Comments

• disla

  April 28

  Hi sofien,
  
  Based on this:
  
  ->reactions are the same (I assume you mean in the static solution)
  ->mode shapes and frequency are the same
  
  I can only think that you are using a larger number of modes for the modal superposition which would include more mass into the solution. ¿Did you check in Robot how many modes are considered compared to Mecway and how many mass is movilized on each solver?
  
  I would also check if Robot might be constraining the torsional modes.
  
  Regards
• sofien_73

  April 28 edited April 28

  Thanks Disla, See in attachements ROBOT modal analysis results and options.
  
  Same modes number (50)
  
  as you can see there's only 40.42% of the Z mass fraction that have been taken into consideration on ROBOT model but without the MECWAY dynamic results (see below) we will not be able to conclude.
  
  Actually it's not possible to extract the same results from MECWAY models because of beam-release or contraint equation there several on my model.
  
  I think that ROBOT have an accurate solver which filter torsionnal modes (you can switch it) but I will be suprised that it's the only reason of such a gap on the bending moment.
  
  I don't have any clue.
  
  Best Regards.
  Sofien

  resultat_auvent_ROBOT.xlsx 423.2K
• Mishal

  April 28

  I did an audit 2 years ago on a crushing and screening building done on Robot. There were several issues with the results, it was also impossible to develop an accurate model of the system for a natural frequency analysis.
  
  My general opinion, I would not trust structural softwares (Staadpro, Robot, spacegass etc…). The caliber of people coding those software packages is really poor, almost all the time they implement simplifications which aren’t followed by other software vendors which develop pure FEA software
  
  (Ps I haven’t looked at any of the results as yet)
• disla

  April 29

  According to the manual Mecway can also output the Modal effective mass fraction which is modal effective mass expressed as a faction of the total free mass.
• sofien_73

  April 29

  Thanks disla, according to my previous message it don't works, I don't know how to output modal analysis results in this conditions. If you have a trick, let me know.
  
  Regards.
  Sofien
• sofien_73

  April 29 edited April 29

  Thanks a lot Mishal,
  Commercial code have been developed for more than 30 years by pro devs for some of them... There are widely used by people who had a full trust of the efficiency of their software which it's cost a lot to buy and upgrade every year.
  
  There is few people who goes deeper with the commercial software deeper and I agree with you there is some simplification taken by devs (for example U shape, or torsion loading in accordance with EC3 standards)
  
  Anyway I bet that I am wrong with my MW model and ROBOT provide the correct results.
  I am going to play the same studies with RSTAB v9 to clear my mind.
  
  If you have an idea it should be great.
  
  Best regards.
  Sofien
• disla

  April 29

  I don't know how to output modal analysis results in this conditions

  
  
  @sofien_73
  
  That should be under solution/MechanicalMisc/Modal effective Mass and Modal effective Mass fraction for a frequency analisys. I haven't tested on your problem so I don't know if there could be some limitation due to your actual set up.
  
  Regards
  
  
  
• sofien_73

  April 29

  Thanks disla, I've tried it yesterday anyway I don't works. According to the manual, it doesn't work because of several constraint equations on my model.
  
  Maybe it could be fixed for the next version of MECWAY.
  
  BR
  Sofien
• Victor

  April 29

  I'd delete half the model repeatedly until you're left with something very simple but still different from ROBOT. That should show what's causing it.
  
  Mecway's beams do have torsional rotational inertia, so is that turned on in Robot?
  
• sofien_73

  April 30

  Hello Victor, I'm not sure that I have understand. You mean suppress all these loadcases, I am interested by the loadcase n°18.
  Or are you asking me to suppress all MPC?
  
  For torsionnal masses I am speaking about modal analysis, ROBOT can suppress/avoid torsionnal modes which may have an impact on the global response of the structure.
  But I am not sure that torsionnal mode may increase by more twice the bending moment on the head of the columns.
  
  BR
  Sofien
• Andrea

  April 30 edited April 30

  Have you taken into consideration the Ductility Class (Eurocode 8)? The elastic spectrum is reduced by a factor of at least 1.5. Please check the spectrum parameters in your structural analysis software. Keep in mind that for DCH (Ductility Class High), the q factor can be around 4 (reducing the seismic action to 1/4). In my opinion, you should conduct a round-robin test considering the elastic spectrum (q=1), as it represents the highest seismic demand.
• sofien_73

  April 30

  Hello Andrea, yes I have taken a ductility class factor q = 1.5 for both ROBOT and MECWAY by divide by q the accelerations of the elastic spectra.
  
