Robot Structural Analysis Forum
Welcome to Autodesk’s Robot Structural Analysis Forums. Share your knowledge, ask questions, and explore popular Robot Structural Analysis topics.
cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 

User Defined Compound Members

16 REPLIES 16
Reply
Message 1 of 17
RoboExploiter
3242 Views, 16 Replies

User Defined Compound Members

Hi,

 

I'm modelling another strut and waler system.  As befor,e the anticipated horizontal earth pressures are likely to be very large, the structure demands struts and walers in pairs.  

 

When I previously modelled my structure (Refer to Post),  I modelled the pairs as single struts and modeled the lacing to ensure the 2 struts were acting as a single pair.  This made the steel design calculation painfully slow because each strut was restrained from buckling at each connection to the lacing.  Learning from past experience I have decided to take on board the advice of Artur, 

 

"...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..."

 

I've tried to do that, but when I use the section in my analysis, the local axes are defauliting, and not allowing buckling analysis to be performed about the minor axis (specifically for member 2, the long transverse strut).  I've tried changing the orientation of the bars axis, and i've tried creating the compund strut with 2 orientations. But the problem is the same:  It refuses to do lateral buckling check.  Am i doing everything correctly?

 

Secondly, in my previous models calculations i was able to see all the terms of the member capacity equations, so i could see which member property was failing the member and incresae the member size accordingly. Now it just displays the equation and the result.  How do i turn that back on?

16 REPLIES 16
Message 2 of 17


 

 

When I previously modelled my structure (Refer to Post),  I modelled the pairs as single struts and modeled the lacing to ensure the 2 struts were acting as a single pair.  This made the steel design calculation painfully slow because each strut was restrained from buckling at each connection to the lacing.

 

To be honest I don't understand why the design was so slow. The only idea that comes into my mind is that you created many nodes along each of the parallel beams and in addition you kept number of calculation points high. As the nodes divide beam into corresponding number of calculation elements (from node to node) and then each of the calculation element is checked in the number of points you entered in the configuration dialog this may be well the reason for slowing down the process. If this is not the case I need to look at the file to say anything more - if this is the solution is to reduce the number of verification point for these beams.

 

 

I've tried to do that, but when I use the section in my analysis, the local axes are defauliting, and not allowing buckling analysis to be performed about the minor axis (specifically for member 2, the long transverse strut).  I've tried changing the orientation of the bars axis, and i've tried creating the compund strut with 2 orientations. But the problem is the same:  It refuses to do lateral buckling check.  Am i doing everything correctly?

 

Before trying to explain I would like to ask how would you like to approach the lateral buckling verification for the doubled H beams? If you want to treat them separately then creating g a compound section will not allow for this (compound section means that the beams are connected together along their entire lengths forming sort of structural tube instead). I would  say that the compound section approach is good when you have something welded/bolted to the flanges which prevents LTB from being factor in the design,

 

Secondly, in my previous models calculations i was able to see all the terms of the member capacity equations, so i could see which member property was failing the member and incresae the member size accordingly. Now it just displays the equation and the result.  How do i turn that back on?

 

What was the section type you selected saving the section to the user database?

 


 



Artur Kosakowski
Message 3 of 17

"The only idea that comes into my mind is that you created many nodes along each of the parallel beams and in addition you kept number of calculation points high. "


This was indeed the case.  I was trying to model the reality, that is struts 1m apart, with lacing every 1m,  X  24m long struts (average length)  X  70 struts (approximately).  I only had 2 verification points.

 

"...I would like to ask how would you like to approach the lateral buckling verification for the doubled H beams? "

 

I would like them to be analysed and designed as a single element to avoid the need to model 2 seperate struts and their lacing. This will change the geometrical properties of the strut and give it greater strength about the individual struts minor axis, something I couldn't capture when I was modelling them as individual struts with lacing.  This will also make manipulation of the geometry easier.

 

"(compound section means that the beams are connected together along their entire lengths forming sort of structural tube instead). I would  say that the compound section approach is good when you have something welded/bolted to the flanges which prevents LTB from being factor in the design,'

 

Agreed.  This is the reality.  I will have lacing connecting the top and bottom flanges of the 2 struts at 1m increments.

 

"What was the section type you selected saving the section to the user database?"

