Non-Reopening Check Valves

What is a non-reopening check valve? This element is a modeling convention, as opposed to an item that can be ordered from a manufacturer, because by definition a check valve is designed to open and close in response to flow and pressure conditions. The way Pipe2016 models an “ordinary” check valve is by allowing it to open when the differential pressure across the valve reaches any positive value, even as small as +0.00001 ft of head. The reverse is also true, the valve starts closing when the differential pressure drops below zero, even very slightly. “Ordinary” check valves in Pipe2016 are those which are modeled as an internal node, or as a feature associated with a pump, an active valve or a loss element.

However, a real-world check valve has inertia and cannot move as freely as a modeled check valve. Most real-world check valves need a minimum pressure differential of approximately 1 to 2 psi to start opening once they are in the closed position. In many installations, this kind of differential pressure may not be available after a pump trip event once the check valve is closed. In such installations, while the check valve remains closed in reality, the modeled representation of the check valve can behave differently.

Specifically, the modeled check valve may start opening and closing because of tiny differential pressures caused by small variation in flowrates arising from local transient wave action. Use of a non-reopening check valve prevents this kind of check valve “chattering,” or pressure spiking, which can be useful to differentiate between reality and artificial numerical-instability-related chattering. When the user wishes to differentiate between these two effects, the model can be run using ordinary check valves, and then rerun using the non-reopening check valve option as a diagnostic tool, allowing the user to compare the different outputs.

Some real-world check valves do approximate non-reopening check valves: One example is that of a check valve located at ground level on the discharge side of a vertical turbine pump sitting in a sump well below ground level. In this case the pressure is very low or even below zero gauge in the pipe between the pump and the check valve, and the water pressure on the outlet side of the check valve can keep the valve closed. Other situations can occur such as when pumping to a very high elevation, excess pressure on the outlet side of the check valve can prevent it from reopening. A third example is a check valve with a high spring constant or heavy counter weight.  However, sometimes the assumption of a non-reopening check valve is too conservative and may not represent real-world conditions, even in situations where it may appear that the check valve should remain closed.

As an alternative to the ordinary check valve, a check valve element is available, which is an end node that was added with the Pipe2012 release.  With the check valve element, the user can specify two criteria which more accurately model check valve inertia: the first is percent steady-state. This represents the percent of the steady-state flow (i.e., the flow present before the initiation of the transient event) that is required to start the valve closing. So at a percent steady-state = 5, and a steady-state flow of 100 gpm, the check valve will begin to close when flow falls below 5 gpm The second criterion is the differential pressure above which the check valve will begin opening, in feet (or m) of head. These criteria are in addition to check valve closing time, which is a feature of all Pipe2016 check valves. This check valve element can be used by itself, or it can be added immediately downstream of a pump, an active valve or a loss element, as opposed to using the check valve feature which is internal to those node types.

As a final note, we advise users not to place a check valve element downstream of a pump, active valve or loss element that also has an internal ordinary check valve: this can create artificial flow conditions that are not representative of the system being modeled.