Discussion Groups

Robot Structural Analysis

Reply
Active Contributor
RoboExploiter
Posts: 44
Registered: ‎09-04-2011
Accepted Solution

Structure Design - Steel Design - Compound Members

3436 Views, 15 Replies
10-27-2011 12:32 AM

Hi,

 

I've created a model composed of steel waler beams, struts and runner beams.  As the anticipated forces are large, the structure demands struts in pairs, walers in pairs and runner beams in pairs.  

 

After I've run the model and forces have been generated, i try to use the steel design module to check the elements, and i get this error message:

 

"Bars 1to31By .... have not been calculated because they are composed of sections not available in the current code."

 

The sections i've selected are available in the standard UKST database that comes built with robot.

 

Now the clue is that Robot has analysed the splays which are single sections (same size sections as the struts).

 

Therefore i think it is a question of Robot not being able to handle compound members.  Is that correct?  Can anyone share similar experiences?  

 

If my understanding is correct, does anyone know of an addin or a spreadsheet with API that I can use to design compund elements given all the information available in Robot?

 

Please use plain text.
Product Support
Artur.Kosakowski
Posts: 5,105
Registered: ‎12-17-2010

Re: Structure Design - Steel Design - Compound Members

10-27-2011 12:56 AM in reply to: RoboExploiter

Could you send me the model please? Assuming that you set BS as the steel code the compound section built from two H sections should be supported.



Artur Kosakowski
Please use plain text.
Active Contributor
RoboExploiter
Posts: 44
Registered: ‎09-04-2011

Re: Structure Design - Steel Design - Compound Members

10-27-2011 12:59 AM in reply to: Artur.Kosakowski

Hi Artur,

 

As requested.

 

Thanks.

Please use plain text.
Product Support
Artur.Kosakowski
Posts: 5,105
Registered: ‎12-17-2010

Re: Structure Design - Steel Design - Compound Members

10-27-2011 01:33 AM in reply to: RoboExploiter

The double H sections are supported provided they are really acting as a single profile (are welded along its contact edge) therefore you should modify the Waler section in the way shown on the attached picture. For the others you should make the change in the way you created your model and use two separate parallel bars each with a single UB section assigned instead of using the single bar with the compound section as it is now.

 

If you find your post answered press the Accept as Solution button please. This will help other users to find solutions much faster. Thank you.

 

 



Artur Kosakowski
Please use plain text.
Active Contributor
RoboExploiter
Posts: 44
Registered: ‎09-04-2011

Re: Structure Design - Steel Design - Compound Members

10-31-2011 03:19 AM in reply to: Artur.Kosakowski

Hi Artur,

 

I took on board your suggestion.  I've now modelled the waler beams and runner beams as welded together, and modelled the struts individually. 

 

As the struts are not behaving as a single element anymore i've had to introduce lacing between the struts.  Making my model alot more complex.  

 

I'm getting frustrated because the lacing i created did not snap to the centreline of the strut, resulting in microscopic gaps.

 

Why can't the snap tool, snap precisely to the centreline or the node?  :robotmad:

 

When I try to move the node back to the centreline it snaps to the other side of the centreline :'(

 

Is robot intelligent enough to analyse the members as being connected at the centroid, despite these tiny mismatches/gaps?

 

 

Please use plain text.
Product Support
Artur.Kosakowski
Posts: 5,105
Registered: ‎12-17-2010

Re: Structure Design - Steel Design - Compound Members

10-31-2011 04:13 AM in reply to: RoboExploiter

RoboExploiter wrote:

Hi Artur,

 

I took on board your suggestion.  I've now modelled the waler beams and runner beams as welded together, and modelled the struts individually. 

 

As the struts are not behaving as a single element anymore i've had to introduce lacing between the struts.  Making my model alot more complex.  

