[Ardalan]

I'm trying to simulate the discharge of compressed air from a pipe. The pipe is 4 meters long which is initially is filled with compressed air with the pressure being equal to 1.5 atm and the initial temperature is 100C. also two ends of the pipe are closed. The transient flow begins when one end of the pipe is completely opened and subjected to atmosphere. As I can see from my solution a rarefaction wave propagates from the open end towards the closed end. The pressure behind this wave is equal to atmosphere pressure. Everything seems reasonable until the time at which, this wave reaches the closed end.As this wave reached the closed end, the pressure at the closed end reduced and reached about 0.6 atmosphere! then the rarefaction wave was reflected and traveled toward the open end reducing pressure in the pipe below the atmosphere pressure. At this stage the velocity at the open end became negative, in other words the flow was drawn inside the pipe from the surroundings. When the wave reached open end it was reflected as a compression wave ( pressure increasing). When this wave was reflected from the closed side and returned, then it changed again to rarefaction nature. This process continued till the pressure along the pipe reached the value of atmospheric pressure and the velocity became near zero.
I solve Euler equation with perfect gas EOS employing AUSM method. and I use some kind of characteristic boundary condition for the open end and mirror technique for the closed end.
In my idea this can not be a physical condition. In your idea, where I have made mistakes? Is the problem with boundaty conditions?