Thanks for your response. The reason I thought using NVT the box size must change is that the atoms are compressed in z direction, so they need more space. But they are limited in z direction with wall so they have to move in x and y direction. But anyway, even the NPT in x and y direction did not work.
I have large amount of deformation in my system (strain=0.5 with strain rate 5.5e9 1/s).
I believe I realized the problem but I do not know how to solve it. the system I study is Aluminum and the units is metals. But in fix wall/lj126 I should add the velocity in lj units which is (sigma/tau).
A few things. Your box will not expand in the X,Y direction when you are using fix NVT because there is no driving force for it to do so. Saying you are doing NVT and forcing the wall to move is not an appropriate set up because then you are no longer conserving volume. How large of a deformation do you expect? What type of a material is it? What is the rate dependence? Another thing to consider is physically think about what you are doing and what the rate is. I don’t think your error is all that unexpected.