 

Putting aside the questions how to secure two H beams spaced by 1000 mm working together for horizontally applied load (imposed displacement) or torsion as well as for lateral buckling (unless they are prevented form it) when in the steel design module the load is not 'divided' into two sets and each of the two beams is not going to be analyzed separately - or in other words assuming that this set of beams works as a single compound beam you can 'force' Robot to design it. All you have to do is to create the section using the Section Builder module of Robot rather than the compound tab of the sections dialog. After calculating the properties of such created section you can save it to the user section database selecting its section type (the rules of the verification to be used by the Steel design module of Robot - see http://forums.autodesk.com/t5/Autodesk-Robot-Structural/How-to-create-thin-walled-sections/td-p/3107... ). The problem is what the section type to use (how this compound section is going to work in the model) - should it be a structural tube? Anyway as my intension is to show that such approach is possible as well I'm going to leave this particular question unanswered.

 

I'm getting frustrated because the lacing i created did not snap to the centreline of the strut, resulting in microscopic gaps.

 

Why can't the snap tool, snap precisely to the centreline or the node?  :robotmad:

 

When I try to move the node back to the centreline it snaps to the other side of the centreline :'(

 

Perhaps you have too many snap types switched on at the same time. Perhaps the following approach will help:

1. Switch off display of grid

2. Generate auxiliary nodes along the beams using the Division option without dividing the bar itself

3. Use snap to nodes creating auxiliary bars

 

Otherwise you can correct the model using either the detailed correction option (see http://forums.autodesk.com/t5/Autodesk-Robot-Structural/Warning-after-Structure-Verification/td-p/32... )

 

or after selection of nodes typing the values of coordinates in e,g. Inspector

 

Is robot intelligent enough to analyse the members as being connected at the centroid, despite these tiny mismatches/gaps?

 

Providing the inaccuracy is 'small enough' or within the default tolerance range - yes it is. Otherwise you can increase the tolerance (within the reasonable limit of course) :smileywink: You can check if the calculation elements are created between the auxiliary nodes as explained in the http://forums.autodesk.com/t5/Autodesk-Robot-Structural/Wind-crosses-and-and-DSC-algorithms/td-p/319... .


Hope this helps.



Artur Kosakowski
Please use plain text.
Product Support
Artur.Kosakowski
Posts: 5,105
Registered: ‎12-17-2010

Re: Structure Design - Steel Design - Compound Members

10-31-2011 04:16 AM in reply to: Artur.Kosakowski

The last attachment for the previous answer.

 

If you find your post answered press the Accept as Solution button please. This will help other users to find solutions much faster. Thank you.



Artur Kosakowski
Please use plain text.
Active Contributor
RoboExploiter
Posts: 44
Registered: ‎09-04-2011

Re: Structure Design - Steel Design - Compound Members

11-01-2011 01:05 AM in reply to: Artur.Kosakowski

Thanks for your suggestion on creating new user defined sections.  I might tinker with this at some point in the future.  The only problem that I foresee is that after design of the elements, if the structure is under capacity (or severely over capacity)  it won't be easy to swop in a new user defined beam, as I'd have to create it.

 

I took on board your suggestion to split up the beam using the division option.  It was tedious, but at least now its thorough.  

"Otherwise you can correct the model using either the detailed correction option (see http://forums.autodesk.com/t5/Autodesk-Robot-Structural/Warning-after-Structure-Verification/td-p/32... )"

 

I read this post you sent me but its not relevant to my problem.  The post title is "Warning after Structure Verification".  But my problem is about autocorrection of an inaccurate model.  Are you sure you gave me the right link?

 

"or after selection of nodes typing the values of coordinates in e,g. Inspector"

 

This is not as easy as it looks.  As you'll see in my model one wall in my model is inclined at 26 degrees to the other wall.  I've had to rely on Robots translation and rotation tools to create the model geometry.  Therefore some nodes might have positions defined to less than a mm.

 

"Providing the inaccuracy is 'small enough' or within the default tolerance range - yes it is. Otherwise you can increase the tolerance (within the reasonable limit of course) :smileywink:"

 

Ok, noted.

 

:smileywink: You can check if the calculation elements are created between the auxiliary nodes as explained in thehttp://forums.autodesk.com/t5/Autodesk-Robot-Structural/Wind-crosses-and-and-DSC-algorithms/td-p/319... ."

 

What auxillary nodes are you referring to?

