Sealed Hydropneumatics Tank Initial Gas Volume

I am modeling a pump trip event in HAMMER. I have a sealed hydroneumatic tank with a total volume of 5,000 gallons and have set the initial gas volume to 2,500 gallons. 

My pump trip occurs after 20 seconds into the simulation. As a result, I expect my gas volume in the tank to be 2,500 gallons within first 20 seconds when nothing has happened yet. However, the gas volume shows an upward trend before pump trip.

To see if the initial gas volume would stabilize in the absence of any event (such as a pump trip), I ran the model for 300 seconds with no pump trip (or any other event). The initial gas volume in the tank rises during the first 50 seconds of simulation and stabilizes at around 2,800 gallons. 

Why is this happening? I was under the impression that if I set the initial gas volume in a sealed tank, HAMMER would assume that volume is under steady state pressure (whatever it is based on initial conditions since there is no field for me to enter the initial gas pressure the way you can do for a dipping tube).