 

I drew the section to the dimensions I required in the section definition and then saved it to the database.  Following that after using my user defined section, I defined the member type as a column, not a beam.

Message 4 of 17

I think that there are two factors here:

 

(I assume that saving the section to the section database you selected H shape instead of leaving the default undefined (?) shape).

 

1. The Section Builder module saving section to the database calculates its properties in the principal coordinate system of the section. Regardless of how you 'draw' it it is saved in such a way that Iz > Iy. It may happen that (due to this assumption) the weak axis is just the one parallel to the webs rather than to the flanges.  The steel design module verifies LTB only against My bending moment whereas in your case it is Mz instead. There is a development request for having the possibility of saving the section to the database as 'rotated' so that the steel design module can handle LTB but this is not available in the current version of Robot.

 

2. Even if the section was saved 'correctly' then mind that for calculations of Mb the section properties of the 'whole' section (doubled H shapes) would be used so for checking of LTB of a single one is not what you would like to see.

 

These the reasons I think the use of two single bars is the best choice. For additional strength you mentioned can be obtained if you model lacing as rigid links between these bars.

 

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
Message 5 of 17

Yes, I save it as a H Section.

 

1. Robot surprised

 

2.  Thats exactly what i want.  I specifically want the 2 struts to be checked for LTB while they are working together as one strut.  Because that is the reaility.  In my previous model the struts were checked individually, but failed LTB because they were unrestrained about the minor axis.  I then restrained each single strut at the point where it connected to the lacing. That is also not correct, because the lacing is connected to the complementary single strut at its other end, and therefore is not truly acting as restraint against LTB.

 

Modelling the lacing as rigid links will restrain each single strut in the analysis, but it still won't get around the above mentioned problem (correct me if i'm wrong). I.e. i can either restrain each strut at every node along its length (which is wrong) or it is completely unrestrained at all points along its length (which is also wrong).

 

My strut detail is shown in the attachment.  

 

Can robot analyse and design this member to BS5950?

or

Should I analyse using robot, and design to BS 5950 by hand?

Message 6 of 17

I'm very sorry but I have difficulty with understanding the way you want to calculate (manually) LTB for the separated double H sections defined as a single profile. Looking at the screen capture of the model you attached and the description you provided I assume that your intended approach is to check LTB as for the single H section using the properties as calculated for the double H sections. If this is right then the possible solution could be to save the standard H section that you want to use for the user defined double H section in the user section database and then using the section database editor replace its geometrical properties (as e.g. area, moments of inertia etc.) with the values you already have for the double H section (the values that you saved in the user section database creating this section). In such such this newly created section will be recognized as a standard H section but verified with the parameters of the double H one. The only shortcoming will be that instead of the double H beam you will see a single one when you display the profiles on the screen but the screen capture for the documentation can be taken from the model you already have Smiley Wink

 

Most likely I'm missing some point and still don't understand how these beams are intended to behave in realty but I would model all the posts and diagonals (if the distance between struts is larger) or just the posts (if the distance is small) and then assign the design parameters as shown on the attached picture. In case there is no load in the horizontal direction (in the plane of the 'paper' on which the scheme of the strut is drawn) I would consider not modeling the bracings keeping the design parameters the same as for the model with the bracings. If you e.g. left the unrestrained LTB length as equal to the length of the entire strut (e.g. for the bottom flange and the check was done at the support where there was a negative moment) then I believe that LTB has not been checked as passed.

 

I hope this helps.



Artur Kosakowski
Message 7 of 17

Andrew,

 

The paths to standard section databases

 

C:\ProgramData\Autodesk\Structural\Common Data\2012\Data\Prof

 

and user section database

 

C:\Users\username\AppData\Roaming\Autodesk\Structural\Common Data\2012

 

for W7.

 

 



Artur Kosakowski
Message 8 of 17

Hi Artur,

 

I've tried the suggested work around that we discussed.  I assigned compund strut properties to a user defined single strut.  I don't think it worked.  When i assigned bar 2 (the longest spanning strut) the user defined compund strut properties, it still placed it with the deeper axis vertical.  Did i do everything correctly?  Is there a specific property in the section properties that robot looks at before inserting into the model?

 

Message 9 of 17

The section you made screen capture of is as far as I can see defined correctly (Z axis along the web). I'm not sure if I understand your description of the way the model is now set up correctly but perhaps all you need to do is to check if you have some rotation (gamma angle) assigned to the bar (not its section) itself. If this is set as 0 send me the model and indicate which bar to look at please.



