Modeling the filling of an empty pipe in HAMMER

Applies To
Product(s): HAMMER
Version(s): 08.11.xx.xx, 10.xx.xx.xx
Area: Modeling
Original Author: Scott Kampa, Bentley Technical Support Group


How can I model a case where an empty pipe is filling using HAMMER?


There are cases where a user may want to model an inrush event, or a case where an empty pipe is filling. A transient can potentially occur when the air in the empty pipe is expelled. It is possible to model such a case using the Discharge to Atmosphere element, but the user must be aware of some limitations with regard to the tracking of air pockets in HAMMER.

When a pump turns back on in a water system or a sewer force main system, it may expel some air from the downstream end. The headloss through the discharge opening causes a resistance that can result in a severe upsurge once the water column reaches the opening. For example, if there is a small orifice at the end of the empty pipe, an upsurge can occur when the flow reaches it since the water cannot exit the pipe fast enough.

If you need to analyze the transient effects of the air being expelled, this can be done by using the Discharge to Atmosphere element, operating as an orifice. The initial conditions must describe the low head condition (zero pressure at the discharge to atmosphere element) and you must enter a volume of air in the "Gas Volume (Initial)" field. You would then have the head increase during the transient simulation (pump turning on or periodic head element on the upstream side with head value increasing, for example.) The "Flow (Typical)" and "Pressure Drop (Typical)" of the discharge to atmosphere would be estimated using the orifice equation. Basically the higher the "Pressure Drop (typical)", the smaller the orifice size, and the more resistance to flow, resulting in a higher upsurge after the air pocket is expelled.

It is important to note that the larger the air volume you introduce into a HAMMER model, the less accurate the results may be, since you are deviating further from normal transient theory. Information on the assumptions made in HAMMER related to air and/or vapor in a transient simulation can be found in the theory section of the Help documentation and in this article: Assumptions and limitations of tracking air or vapor pockets in HAMMER

Also, the value entered for "Gas Volume (Initial)" will impact the timing of the release of the air. The value you enter will be up to your engineering judgment, but a good starting point may be the volume of the "empty" pipe. A larger volume of air for the same size orifice will take longer to be expelled from the D2A. This, in turn, will impact the head increase at the source. The most important impact on the system will occur with the air is fully expelled, which is when the transient would occur. So while a large air volume will take longer to expel, the setup and size of the D2A may prove to the be most important part of the transient event.

Note also that if you're interested in simulating the time it takes for an empty pipe to fill, the timing of that will be influenced by the assumption that the air pocket is located at a discrete point (the D2A / high point). Using the D2A approach with the initial air volume set equal to the empty pipeline volume, the upstream pump would not "see" the correct head. Meaning, when it starts up, it will be pushing against an HGL equal to the high point, whereas in the real system, it will be pushing against a very lower head (since the pipe is drained) which will gradually increase as the rising pipe fills with water. This effects the flow rate from the pump operating point and therefore the time taken to fill the pipe. An alternative to consider would be to use the Implicit dynamic solver in SewerGEMS.

See Also

Modeling Reference - Discharge to Atmosphere

"Initial pressure less than vapor pressure. At the pipe end(s), the elevation(s) or head(s) are incorrect"

Assumptions and limitations of tracking air or vapor pockets in HAMMER