Supersonic twin jets
 

I am trying to model supersonic twin jets(Mach 1.5) impinging on a flat plate using LES. My domain is basically a rectangular tank (17.55cmX11.7cmX3.9cm). The top surface has two circular faces(1.95cm diameter) for the jet inlets. so, the boundary conditions I chose are:
bottom surface-wall
top surface except the inlets-wall
inlets-pressure inlet
all the side walls- symmetry

I set the solution to run for 2500 time steps, the step size being 2e-06 seconds.
Fluent autosaves my solution every 100 time steps. So, now after 800 time steps, I checked the Mach contours to see how the solution is progressing and I found that the inlet conditions are changing with time. i.e Near the inlet faces, where the jet Mach number is supposed to be 1.5, the mach number is different at different time steps. I am unable to understand why this is happening. I checked again and again and all the inputs seem to be correct. In the materials section, I set the density of air to ideal gas, in the pressure inlet boundary condition, I gave the correct values for total pressure, static pressure and total temperature. So, by isentropic equations, my Mach number must be fixed at 1.5 near the inlet. But my solution is showing different values at different time steps. Please help.