We are modelling a series of existing buildings adjacent to an new deep excavation to assess the impact a vertical settlement will have on the frames during construction. To do this we are applying vertical settlements (using imposed displacement load) to the columns in the building.
The settlements we are inputting are based on zero displacement at the rear of the building to a defined value at the front, with a linear distribution between. We are not modelling any horizontal displacement at column bases. However, we do not want the building to just "rotate" about the rear column base and react as rigid body. So, by preventing all the column bases from moving laterally we expected to induce forces and moments in the members in order for the frame to deform to resist the vertical movement.
However, when looking at results we are get zero forces in all members from applying such conditions. This is despite the column base being fixed laterally as we intended. To try to compare we have simplified the model down to a simple portal frame (see below). Again, in the model we would expect to see forces in the members, but again we get zeros only. It appears that Robot is elongating the members to form the required shape without causing the frame to deform (which would put forces into the members).
Can anyone shed any light on the results we are getting and offer a method to solve or an alternative method?
Thanks in advance.
Solved! Go to Solution.
Solved by Pawel.Pulak. Go to Solution.
This behaves as a linear rigid body so displacement perpendicular to support line will generate no internal forces.
Some examples below.
(In case of nonlinear calculations you will get forces).
Rafal,
Thanks for the response but I don't quite understand what Robot is doing here. Surely the frame can only behave as a linear rigid body if the right side support is allowed to move inwards (on a roller support) and therefore purely pivot about the left support. If the right side support is restrained laterally surely there should be forces and moments in the members as the frame will need to deform to accomodate the vertical movement?
In the linear domain the displacement perpendicular to the line connecting supports will not result in changing the distance of supports so no internal forces will be generated (the 2nd and 4th structure on the Rafal's screen capture) - even if structure is hyperstatic (no roller supports).
As written by Rafal it is necessary to consider nonlinear effects to obtain internal forces in such case - in more detail it is related to the 3rd order effects activated in Robot by "P-delta analysis" check-box (somehow misleading name discussed in other posts).
Regards,
Pawel,
Thanks for the response. We have now run our models in the non-linear/P-delta case and have been able to verify our results from Robot against hand methods for simple portal frames, and are happy with the results.