Pump operational controls

Hello Nastassja,

Is the tank connected to another source? Or is there any way the tank could be filled by some other tank / source in the system? Typically in such cases it would be a good idea to first graph the tank HGL vs time and see how it is changing. Check the pipes connecting to and from the tank to see if there is any reverse flow which is occurring which is filling the tank rather than emptying it.

Please share the model files so that we can check out what's happening;

Sharing Hydraulic Model Files on the OpenFlows | Hydraulics & Hydrology Forum

Hope this helps.

The model consists of a pump that pulls from a reservoir. Water travels to two tanks along an isolated transmission line. Both tanks are connected to the distribution system. There are no other water sources.

PDF

I will upload the file. I am working in Scenario: EPS_AVG Day Baggs

Hello Nastassja,

Thanks for sending the model. This is a hydraulically close tanks problem.

During the time when the pump is off, you see T-4 appear to fill, because it is filling from the other close-by tank T-2. Tanks T-4 and T-2 are so close, that within one timestep (set to 1.0 hr), one tank fills and the other tank drains due to the hydraulic grade difference, which then reverses back and forth between the large timesteps. This problem is described in more detail here:

Rapid flow oscillation between hydraulically close tanks

You can see this if you graph the tank hydraulic grade along with the flow in the pipe between them. This rapid oscillation (from the flow between the tanks) causes the tanks to overshoot the pump control range, which is why you see them fill beyond the elevation of the control.

As a quick way to illustrate the impact this has, if you change the timestep to something very small such as 0.01 hr, the problem no long occurs, because the tanks do not "flip-flop" the HGL between one timestep (in other words, they are able to "settle" to roughly the same HGL). Ideally per the above article, the tanks would be combined together into one single equivalent tank. 

With the 0.01 hr timestep, the upstream pump controls work as expected:

Hello Nastassja,

Thanks for sending the model. This is a hydraulically close tanks problem.

During the time when the pump is off, you see T-4 appear to fill, because it is filling from the other close-by tank T-2. Tanks T-4 and T-2 are so close, that within one timestep (set to 1.0 hr), one tank fills and the other tank drains due to the hydraulic grade difference, which then reverses back and forth between the large timesteps. This problem is described in more detail here:

Rapid flow oscillation between hydraulically close tanks

You can see this if you graph the tank hydraulic grade along with the flow in the pipe between them. This rapid oscillation (from the flow between the tanks) causes the tanks to overshoot the pump control range, which is why you see them fill beyond the elevation of the control.

As a quick way to illustrate the impact this has, if you change the timestep to something very small such as 0.01 hr, the problem no long occurs, because the tanks do not "flip-flop" the HGL between one timestep (in other words, they are able to "settle" to roughly the same HGL). Ideally per the above article, the tanks would be combined together into one single equivalent tank. 

With the 0.01 hr timestep, the upstream pump controls work as expected: