Hi every one

I build a model containing two pyrene molecules .The structure went through geometry optimization and dynamics run at 1600k with DFTB+ tools.Then I apply a velocity to one of them by script and start dynamics run again,so they can collide when the dynamics is complete.However,the non-bond energy is 0 when the dynamics is complete.And I do not think the non-bond energy should be 0 due to  the dispersion correction which I have used in script.And my question is

(1)why the non-bond energy is 0 when dispertion correction is used?

Here is my energy plot ,the green line is non-bond energy.