Pump Emergency stop by Inertia

Hi Friends,

I have three running pumps in my model and for pump emergency (abnormal) shutdown I did the following setting:

the full stop time for one pump shutdown scenario (speed drops from 2900 to zero) is about 2.85 sec, 

for 2 pumps simultaneous shutdown, the stop time is 4.5 second

but the full stop time for the 3 simultaneous pumps shutdown is about 26 seconds

I just confused about the results, since each pump stops based on defined inertia, so the stop time should be the same.

because of this slow stopping in three running pumps scenario, there is no cavitation and negative pressure in the system while for a single running pump there is cavitation and vapor generation. This means that three simultaneous pumps shutdown in better than two simultaneous pumps shutdown and two is better than one, but in reality, it is not correct, more pumps shutdown worse conditions.

I think the stop time should be same for each pump and independent from how many pumps are running simultaneously, 

have any idea how can I fix this problem or is this normal?

Thank You.

Parents Reply
  • but I expected the more negative surge pressure (more vapor generation) for the 3 simultaneous pumps shut down rather than two and one, but the result shows that the one pump shutdown is the worse case, is it correct in reality or this is the software issue?

    In my experience, users often see surprising results in transient simulations that they may not have expected. This is due to the complex interactions and phenomenon that can occur. In your particular case, shutting down three pumps instead of one or two may change the hydraulics of the network such that the check valve closes at a later time, when the velocity is different, causing a different response to that closure. There could be a complex interaction going on between your pumps, the check valve, the velocity in the pipes, the inertia of multiple pumps vs. the inertia of one pump alone)

    So, the short answer to your question is that the results should be accurate as long as your data input and assumptions are accurate (and as long as the user notifications do not indicate any major problems), but without a deep examination of the model, I cannot say specifically why the results are different other than my observations above. I would recommend taking a close look at the interactions occurring in your model using profile animations as mentioned in my previous reply, to get more confidence in exactly how the network is reacting in each case. (with close attention to the check valve closure)


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

No Data