The results of a transient run are not matching the steady state results of a valve set to a given setting.
As a test to assure that the transient results are working as expected, you may run the initial conditions for a model not only on the starting position of the run (which is used at the beginning of the transient simulation), but also the end position. For instance, you may have a series of valves that are closing in a model and want to see how the model should look at the end of the simulation, so you may run a steady state with the valves in their final position. In some cases, you may see results from that final steady state differ from the end of the transient run.
This can occur because of assumptions related to how valves operate in HAMMER. Most valves will not modulate or see a significant change in the initial setting, even if there is some sort of pattern or rating curve associated with it. In some cases, like a PRV or FCV, the solution may be to model this as a TCV. The following link has more information on this: Modeling existing valves as Throttle Control Valves in HAMMER.
If you are not able to find a suitable way to calculate the discharge coefficient for the valve, or if you want to try to simulate how valve follows a given pattern or rating curve, another option may be to use the Orifice between Pipes element. For example, suppose you have a GPV with a headloss curve. If you were to compute a model, you may see a given results for the headloss through the GPV. But since HAMMER will not following headloss curve, it will simply use the starting value from the initial conditions through the entire transient run. One way to workaround this would be to use the Orifice Between Pipes element. The Orifice Between Pipes element allows you to enter the "nominal" headloss characteristics and flow conditions. It then uses this to establish the headloss characteristics through the valve, and use those during the transient simulation to model how the headloss changes with flow.
Modeling Reference - Valves
Modeling an Initially Partially Closed Valve
Transient pressure wave not dampening or unexpected lack of headloss