Pipe flow goes negative

In the model above, I am expanding an existing water system to a new building.  Per the results of hydrant flow tests, the available flow rate at R-4/PMP-4 is 1,060 gpm and at R-3/PMP-3 is 840 gpm (per a prior question, I am using an approximation method to model the hydrant flow tests from an article that was provided Bentley). My question: why would the flow in pipe P-7 go negative and why would the system try to pull more water from R-3/PMP-3.

See attached for two clips from the software.

PDF

Parents
  • Hello Maria,

    The negative flow is because the start and stop nodes for P-7 is reversed. The flow direction is correct, so this isn't really an issue. You can reverse the start and stop nodes in the pipe properties or use the Flow (Absolute) field to report for the flow.

    For the flow from R-3 and PMP-3, this could be related to how the system is being modeled. The approximation method from this link assumes that you have one connection point. If you have more than one connection point, the approximation can break down as the pumps work against each other. See the section on limitations and troubleshooting at the end of the link. 

    If that is not the methodology you used, it could be a matter of calibration. The pumps will generate enough flow to satisfy the demands, but since there are two pumps there may be some impact on the headlosses in the pipes between the pumps and the demand node. Adjusting the roughness values may have some impact on the results.

    Regards,

    Scott

Reply
  • Hello Maria,

    The negative flow is because the start and stop nodes for P-7 is reversed. The flow direction is correct, so this isn't really an issue. You can reverse the start and stop nodes in the pipe properties or use the Flow (Absolute) field to report for the flow.

    For the flow from R-3 and PMP-3, this could be related to how the system is being modeled. The approximation method from this link assumes that you have one connection point. If you have more than one connection point, the approximation can break down as the pumps work against each other. See the section on limitations and troubleshooting at the end of the link. 

    If that is not the methodology you used, it could be a matter of calibration. The pumps will generate enough flow to satisfy the demands, but since there are two pumps there may be some impact on the headlosses in the pipes between the pumps and the demand node. Adjusting the roughness values may have some impact on the results.

    Regards,

    Scott

Children
No Data