Pump Trip - Correct method to model

Hello JP,

Can you elaborate on what you exactly wish to accomplish?

As per my understanding, you are conducting Fire Flow Analysis for a building in which your pressure at the top floor is exceeding the limit of 700 kPa. Also in this analysis you are simulating Pump Trip / Shutdown to observe the transients in your system. Am I correct in my understanding? If not please clarify.

The solution two this is in two parts;

First you can develop your model in WaterGEMS / WaterCAD and try using the Automated Fire Flow Analysis Tool. With this you can specify your fire demands and analyze your system to understand how the fire demands are fulfilled and the resulting pressures due to high demands. You can put constraints on the flows & pressures in the Fire Flow Alternative. You can refer the chapter "Automated Fire Flow Analysis" in the Quick Start Lessons.

Once you are satisfied with the results, you can bring the model in HAMMER to analyze for a pump shutdown scenario and observe the transients produced in the system. From the model you shared (Pump Trip), you have created several scenarios for different conditions in HAMMER. However, HAMMER is used for transient analysis. By running the different scenarios you have envisioned, you will be able to observe the negative pressures (the the vapor produced due to water column separation) produced due to a pump shutdown or sudden valve closure.

Is the above mentioned workflow as per what you wish to accomplish?

As far as the model is concerned here are my observations;

1. In the scenario "With D2A", the D2A element is directly connecting to the Reservoir R-74. The status is set to "Open" initially hence the D2A element "Closes" after 5 secs (as prescribed) resulting in high pressures (magnitude of 1400 kPa).

2. In the scenario "With Pump +TCV", the check valve of the pump is closing after 5 seconds. The TCV is reducing pressure till 5 seconds only after which due to the check valve closure, there is no flow to the junction. Hence pressure reduction due to TCV is upto 1158 kPa only before the pump shuts down.

3. In the scenario "With Pump" the pressure is initially zero and after the pump stops after 10 secs, the pressure variation is erratic (moving from negative to positive).

4. In the scenario "With Pump - node elev" the pressure is dropping from 900 kPa to about 300 kPa.

In all of these scenarios, the negative pressures or the air vapor volume produced is of small magnitudes.

It would help if you elaborate on what you wish to accomplish using the above results. If you are using HAMMER, then simulating some extreme event (such as pump shutdown or sudden valve closure) would help you understand the maximum and minimum pressures produced in the pipeline and you can try out different surge mitigation methods to reduce these transients. Mostly such studies are done for rising mains in water supply system.

Here are some reference articles which you may find useful;

Modeling Reference - Discharge To Atmosphere

Modeling a pump shut down transient event

Hope this helps.

I agree that we need a bit more information in order to help you determine which approach is "correct". This will depend on exactly what you are trying to model. In general, the "correct" method is to model the system as close as possible to the actual system. So, understanding how to do this in HAMMER requires us to know exactly what the real system will be, and what you want to assume will happen in the transient event.

Modeling a pump "trip" is done by setting the pump's "Pump Type (Transient)" to "Shut after time delay" (see the article Yashodhan linked to on modeling a pump shut down event). This simulates power being cut at the time entered for the "Time (Delay until shut down)". After this time, the pump impeller will take some time to ramp down, based on the Inertia you enter in the Transient tab of the pump definition. Also, unless you specify a check valve or valve closure via the pump's "Pump Valve Type", flow can pass through the pump if the hydraulic conditions warrant (which can explain flow continuing after the pump shuts down). Furthermore, if air or vapor pockets form in the system, the water column may still move away from the pocket as its size increases. See: Positive flow in pipe downstream of closed element or air valve during transient simulation

If you are trying to model a building whose sprinkler system has turned on along with a pump to supply the sprinklers, and you want to see what happens when the pump suddenly turns off (to develop changes if the resulting pressure envelope is not acceptable), then you could model the sprinkler outflow as demands at junctions.

When the pump shuts down and the pressure at the sprinkler demand locations drops to zero, the outflow will drop to zero. This is because junction demands in HAMMER are always treated as pressure dependent during the transient simulation. See: How are demands treated during the transient simulation in HAMMER?

When looking at your model, I do not understand what the different scenarios represent. For example in the scenario "With Pump - node elev - 10 sec", there is a single pump connected to a single downstream demand, with several reservoirs upstream. Does the single demand represent the sprinkler system you're pumping to? What do all the reservoirs represent?

In the scenario "With Pump + TCV", the pump and demand are still there, along with all the upstream reservoirs, but there is a TCV valve between, configured to close immediately at 5 seconds. The pump was already set to shut down at 5 seconds, so I do not understand what the TCV is for. If you have a valve built-in to the pump that also closes when the pump shuts down, you can model that by way of the "Pump Valve Type".

In the scenario "With D2A - 10 sec", the single demand junction is replaced with a D2A, but the pump is gone and there are only the reservoirs. The D2A is set to start open and then close, which would depict a transient event from a valve closure. Were you trying to simulate the pump shutdown with this? Using the actual pump element would produce more accurate transient results (vs. the D2A or TCV method). Again, you want to model things as close to reality as possible.

If you can explain a bit more about what these elements represent, we can provide more specific assistance regarding how to model this transient simulation.