 

Ok, so finally the entire model is looking like it should, and loaded like it should be. I hit the calculations button, and my model is not converging.

 

Firstly, I haven't specified my model to perform a non-linear analysis, so I don't know why Robot has chosen this method over a static calculation.  Secondly  I don't know how to change it back to a static analysis.  I've had this problem before with previous robot models, and the only way to circumvent this error is to start again.  Please, please tell me there is an easier way. 

 

 

 

Please use plain text.
Product Support
Artur.Kosakowski
Posts: 5,105
Registered: ‎12-17-2010

Re: Structure Design - Steel Design - Compound Members

11-03-2011 01:19 AM in reply to: RoboExploiter

RoboExploiter wrote:

Thanks for your suggestion on creating new user defined sections.  I might tinker with this at some point in the future.  The only problem that I foresee is that after design of the elements, if the structure is under capacity (or severely over capacity)  it won't be easy to swop in a new user defined beam, as I'd have to create it.

 

I'd say you are in the exactly the same position when you use the compound sections (the automatic search for the right size of the beam is done for the sections stored in the section databases). What you can do is to create a number of sections that you would consider as possible to be used for a particular model and save all of them in the user section database within the same family. Then you can use them in the design process in the same way as the standard sections.

 

 

I took on board your suggestion to split up the beam using the division option.  It was tedious, but at least now its thorough.  

"Otherwise you can correct the model using either the detailed correction option (see http://forums.autodesk.com/t5/Autodesk-Robot-Structural/Warning-after-Structure-Verification/td-p/32... )"

 

I read this post you sent me but its not relevant to my problem.  The post title is "Warning after Structure Verification".  But my problem is about autocorrection of an inaccurate model.  Are you sure you gave me the right link?

 

Yes, I think so. The meshing.png picture I attached shows the described above functionality in the reference to this line: "You may start with the detail correct option and run it for the entire structure adjusting it to e.g. structural axis." I'm attaching this picture to this post as well. In your case the adjustment to the line may be what you need.

 

"or after selection of nodes typing the values of coordinates in e,g. Inspector"

 

This is not as easy as it looks.  As you'll see in my model one wall in my model is inclined at 26 degrees to the other wall.  I've had to rely on Robots translation and rotation tools to create the model geometry.  Therefore some nodes might have positions defined to less than a mm.

 

The rotation or translation tool should not influence the accuracy of the model. In addition for the functionality I described (use of the Inspector to correct the coordinates) you can set your local work plane by 3 points in such a way that a bar with nodes along it is length that is arbitrary located in the space is 'set' along the one of the main axes of the coordinate system. Then you can easily type the new value of the coordinate you want to change. I'm attaching the example of the use of this option for the graphical definition of the rotation angle (copy of the bar 1 by rotation it by the angle between bar 1 and bar 5).

 

 

:smileywink: You can check if the calculation elements are created between the auxiliary nodes as explained in thehttp://forums.autodesk.com/t5/Autodesk-Robot-Structural/Wind-crosses-and-and-DSC-algorithms/td-p/319... ."

 

What auxillary nodes are you referring to?

 

I meant the nodes you defined along the bar (as if dividing it into parts) in order to connect it with the one next to it to secure these two bars work together (as if it was the compound section).

 

Ok, so finally the entire model is looking like it should, and loaded like it should be. I hit the calculations button, and my model is not converging.

 

Firstly, I haven't specified my model to perform a non-linear analysis, so I don't know why Robot has chosen this method over a static calculation.  Secondly  I don't know how to change it back to a static analysis.  I've had this problem before with previous robot models, and the only way to circumvent this error is to start again.  Please, please tell me there is an easier way. 

 

 Could you attach the model please?

 


 



Artur Kosakowski
Please use plain text.
Product Support
Artur.Kosakowski
Posts: 5,105
Registered: ‎12-17-2010

Re: Structure Design - Steel Design - Compound Members

11-04-2011 02:22 AM in reply to: RoboExploiter

Have you managed to solve all the issues with this model or is there anything I could help with?



Artur Kosakowski
Please use plain text.