ahhh, makes sense, my bad :"D
Well, assuming that the potential form is correct for what you are trying to do then, what I can see from your script that can be a bit weird is the minimization you are doing after having assigned velocities to the particles in your system (e.g. of this problem: Energy minimisation (CG) at various temperatures). Maybe you can also check the temperature evolution during the nve portion of your system (PS: I dont see why you would need to do some dynamics in the nve ensemble nor an energy minimization for glass transition, but indeed I am not familiar with the topic, so as Axel said this is probably seen better with your supervisor or the team you work with). I think for glass transition you start at a high temperature (at which the liquid phase is the stable one) and then you drop it very quickly or do a ramp where the temperature changes very fast, no? So maybe you can equilibrate your system in the liquid phase first (again, to be seen with your supervisor).
You can also save data files (see the write_data command page) and dump file before the last part of the simulation to try visualize what’s happening and carry out a deeper investigation yourself.
If you dont have anyone around to assist you with the project, you can try asking questions on the Science Talk category about “how people working with glass transition simulate it” or try to see if someone working with this happens to make input scripts available on github or something for you to try to learn from it.