Apply shear stress by 'fix aveforce'

Thank you Eric again.
Yes, in my system there exists a force which resists the system to follow the applied force. Therefore, the system could evolve to a steady state after initial relaxation.
Your idea is nice. It is easier to assign a constant velocity instead of a constant force. But the result is indirect. I just want to reproduce the experiment.
Controlling stress and strain both provide meaningful result. Now I am already changing to control strain, although it is not perfect condition I want to achieve.

Many thanks.