Artur Kosakowski
Message 10 of 17

Hi Artur,

 

The Gamma Angle was in fact 90 degrees. I defined it as 0 in the "New section dialogue", not sure why it was 90 degrees in the table.  Anyway......

 

Good news, the compound strut is now designed to BS5950 with bending in the appropriate direction!  

 

Bad news is, the calculation note parameters do not match the section database.  See attached.

 

Send me an email, and I'll let you have the model.

Message 11 of 17

The Gamma Angle was in fact 90 degrees. I defined it as 0 in the "New section dialogue", not sure why it was 90 degrees in the table.  Anyway......

 

As your previously created user defined section had been rotated by 90 while you saved it in the user database you had to rotate it by another 90 deg to have it correctly displayed in the model. Then you replaced the section with the one that wasn't rotated while saved in the database but the 'additional ' rotation you defined remained.

 

Please mark the values that you found different and send me your user section database as well. Thank you.



Artur Kosakowski
Message 12 of 17

As requested.

Message 13 of 17

Hi Artur,

 

The properties that are not being displayed correctly are, Ay, Az and I think  B. 

 

The elastic section modulus and radius of gyration do not require definition in the user database, but i see it displaying in the detailed calculation. So i'm not sure if Robot is calculating it correctly.

 

Vpy and Vy are defined as the distance to the extreme fibres. Are those used to define or vet the area of shear capacity as well?  I'm just worried it will overestimate horizontal shear capacity.

Message 14 of 17

Looking at the pictures you attached I think that some values have been incorrectly defined. Assuming that the basic section is UB 914x305x289 you can see that the value of AX for your used defined section is the same as for the single UB section which seems to be wrong. The values of Ay and Az are close to the doubled areas of flanges and web but the steel design module follows the code formulas and calculates them itself. As you defined thickness of the web and flanges as for the single UB section then the values of Ay and Az in the steel design note correspond to the values for the single UB section. To change that the thickness of the web should be 2*19,5 mm and the thickness of the flange should be 2*32mm instead. Perhaps you could try another possibility that I have just thought about. It is possible to use a parametric section with additional information for the steel design module as shown on the attached picture.

 

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
Message 15 of 17

Hi Artur,

 

I have tried your suggestion to use a parametric section with no luck.  As you can see from the attached print screen, Robot infers the D and B of the section based on Vy+Vpy and Vz+Vpz respectively. I have tried switching Vy+Vpy and Vz+Vpz, however Robot will always infer D to be the greater of Vy+Vpy and Vz+Vpz.

 

Robot assumes the major axis of this particular compound strut arrangment to be parrallel to the ground, which is not applicable to this strutting system.  This assumption holds true regardless of how you draw the members, or rotate the members in the member definition dialogue. 

 

Therefore, assuming a user is willing to model (has already modelled)  the struts and lacing, robot can do:

 

  1.  Lateral torsional buckling checks for an individual strut (local check between lacing) assuming you defined the LTB restraints correctly.  Robot is unable to do global lateral buckling checks about the compound struts major axis check regardless of restraint definition, because it cannot calculate the correct radius of gyration, due to the limitations I gave above.
  2. Local buckling check (i.e. between lacing) for each strut about its minor axis (i.e. axis perpendicular to the ground) as long as you define the restraints correctly.  Robot is unable to do global buckling about the compound struts major axis check regardless of restraint definition, because it cannot calculate the correct radius of gyration, due to the limitations i gave above.
  3. Local buckling check for each strut about its major axis (i.e. axis parrallel to the ground).  

I look forward to the version of robot that addresses this limitation.

 

Thanks for all your help.


Message 16 of 17

Would it be possible for you to send me the sample hand verification of LTB (with the reference to the code (I assume BS)  paragraphs) as you would like to have it in Robot please?



Artur Kosakowski
Message 17 of 17
Barthi1
in reply to: Artur.Kosakowski

Colleagues,

What is the conclusion after the last conversation? I am going to propose this method if the global buckling check is done for laced strut by ROBOT.

 

Thanks

Barthi

Can't find what you're looking for? Ask the community or share your knowledge.

Post to forums  

Autodesk Design & Make Report