questions about sp command

[TWI]

I have some questions on the analysis module of opensees, regarding single point non-homogeneous constraint.
1. When performing material test of a uniaxial nonlinear material, use 'sp' command to impose a specified displacement path (time series object) to one DOF, then choose 'loadcontrol' as 'integrator', and penalty constraint. So for this case, could the equilibrium path actually pass the limit point? I run some examples and found it actually does converge without any problems at the limit point....
I thought the loadcontrol can not pass the limit point.... or maybe the loadcontrol in this case is indeed displacementcontrol since 'sp' command specifies displacement as the 'load pattern', plus the model only have one DOF, and the material itself does not need iteration.

2. What if the above problem has multi DOFs? Could loadcontrol still work?

3. As the nonlinear equation is , then if non-homogeneous constraint is specified, the load pattern P* would be dependent on stiffness too (using transformation or penalty constraint), which is not constant during iterations. Then when opensees is applying Newton's method to this equation, is there any special treatment for this case or still use the same iteration procedure, but just update the load pattern P* in each iteration?

Hope someone could explain these doubts to me, thanks a lot.

[fmk]

1. you are not using LoadControl in the typical LoadControl sense, i.e. with an actual load applied. All LoadControl does is increment the time in the Domain by delta lambda. In you case there is no load, only a displacement path. The displaceent at the node controlled by the SP will be updated by some increment.

If you had a load applied it would not get past your limit point. However you have an SP in there, it is behaving more like displacement control so yes LoadControl can go past the limit point. If you don't understand this hust plot the stress-strain historry of the material.

2. load control will still work. again it is not acting as LoadControl.

3. the equation is not valid for SP's, it is only for loads.

[TWI]

1. you are not using LoadControl in the typical LoadControl sense, i.e. with an actual load applied. All LoadControl does is increment the time in the Domain by delta lambda. In you case there is no load, only a displacement path. The displaceent at the node controlled by the SP will be updated by some increment.

If you had a load applied it would not get past your limit point. However you have an SP in there, it is behaving more like displacement control so yes LoadControl can go past the limit point. If you don't understand this hust plot the stress-strain historry of the material.

2. load control will still work. again it is not acting as LoadControl.

3. the equation is not valid for SP's, it is only for loads.

Thanks Frank. But for the question 3, if using transformation handler, how did Opensees form the nonlinear equation then?

[fmk]

if using the transformation constraint handler with SP's .. the handler goes and sets the disp at the nodes directly.

[TWI]

if using the transformation constraint handler with SP's .. the handler goes and sets the disp at the nodes directly.

Yeah, below is my guess, please let me know if it is correct. The single-point displacement would be condensed to the right hand side of the equation to be treated as a load. So my real question is how to solve the nonlinear equation which has unknowns on both sides? just use the common Newton's method to iterate based on the unbalance force or there is some other special treatment for this case?