Background: So I have a transient structural analysis that uses a large number of load steps (>1000) to simulate a system with load feedback (an aerodynamic model which takes velocities, angles, etc. and applies a load). The use of load steps over substeps is so an APDL command snippet can be exectued at each load step to apply the feedback load. (I'm not using USRCAL as I don't know how).
The is problem is, to define the load step time data, I have to externally create a vector, e.g. [1e-4 2e-4 .... 0.1], and insert that into the step control tabular data (Analysis Settings). I have to re-edit this, and the substepping options, whenever the timestep, etc. is changed. This is a pain, and I cannot parameterize any these variables (timestep, etc).
A clunky workaround I have is to leave an dummy loadstep time vector in analysis settings, e.g. [1 2 .... 1000], and then use:
*GET, nloadstep, ACTIVE,,SET,LSTP
settime = dt*(nloadstep+1) ! (+1 is needed)
executed at all load steps. This overwrites the dummy time value with a new value, based on a timestep, etc. that can be parameterised with snippet arguments (dt = ARG1). However I'm inherently skeptical of this, as the dummy time value persists for a short time in each loadstep, before the snippet is executed, and I can't be sure it's not affecting anything. Also, the solver progress bar doesn't display properly, and solution info folder is always marked as unsolved (both, I guess, b/c solver is tracking progress against a dummy end time: 1000 s). And I still can't parameterize the number of loadsteps!
Question! Does anyone have any ideas for creating a system in which I can parameterize the number of loadsteps, and loadstep time increment / end time? Is there an APDL command to specify complete loadstep time data at the start of the analysis? (Workbench is doing it somehow!) Or a routine to replicate the transient analysis solution process in a snippet, so I can control the timestep vector that is used? Or would USRCAL solve all my problems?
Any help is welcome!