This discussion has been locked.
You can no longer post new replies to this discussion. If you have a question you can start a new discussion

Pipe fills tank modeled like reservoir from top and shows vacuum in Hammer simulation?


I have system, consisted by pump, rising main and in the end reservoir (according to the examples, it is modeled with reservoir, not tank). The reservoir fills from top, or I thing that I have modeled like that.

Elevation - 721 m; Elevation (inlet/outlet) - 721,5 m.

After simulation in Hammer with no protection there is a vacuum at the end of rising main. The strange thing in my opinion is that there is a vacuum in the rising main near the reservoir. Is it the software take 

in account that the pipe ends with contact with the atmosphere (it flows above water level) and may be in practice it is some kind of big air valve, or in other words the end is one big inflow orifice?

  • Hello Boyan, 

    Could you please upload a screenshot of layout of your model? What kind of transient simulation you are performing like pump shutdown, pump startup etc.? 

    Generally,  the wave front causes the pressure to drop below zero as it approaches the reservoir and is then reflected backwards. Where is the location of the pump, is it near to the point where you see this vapor formation? 

    We recommend enabling the animation option if you have not done so already, so that you can animate the profile path and better understand what is happening here. See the information in this wiki article .


    Sushma Choure

    Bentley Technical Suppport

  • Hi, Sushma,

    In the the first figure below is the layout. For my convenience this looks like mix between longitudinal profile and  axonometric scheme mainly for pump station).

    In the second figure there is the section of pipeline near to the reservoir.

     KBH means "elevation of water level", the invert of inflow pipe is above the water level

  • Hello Boyan,

    The "Elevation (Inlet/Outlet)" for the reservoir you mention is above the elevation of the reservoir. This elevation would typically be below the elevation. If the inlet/outlet pipe is above the water surface elevation of the reservoir, as you mention, this would more likely act like a top-fill tank. See the section at the bottom of the link for information as it related to HAMMER.

    In HAMMER, these are typically modeled with the discharge to atmosphere element. This is because any transients would no longer occur at the discharge point.



  • Hi, Scott,

    The element D2A has to be the correct solution of pipe discharges in the atmosphere. 

    But unfortunately when I try to model D2A, I don't have any significant difference with the reservoir case results, i.e. I have big vacuum and big pressure at the end of pipe. May be I make a mistake somewhere (but not in transient operational pair for P and Q, because I made it).

    After this fail I tried to achieve the effect of air discharging by emitter node (with various emitter coefficients) and air valve near to the end of pipe. Again the effect is negligible. But if the emitter node is moved somewhere in the middle of the pipeline, the big pressure  decreases. And with the air valve is the same - vacuum disappears in near proximity when it has been moved somewhere in the middle of pipeline. That is why I think that there is some problem when air valves and emitter node is very close to D2A.

    Finally I tried to model the end of my pipe with demand node, because I read in the suggested by you material (for tank) that demand is always pressure depended when a transient is modeled. But then arises another problem - error message that there is negative pressure in this end node.

    Is it possible to send you by email the model with D2A?

    And another question or something for confirmation: Emitter node decreased the pressure when was away from the pipeline. The mitigation is expected effect, but I noticed that it also decreased the vacuum in the near proximity, which is also logical. I didn't know that emitter node in Hammer could works in the reverse direction, i.e. intake an air in the pipeline ...

  • Hi Boyan,

    Can you clarify what you mean by "emitter node"? Do you mean a junction with an emitter coefficient? If so, HAMMER treats this situation the same as a regular demand node during the transient simulation - it sees the initial pressure and outflow from the initial conditions, and uses that as the transient outflow to construct its own discharge coefficient that relates pressure to outflow. Meaning, the total calculated demand for a junction (including both regular demands and emitter demands) is treated as pressure dependent during the transient simulation. See this article: How are demands treated during the transient simulation in HAMMER?

    The D2A element is similar in that it is treated as a pressure dependent demand - it is just a different way of establishing the initial pressure and outflow (as a pair of "typical" pressure and corresponding flow), and it permits air intake if the pressure drops below zero. (at the D2A node itself, as noted in the D2A technical reference article previously mentioned) In the profile path from your screenshot it appears air intake is not occurring as the air/vapor volume noted above that point on the right side is zero.

    Note that junction demands do not permit air inflow (a vapor pocket will form if the pressure at the node drops below the vapor pressure limit). The difference in transient mitigation depending on the location of your "emitter" (junction with emitter coefficient, I assume) is likely due to the complex interactions that occur. Meaning, the change in pressure may have changed the resulting initial outflow, changing the amount of flow leaving that point, changing the velocity in the nearby pipes, changing the transient results from changing that velocity, and so on.

    Generally speaking a wave will reflect differently for different boundary conditions (outflow such as a demand or D2A, dead end, or open reservoir) and it is important to animate the profile path to get a better understanding of the complex nature of what happens during a transient simulation. This will give you better confidence in your results. You may find that the wave front from the transient event will reach a certain point before it reaches the end of the system (for example cause vapor pressure conditions to occur before the wave even reaches the end). See: Using Transient Results Viewer animations for visualizing a transient simulation

    With that said, if you still need more help understanding the results, please use one of the two options in the following article to send a copy of your model. Please be sure to include details and steps to reproduce the issue. Sharing Hydraulic Model Files on the OpenFlows Forum


    Jesse Dringoli
    Technical Support Manager, OpenFlows
    Bentley Communities Site Administrator
    Bentley Systems, Inc.

    Answer Verified By: Boyan Borisov