BIG PROBLEM IN CODE CHECK OF BARS

[StefanoPasquini6790]

Hi everyone,

I absolutely need to understand if Robot had a bar-limit in verification of steel members.

Well, I try to explane my problem. I have a very big model with 86367 nodes, 84913 bars and 1778 panels

Well, when the calculation is done I run the verification of the bars to define the reliability intervention on the whole structure. This morning I've discovered this surprising bug....

When I run verification for whole bars I have a complete table with the list of the non verificated ones...I've ordered this table by the verification ratio and, for the same bar that have the maximum ratio I've run the verification again.

My big big Surprise is that the value of the internal forces are very different, and in case of one bar verification this bar is ok otherwise in case of multiple verification this bar is absolutely wrong.

Well, my question is: IS THERE A LIMIT FOR THE NUMBER OF THE BARS IN VERIFICATION LIST? I need a very urgent answer to this question to define how I have to go on with this project and to chek the real quality of the work just done!!

Greetings

[t.sautierr]

Didn't you change your bar type between the two checkings? for example normal bars for the frist one and tension/compression for the second? would explain that your bending moment has disappeared.

[StefanoPasquini6790]

Hi, the verification was done in the same calculation session. The results in bars was the same for the two verification session, but the report show different results

[Artur.Kosakowski]

Can you send the model (let me know if you don't want to have the public access to it) please?

[StefanoPasquini6790]

Hi Artur, obviously this model is extremely big, 5gb the model rtd file and 4 gb the _RT file. The calculation time is one day for the analysis and more than one day for the verification. I'd like to don't share this file with the Community for the privacy of my Customer, tell me how can I share it only with you. You can write to my personal mail.

[Artur.Kosakowski]

I'll contact you offline.

[StefanoPasquini6790]

Sent.

[StefanoPasquini6790]

Hi, I need to update this thread. Now I made a verification group for all the section type in the model and the result it was the same. In the guop verification table there is a non verificated bar. Well, when you investigate the single bar non verificated in the group check the result is absolutely different. I can't understand this strange behaviour, i need to know how I have to go on with my project.

Greatinga at all and happy Easter.

[tony.ridley]

Can't you just perform verification in two or three runs? 

Anyway, one workaround.........once you have run calculation, you can save part of the structure as a "sub-structure".  

The analysis results remain vaild.  

Then you can code check small number of bars in the sub model.  Any change to member(s) in the sub model can be re-loaded into the main model via "update from sub-structure"

Hope this helps........

Tony

[StefanoPasquini6790]

Hi Tony, many thans for your suggestion. At this time i had very big improovement opening my file in repair mode. The model size go from 5.5 gb to 550 mb. I think that a lot of calculation made to check the reliability solutions have created some problem on the link with the result file. When I close this check stage, I let you know the results.

Greetings and thanks for your suggestion.

[Artur.Kosakowski · Accepted solution]

Hi Stefano,

I have run the verification for all bars and the results I obtained for the bar you obtained is the same as for the verification you made for the bar with the largest ratio. Looking at the bending moment diagram it sems that the verification  has been done correctly and I can't see any issue here. The difference I noticed is that I have SP5 installed whwreas you have Robot 2014 SP4. The other factor is the size of the model which results in the need for large RAM being available. What is the size of RAM on your computer?

After sorting the results by ratio I compared the values obtained for the verification done for all bars and a single one with the larggest ratio and I can't see any issue here either.

Hi, I need to update this thread. Now I made a verification group for all the section type in the model and the result it was the same. In the guop verification table there is a non verificated bar. Well, when you investigate the single bar non verificated in the group check the result is absolutely different. I can't understand this strange behaviour, i need to know how I have to go on with my project.

I'm not sure if I understand your description good enough to be able to investigate this situation. Could you give me more detailed one (step by step to follow) please?

[StefanoPasquini6790]

Hi Artur,

My workstation have 22 Gb of RAM and I don't think that this can be the problem.

Well, I've opened the file in repair mode and the dimension of it has gone from 5.5 Gb to 550 Mb, then I've recalculated the structure and run once more the check of the bars. In this case everithing goes ok, I've done the same procedure you've described in the previous post.

Well, at this point I've understand thet when the file became "too big" it's very good thing to open it in repair mode ang go on with the recevered model. The real question is "When I need to do this??". As you can see the time to completely analyse this model is too large and this procedure make me spend a lot of time...

Greetings and thanks

[Artur.Kosakowski]

Saying that the file is 5.5GB do you mean that you saved the results in the file itself or is this size the sum of the 'model' file (rtd) and the results file (rt_) ?

[StefanoPasquini6790]

Hi Artur,

Only the model (*.rtd) was 5.5 Gb, and the result (_RT) was 3.8 Gb. After recovered the model this less then 1 Gb, and the result file was the same:

Greetongs

[Artur.Kosakowski]

My assumption is that at certain stage of working with the model the results were saved inside the rtd file (this is why it is such large). Do you remember if you saved the model (which was previously calculated and saved with the results) under the same name with save with no results option?

[StefanoPasquini6790]

No Artur, I don't remember, but this is an operation that I've never do! Everytime I save the file without results I always change the name of it !!!

Greetings

[Artur.Kosakowski]

I such case try to upload the file with the results (the  one which you got different results of verifications). I'll check if I can simulate the same situation and if this allows us to determine its cause.

[StefanoPasquini6790]

Hi Artur,

I'm uploading the model file (and the results ones) that give me the described problems in the ftp. 

As I told you, I've solved this problem using the "open in repair mode", but this it work only noe time....in the second analysis runned in a repaired model I have the same problem...then I've opened in repair mode this file and, with my big surprise, all the code check rule are disappeared although each bars have the code chek rule assigned!!!!!

I'm sure that this depend by the dimension of the model...but I can't understand this behaviour.

greetings and wait for my model on ftp

[Artur.Kosakowski]

I'll try to download and check the file next week. In the meantime could you try the following:

1. Move  to another location on the drive

2. Open 

3. Save it under a new name as rtd without results - what is the size of the file?

4. Do this again - what is the size of the file?

5. Calculate this file (model) and save results in the external file using again a new name. What are the sizes of saved rtd and RT_ files?

6. Run code checking and see if you have the same issue. 

in the second analysis runned in a repaired model I have the same problem.

I'm not sure how to understand this. Does it mean that you had run code checking for the first time and the results were OK and then you repeted this check again and the results were wrong?

then I've opened in repair mode this file and, with my big surprise, all the code check rule are disappeared although each bars have the code chek rule assigned!!!!!

Have you set the Job Preferenmces from the opened file as default before using the repair mode?

http://forums.autodesk.com/t5/Robot-Structural-Analysis/Error-message-quot-Encountered-an-improper-a...

[StefanoPasquini6790]

Hi Artur, I'sorry but I can't upload the file in the ftp, there is a lot of interruption in dsl connection and I can't understand why.

However, I've read the post you've linked and I will try to do that steps.

Thanks a lot for your commitment.