Please note that posting questions in word format makes it exponentially less likely to get a response because reading them takes additional effort (over plain text files),
If you want help, you need to make it as convenient and easy to assess your issues as possible.
Two comments on your simulation.
- There is a minimization but no equilibration in your simulation. Rather you start with the deformation immediately. That is not likely to produce reliable results
- Pressure/stress fluctuates. Sometimes a lot. And more for smaller systems and also for less compressible condensed systems. Thus you need to average. If your system is in equilibrium, you can use fix ave/time, but when you are deforming that would only be possible for limited intervals due to the ongoing modification of the system. So you need to average either by creating bigger versions of your system (e.g. with replicate 2 2 2) until its fluctuations are within the desired limits or by running multiple independent trajectories (e.g. by running your equilibration longer than technically needed and storing individual data files are at sufficiently distant intervals like 100ps apart so they are decorrelated and then start multiple runs from those data files and average over those results) and take the average of those results or a combination of the two.
For more information on either subject I suggest discussing with a local expert on MD simulations.