cannot find "option" // cp949 codec can't encode

[gookki]

from last post: viewtopic.php?f=33&t=2052&p=4116#p4116
I fixed my model with your help

than I try to analysis, but it can't write tcl file with error.
I think it is an error caused by element property. but I dont know how to fix it...
I just remove the element property and get another error.
What kind of element type to solve this problem?
here is my scd file.

[marafini.f]

Dear Gokki,

both .scd file you sent were not merged.
You have all disconnected geometries.

Did you watch this?
https://www.youtube.com/watch?v=w8FCrWkeSyI

[gookki]

I'm sorry. I ordered merge and made mesh again, but I didn't press save and uploaded it in that state.

After proceeding with the merge, there is an error like that.

Just in case, I removed the wall and used the hinged beam element in the 0902 model, but the same option/codec problem comes out.

[diego.talledo]

Dear Gokki,
I suggest to take the following modifications to your model:
1) I see that you created several surfaces for surface load. Can you create only one surface for each span of your frame? The surface load distributed the load to its edges (unless you exclude them explicitly). If you create multiple surfaces discretizing them, simply it cannot work.
2) As a consequence I see that you divided your beam elements. The longitudinal reinforcement changes along the sections. You can use the RCBeamColumnProperty specifying different sections along the element. You just need to deselect the option "Use same sections on all integration points" and then specify the section for each integration point. Please keep in mind that the module RCBeamColumnProperty is under development: it works correctly right now, but the automatic computations that you see (for instance for shear) are not available yet. If you need more than 5 integration points (although in my experience if you add more integration points you need to pay more attention on energy regularization for mesh-objectivity of results), you can use a custom integration (that basically is what RCBeamColumnProperty does behind the scenes) and specify for each integration point the section and the weight.
3) You used some Hinged Beams in the physical properties (this is a special purpose compound that wraps an integration for the internal beam with two zero lengths at the endings of the beam: it's important to note that you selected a fiber section for the internal beam, but you should select an integration scheme - e.g. Lobatto 5 points); the problem is those elements does not have a Hinged Beam element property. You find it in Element properties -> Special Purpose -> HingedBeam. Note that you also need to define a zero length element property. This actually is the source of your error.

Attached the file where I corrected the problem 3). Now it is solved and the tcl is written. The only error that is still there is the one related to the surface loads with several elements (Problems 1-2). Please make these modifications and then we can start from there to check other possible problems.

analysis0901_test_edits.zip

(774.21 KiB) Downloaded 212 times

[gookki]

thanks for your help.
I will try to fix my model

but aside from modeling, still I have codec problem with surface load condition. so I can't make analysis.tcl completely.
The problem is that the basic encoding value of the Korean Windows system is cp949 (ANSI/OEM - Korean). from a little search in web, I think I need to declare in utf-8 format when opening the file to be written.

This functional problem has also been identified in recent webinar files. But I don't know how to solve it. For now, other functions other than surfaceload.py seem to work well.
So, I am thinking of calculating the load that acts on the beam and girder separately while modifying the model and putting it in line load.

[marafini.f]

Dear Gokki,

sorry both me and diego focused on the errors we were also encountering on our machines.

I think the problem lies with korean characters yes.
On which machine are you running STKO? Cause the solution could be different.
The idea is to set Python encoding to UTF-8.

I found this post maybe it can be of use.
https://stackoverflow.com/questions/509 ... ing-by-def

[diego.talledo]

I think the error is due to the non latin characters in your system.
Maybe it cannot decode "/" character?

When I run write_tcl in your file I obtain the error on surface load that you can see in this picture:
https://ibb.co/T476Qkp

The error is due to the fact that you created multiple surfaces.
If you observe the strange characters in your image appears where I have the slash character "/".

I think that the decode error happens to you in the raising exception phase: basically when there is an error, the system catches it and provides the window you see. In my opinion in that phase you cannot decode those characters. If there is not an error you don't see that decode thing. The prove is that now you don't see it from beams.

So basically you need to do what I told to solve the problem (have only one surface for each frame span): just as a test I reduced your model in the way it should work; you can try if it creates the tcl without the error.

analysis0901_test_edits_red.zip

(283.73 KiB) Downloaded 230 times

Of course you can always compute manually the linear load and assign it to the beams so removing all the surface loads.

Besides that follow Francesca's directions to solve the encoding problem.

Let me know if this works.

[gookki]

Thanks for your help, I can solve the encoding problem. now I will try fix the surface with your advice.

I looked through the Python code and all the other lines in the same format went back to normal, so I erased the 942nd line and wrote it down again just in case. That's exactly same typing, but the problem is solved. According to the finding code difference on https://www.diffchecker.com/diff, the characters in the 'face' is different, but I don't know why. It's fixed anyway. I will upload screenshots and modified py files to see if it will help.

[gookki]

I revised the model overall after hearing your advice. Using "RCBeamColumnProperty", the cross section of the member is distinguished, and when element is applied as forcebeamcolumn, an option error occurs again. I don't know what I'm thinking wrong.

here is the another model file

[marafini.f]

Dear gokki,
there are some fiber section physical properties that you assigned directly to the edges without a beam section property and some edges with no assignments.
Please check all physical and element properties, you turn on and off the geometry with some specific assignments to help you identify mistakes.

Enjoy your modelling
Francesca