is a section about the BESO for beams component.

Some additional information can be found in the attached paper 16_PREISINGER%20-%20Evolutionary%20Structural%20Optimization%20Revi... which was presented at the 7th seminar of the Structural Morphology Group in 2011.

Best,

Clemens

Permalink Reply by Simon Laporte on May 12, 2019 at 1:30pm

Hi Clemens,

Thanks for your response, it was very helpful.

As a follow up question:
I've been analyzing a 2D truss with BESO-beams but whenever i have less than 300 elements my model seem to work stochastically rather than deterministic which I would expect.
(IE. every time i give the algorithm the same input it gives a different solution, for example when you click the Target Ratio slider).
If needed you can counteract the effect by adding additional boundary conditions.

Is this the expected behaviour with a low amount of elements, or have i set the script up wrong?

I've attached a script setup to work stochastically and a distribution of the effectiveness of the solutions to the algorithm if it's feed the same input 200 times.

Attachments:

BESO-Stochastic.gh, 60 KB
Effectiveness distrubution.JPG, 27 KB

Permalink Reply by Karamba3D on May 13, 2019 at 1:03am

Hi Simon,

the algorithm should be deterministic. If not it is a bug.

I will take a look at it - it may however take some time since I am away from office for a few days.

Best,

Clemens

Permalink Reply by Simon Laporte on June 13, 2019 at 3:33am

Hello again,

Did you ever look into this issue?

Permalink Reply by Karamba3D on June 13, 2019 at 7:59am

Hello Simon,

this is still on my TODO list.
Do you have a small definition which I could use for debugging?

Which version of Karamba3D do you use?

Best,

Clemens

Permalink Reply by Simon Laporte on June 14, 2019 at 5:18am

The model attached in my previous message was the smallest definition I could do showing the bug clearly. The module functions normally under other conditions.

I'm using Build 0.9.0076 Student version.

Permalink Reply by Karamba3D on June 28, 2019 at 4:26am

Hi Simon,

There has been a bug in Karamba3D 1.3.1: The BESO weights for truss elements were calculated incorrectly.

The problem is solved in the current build of Karamba3D 1.3.2.

Best,

Clemens

RSS