Loading imperfections, b.c. imperfections. Then you have the stiffness imperfections which can refer to the fiber angle deviations, material properties imperfections, fiber volume fraction deviations...
Thanks @Regina Khakimova for your answer , now my question is how can I introduce these all types of imperfection in my model while performing a non linear analysis ? , Is it possible to introduce them directly by modifying the keywords or I will have to write a subroutine for that ?
If I do this research, I use Microsoft Excel to introduce the imperfections. First, make a bulk data without imperfection. And I define imperfections as functions of the coordinates. Then process the bulk data in Excel. In the process, the imperfections are given to the bulk data---modifying the grid position, fibre orientation or others. Finally, save the data as CSV format files. The files are the bulk data with imperfections.
It is interesting to give the imperfections according to a specific buckling mode, because it maybe the "worst case".
Yes, you should write some simple routines in python for that. As Yoshiki mentioned, you will have to work either with txt. or with csv or similar data from where you read the position of the deformed nodes to rebuild the Imperfect model.