K2Engineering Bending display component

Replies to This Discussion

Permalink Reply by Cecilie Brandt-Olsen on September 15, 2017 at 1:25am

The rod goal, which calculates the bending moment, is based on a model that uses two consecutive line segments and describes the moment at the shared vertex. That's why you don't get one value per line segment but rather one value per shared vertex (end hence also why you don't get a moment at the end points).

Wrt the display, it is only for visual purposes - don't use the length of the lines to compare to your moment values. The lengths are mapped values from the bending moments (but there are some tweaks to deal with certain situations). You can see how it's done here: 

https://github.com/CecilieBrandt/K2Engineering/blob/master/K2Engine...

If you want to compare your results with the ones from K2E then compare to the M output values.

Permalink Reply by Jon on October 6, 2017 at 6:16am

as Alex said, yes pls open a K2Eng group for more discussion...,-)

By the way, thanks for example file. It really help to understand.

I was in ur AAG 2016 workshop and very impressed of K2Eng.

Hope can see more example...e.g. BucklingAnalysis and Wind Load(var. scale...maybe based on Perlin Noise...)

Permalink Reply by ng5 Alex on September 15, 2017 at 4:12am

Hello Cecilie,

thank you for the answer and for the plugin.

I am sorry, my question wasn't clear in the first place. I am calculating bending stress using the formula σ=(E*t)/(2*r). the results i am getting are very similar to the rod bending stress output (two to three decimals minor differences). The r is coming out of an evaluate curvature component, which evaluates an interpolated curve of the polyline points. These results are used as length for lines, that are perpendicular to the curve.

Doh, i did not think to take a look at the code posted in github. this will answer some following questions i had, about how the calculations are done etc.

I have read in this page, that you did some benchmarking of k2engineering. Are the results published? I would to cite your work in my thesis.

thank you again for your response.

best

alex

ps.

please consider creating a group for k2eng plugin.

Permalink Reply by ng5 Alex on September 15, 2017 at 5:07pm

if you get a chance to answer the previous question, i was wondering also if it is possible to input more than one E value in the rod and bar components. it seems now the components work on single value, so that it is not possible to distribute along one rod, different E's to selected line segments.

thank you

best alex

Permalink Reply by Cecilie Brandt-Olsen on September 18, 2017 at 1:45am

Hi again Alex,

Yes, you can find some benchmark studies in the appendix of this document:

http://www.ceciliebrandt.com/cases/calibrated-modelling-of-form-act...

You can definitely have different E,A,I values along a polyline. Just make sure that you input as many line segments as E,A,I values then it should work (i.e. parallel lists of geometry and properties). I hope that makes sense.

Ps. you're right, I should create a group for K2Eng.

Permalink Reply by ng5 Alex on September 18, 2017 at 4:01am

Ok!,

at first i was partitioning my lines list into 4 groups and then provided for E,I,As such as for example

lines

0;0  5 segments

0;1  10 segments

0;2   8 segments

0;3  12 segments

E

0;0 8000

0;1 3500

0;2 5000

0;3 1500

etc

this did not work, but then figured to provide 1 by 1 lists like you suggested, by duplicating accordingly E,I,As.  It would be nice though the components to work on list or tree access, unless this causes problems.

An exception is the bar self weight component, which although provided an 1 by 1 lists input it outputs the following, working in cross reference i think,

my workaround was to graft inputs, flatten the output, in order to match the items one by one.

than you again

best

alex

Permalink Reply by Rudi on October 6, 2017 at 2:10pm

I tried the samples included in your download of K2Engineering. But it finds unrecognized objects: Materials and Pretension. Any hint?