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

Hammer model setup/errors

Hi,

I'm trying to do transient analysis for a water distribution network. The network comprises a reservoir, a pumping station (6 duty pumps at max. flow condition), and a 37 km pipe network, sizes ranging between DN100 and DN800. The demands are modelled as emitters. I'm running all pumps trip scenario

This is the first time I use Hammer. I'm finding it frustrating on many levels compared to other software I used before, but the main issues I have are:

1) The wave speed/pipe length adjustment. When I ran the model, the time step was automatically calculated as 0.06 s, and I was getting the message that says there is an excessive adjustment to the wave speed/length. Therefore, I made sure that the min. pipe length in the model is equal to (2x0.06xpipe wave speed) by eliminating many junctions. This resolved the issue for system with no protection. However, I wanted to add vacuum breakers. I added branch pipes of 1m length to the main network and added the vacuum breakers at the end of those pipes, but consequently the wave speed/length message reappeared (presumably because of the short pipe lengths), and the time step now has changed to 0.11. How can I deal with this situation?

 

2) After inserting the first branch pipe and vacuum breaker, the model ran with the previous warnings (wave speed/length  adjustment, negative pressure). But when I added the second vacuum breaker, I got the error message "Invalid results. Please check the input data for the for the connected elements". The highlighted element is a pipe in the network, of which one end is a junction that is connected to a 1m branch pipe and a vacuum breaker. The vacuum breaker elevation is 1m higher that the junction on the main network, and the pipe is 1m long and of 100mm dia. I don't see any wrong inputs here. What is the solution?

Help!

Regards

Fadi Sirrieh

 

  • Hello Fadi,

    I see that you also submitted a service request for this. If it is okay, I would suggest that we keep the communication in the forums as that would allow other users to benifit from potention solutions to this issue.

    1) The wave speed/pipe length adjustment. When I ran the model, the time step was automatically calculated as 0.06 s, and I was getting the message that says there is an excessive adjustment to the wave speed/length. Therefore, I made sure that the min. pipe length in the model is equal to (2x0.06xpipe wave speed) by eliminating many junctions. This resolved the issue for system with no protection. However, I wanted to add vacuum breakers. I added branch pipes of 1m length to the main network and added the vacuum breakers at the end of those pipes, but consequently the wave speed/length message reappeared (presumably because of the short pipe lengths), and the time step now has changed to 0.11. How can I deal with this situation?

    This is related to the short pipes in the model. HAMMER tries to have a wave travel from one end of the pipe to the other end in even multiples of the time step. Therefore, Hammer will try to adjust the length of the pipe to allow this to happen. By default, if it has to adjust the length more than + or - 50% (the default option at the bottom right of the transient timestep options window), then you get that warning. With 1 meter pipes, the length adjustment required in order to have the wave speed. With the 1 meter pipe, it will not take much adjustment in length for this default length adjustment value to be violated, particularly for longer time steps.

    The calculation time step used in Bentley HAMMER can be defined by the user, or the user can elect to have Bentley HAMMER automatically select a time step for them. It appears that you are using the latter method. In general, smaller time steps will allow for more accurate results, since the length/wave speed adjustment will be less. However, smaller time steps also means the run time will be longer.

    If Bentley HAMMER selects the time step, it will attempt to ensure the time step provides a good trade off between solution accuracy and the time taken to compute the simulation. The time step selected by Bentley HAMMER generally requires some adjustment to the pipe lengths or wave speeds. The adjustments are done automatically by Bentley HAMMER, but the user is able to select whether they want the length or wave speed adjusted. Similarly, if a user enters their own time step, Bentley HAMMER will adjust the pipe lengths or wave speed accordingly and once again the user can select which of these parameters is adjusted. See the Help document "Selection of the Time Step" for more information on this.

    In this case, you may want to explore using your own time step. To do this, go Analysis > Calculation Options and double-click the transient calculation options to view the properties. Set "Is User Defined Time Step" to True. A new field will be avaiable so you can enter your own time step. I would suggest trying 0.01 seconds to start. You may still see some pipes that have a length adjustment greater than the value entered in the Transient Time Step Options (Analysis > Transient Time Step Options), but the adjustment should be smaller. You can actually see the adjustment in the pipe FlexTable. Look for the columns Length Adjument and Length Adjustment Percent. If they are not there, you can add them by clicking the Edit button in the FlexTable; from there, you can add the field from the Available fields table.

    If the length adjustment is still too high for your modeling case, you can try a smaller time step. Keep in mind that the model will take longer to compute for smaller times steps. You may also want to do a sensitivity analysis and compare the results you see between the different time steps. If the results are not significantly different for different length adjustments, you may be able to infer that the larger time step will still give you accurate results without taking a long time to compute.