  I think that doing elastic spectrum test shoud induce considerably worst results.
  
  BR
  Sofien
• Andrea

  April 30

  Could the seismic demand be strictly related to the interaction between the earthquake's frequency content and the structure's natural frequencies? Generally, the response spectrum shows peak excitation for natural frequencies around 2 Hz (the acceleration plateau). For higher frequencies (> 5 Hz) or lower frequencies (below 1 Hz, corresponding to long periods), the seismic demand significantly decreases.
  
• Victor

  April 30

  @sofien_73 I mean delete or suppress features like all the released DOFs or all the point masses or big parts of the mesh until it agrees with ROBOT. That way, you can isolate what's causing the difference.
• Andrea

  May 1 edited May 1

  Staying within the field of seismic action, which type of analysis and which solver are most suitable, in principle, for simulating the response of one or more stacks of pallets?
  Below an example of stacked rack.
  
  
• sofien_73

  May 1

  Dear Victor, as you've asked before I have made a model without any MPC and no local masses and only merged nodes full embeded columns fixation on MECWAY and ROBOT.
  
  Here's the results:
  bending moment on the top of the columns (the same as before):
  MW 23929Nm
  ROBOT 13494Nm
  
  for the modal analysis you can open the Excel table attached.
  There are big difference with the frequency and %masses excited.
  
  I also have watched the modal shape in MECWAY we have a lot of torsion modes which have been ignored by ROBOT modal analysis solver.
  
  Columns warping along it's own axis... same for some beams it's not physical.
  
  I am going to make the same design with RSTAB v9.
  
  Anyway, for steel frame design it have to be fixed. ie we should have the possibility to exit torsion mode for dynamic analysis.
  
  and there is some other things that should be improved:
  
  -Nodal reaction not avaible for dynamic response 3d
  
  -we can't perform axial release (u,v,w) only rotationnal for beam not oriented in accordance with global coordinate system (X,Y,Z) which is not an option for us.
  
  -Modal results not avaible for MPC, I really need MPC for steel frame design.
  
  BR
  Sofien

  resultat_auvent_ROBOT.xlsx 440.2K
• sofien_73

  May 1

  Dear Andréa, according to the picture, I will say beam analysis. shell will cost a lot of ressources and there's no reason to go this way.
  There is some specific connections and gusset that will be step out of the analysis, if it's ok for you go for beam analysis.
  is it anchored ?
  
  BR
  Sofien
• Mishal

  May 1

  I haven’t read through all the comments as yet. But, if you ran a dynamic simulation with initial conditions using the Radioss solver, use those results with some signal processing to check the frequencies… it will account for damping which you modal solver doesn’t do but I highly doubt Radioss will be wrong.
• JJJ

  May 1

  @Andrea you can find standars related to these topics (RMI in USA, EN 15512, EN 15635 in Europe). For seismic parameters, you should follow the national codes.
  Depending on the type of profile, could be necessary an specific type of analysis; I mean, if the structure is designed with open cold formed profiles, you should analyse by the finite strip method.
• Victor

  May 1

  Sorry you're having such a struggle with Mecway. It looks like it still doesn't have the features you need.
  
  However, I'm concerned about any possible bugs so I do hope to somehow identify the cause of this difference.
  
  >without any MPC and no local masses and only merged nodes full embeded columns fixation on MECWAY and ROBOT.
  
  Before removing those features the modes were the same in Mecway as ROBOT but after the change they're different. Is that right? That sounds like it might be a 2nd different issue. Still keep removing features or mesh until they agree. If you get down to just one beam element and they're still different, it should be easy to identify why.
• Andrea

  May 3

  @sofien The frames (load units) are stacked; they are resting on one another rather than being rigidly connected.
  
  
• sofien_73

  May 3

  Dear Victor,
  I agree with you. Let's be honest I have too much ambition for a general FEA software which is effective for the scope it have been design for (and costless).
  
  There's no bugs on MECWAY.
  
  We can't compare specialized steel frame structure design software with optimized methods and solver (ROBOT STRUCTURAL ANALYSIS).
  
  The comparaison have no sense.
  
  Anyway, I've made a small model a cantilever columns with the same spectrum analysis.
  
  I have obtained the same gap with Mw and Qv average increasing by twice.
  
  So I think that it's just the optimized ROBOT solver which exclude non physical mode (especially torsion or small displacements modes).
  
  I don't need to go deeper at this time.
  
  Thanks a lot.
  
  BR
  Sofien