Simplified model of lattice towers 1

[tubsdua]

Hello all,

I am modeling a transmission line on SAP2000. The aim is to compare the results of numerical modeling with the data collected from a test line. The sensors are installed on two spans. The model has 3 spans consisting of spans between two end (tension) towers.
As each tower has more than 1000 members the model has a large degrees of freedom. When I try to do a non linear time history analysis for wind loads on the lines the analysis time is exceptionally large. I have tried using multi-threaded solver and also reduced the tolerances. I modeled only the lines without the tower and found that the time is still within acceptable time-frame (12 hrs). So now I wish to reduce the towers to a simplified model with same stiffness and modal characteristics.
The question is: By what method I can reduce the lattice tower to a simplified model. Say, beams representing the tower and the traverse. I am aware the a lumped mass approach can be used to get the mass matrix and the stiffness matrix can be reduced to lesser degree of freedom. However each tower has more than 1200 members and 35-50 section types. So it will be very time consuming to make and then reduce the 1200x1200 matrix. Is there any other method?

 

[Ron]

I have modeled similar in other programs (mTab-Stress and STAAD). Size of the model is not a big deal. It is possible you have duplicate nodes or lack of node closure somewhere...troubleshooting large models is time consuming and sometimes difficult.

A model of this size should run in a matter of minutes, not hours.

Animate your model and see if you can detect anomalies.

 

[tubsdua]

Thanks for your reply Ron. I tried that. The nonlinear dead load analysis and modal analysis are done in matter of minutes. But when I try to do the non-linear time history analysis the solution time is too large. Just to check if there is something wrong in modeling the towers I modeled only the conductors with insulators and tried to run the direct time history analysis and still the solution time is very large. The applied time history is of 10 mins. I am attaching the model of conductors. It is a simple model with insulators and cable elements. Do you see anything wrong in the model.

 

[Ron]

tubsdua....the file you attached is apparently proprietary to your software. I don't use SAP2000, so I can't open it nor can I troubleshoot it. Perhaps some of the members who use SAP2000 are willing to do that.

Why is your time period so long? If you are only looking at wind gusts, why not consider maybe 3 or 4 times the gust period just to see the spectal response? That would put you into a 10 to 15 second analysis frame, rather than 10 minutes...am I missing something?

 

[tubsdua]

Thanks Ron,

Yes we can do as you suggested. However we need to do a deterministic analysis and compare the results with data collected on a test line.
I noticed one peculiar thing about SAP2000. When I run the file after making some changes to the model the time for analysis is very large. However some times same model would run and converge faster. I presume when I import the towers into one file there are some anomalies which are preventing the solution to converge.
Any views on this?