    You can find more information on the length/wave speed adjustments at the following support solution: communities.bentley.com/.../9401.aspx

    Regards,

    Scott Kampa

    Bentley Technical Support

  • Hello Fadi,

    2) After inserting the first branch pipe and vacuum breaker, the model ran with the previous warnings (wave speed/length  adjustment, negative pressure). But when I added the second vacuum breaker, I got the error message "Invalid results. Please check the input data for the for the connected elements". The highlighted element is a pipe in the network, of which one end is a junction that is connected to a 1m branch pipe and a vacuum breaker. The vacuum breaker elevation is 1m higher that the junction on the main network, and the pipe is 1m long and of 100mm dia. I don't see any wrong inputs here. What is the solution?

    I may need to see the model to look into this second issue, but there are a few things you can try. First, you can check to see if a smaller time step, such as the manual time step of 0.01 seconds mentioned in the earlier post, helps at all. This will have some impact on fields such as pipe length and may offer better results.

    Aside from property fields like elevation, you should also review other properties of the air valve. For instance, make sure that the orifice diameter or air flow curve is correct. Check the setting for "Treat as Junction" as well. If this is set to False, try setting it to True to see if that helps.

    After computing the initial conditions, you can also run a validation (Analysis > Validate). As long as the initial conditions are computed, the validation procedure will review the transient settings of the elements in the model in case there are any obvious data entry issues that need to be fixed.

    If that doesn't help, I will need to see a copy of the model in order to look into this. There are two options for sharing your model files on BE Communities. If you would like the files to be visible to other members, compress the files into a zip file and upload them as an attachment using the ‘Advanced Reply editor’ before posting. If your data is confidential, you can follow the instructions in the link below to send it to us via Bentley Sharefile. Files uploaded to Sharefile can only be viewed by Bentley.   http://communities.bentley.com/content/bentleysecurefilesupload.aspx

    Regards,

    Scott Kampa

    Bentley Technical Support

  • Hi Scott,

    Thanks for the quick responses.

    I tried 0.01 s time step. The simulation ran fine but it took approx. 5 minutes. I then tried 0.02s, but I got the same error message. I can not afford 5  minutes simulation as there are multiple scenarios I want to run for this network, and there is a similar network that I also need to analyse after this one.

    Is there a better way to setup the model to avoid this siuation? E.g. if I breakdown each pipe into equal segments by inserting multiple junctions, would that help?

    I've uploaded the mdoel file for you to look at.

    Thanks and Regards,

    Fadi Sirrieh

  • Hello,

    We have looked at the model, trying a few things to see what we could find out. In computing the active scenario (Surge_All Pumps Fail_NO CV), the error message regarding invalid results was generated. We changed the time steps to a custom time step of 0.01 seconds. In doing this, it is clear there is a severe surge after the pump shuts down, with causes vapor pockets to form along the path. With a lot of transient events occurring, this can make the model sensitive to instability. This may be what is happening in your modeling case.

    One top of this, you are using the Extended CAV method to analyze the air valves. This does add a level of complexity to the model calculations as well. I will note that when Extended CAV is not used, the model computes without any errors about invalid results.

    The model is probably failing to compute because of a combination of the complicated interaction between the air pockets at the air valve switching between extended and concentration mode and the many vapor pockets that form and collapse in the rest of the system.

    The best solution for your case is to set "Use Extended CAV" (found in the transient calculation options) to False. Extended CAV allows the user to see some tracking of the air pocket in the pipes adjacent to the air valve. However, since your air valves are connected to a short pipe, the benefit of this method is minimal, especially given the overall instability of the results in general.

    I will also note that you may want to look into using a smaller time step than the one generated automatically by HAMMER. The length adjustment from this relatively large time step is very large. A smaller time step will allow for less adjustment. I would still recommend doing a sensitivity analysis. For the short pipes near the air valves, a timestep 0.0009 seconds (of 1.0 m / 1100 m/s = 0.0009 sec) would be needed to prevent adjustment, but that would make for an unreasonably long run duration. However, if you look at different time steps and see no appreciable change in results, you can then make the case for a smaller time step.

    For more information on air valves and how they operate in HAMMER, please see the air valve TechNote found at the following link: communities.bentley.com/.../modeling-reference-_2d00_-air-valves.aspx

    There are another thing worth looking at in the model. First, the pumps have a control valve setting of 2 seconds, while the pump shutdown occurs after 5 seconds. The control valve is a valve that operates from the start of the transient simulation. With a setting of 2 seconds, that basically means that the valve starts closing at the start of the simulation and fully closes after 2 seconds, i.e. the pump is already closed before it shuts down. If there is no valve on the pump to inhibit close, such as a check valve, it is common to set the control valve time to a large value, like 99,999 seconds. While this doesn't resolve the issues in the model, we thought it worth checking out.

    Please let us know if you have any additional questions.

    Regards,

    Scott Kampa

    Bentley Technical Support

    Answer Verified By: FADI SIRRIEH