How to model tanks with inlets above liquid level - partially filled pipes missunderstood?

Hi everyone,

Update:

See last comment:

problems:

• perpetuum mobilé with partially filled pipe (IL) blocks

OLD:

I am trying to model a system in Simscape and Simscape Fluids which looks like the following scheme:

Somehow I cannot figure out how to face the limitations of the Tank (IL) block asking for all inlets being below liquid level.

Those tanks try to equalize pressure and thus there is volume flow between the tanks even if the pumps are shut off (see attached model). This is not the case in reality when there is free outflow above the liquid level.

Any idea anyone?

[Sam Chak]

Hi @Andreas

Do I interpret the diagram correctly? The diagram indicates a closed-loop piping system, but in open-tank system, the liquid can evaporate. So, in reality, the available liquid in the two-tank system will get lesser over time.

What is minimum liquid level needed to equalize the pressure in the system?

[Andreas]

@Sam Chak thanks for your answer.

I updated the diagram a bit:

i am only looking for pressure / hydraulics in this model, so regard the heat exchangers as simple flow resistance.

The tanks are not open (so there is no loss of liquid - did that wrong in the first scheme), but the upper inlets are above the liquid level (min 0.5 m) and above the pump level (> 2.5 m).

Pump P1 has constant rpm, Pump P2 has a pulse signal on rpm (on / off). The liquid levels should behave as in the following scheme:

Important: There is no volume flow when both pumps are shut down hence the inlets are above liquid and pump level. Without check valve, the liquid level of the upper inlet pipes will fall to the corresponding tank level (level upper inlet tank 1 = level tank 2 and vice versa), but the tanks itself won't equalize.

But I have found no way to model this behavior with tank (IL) and partially filled pipe (IL) blocks so far.

In all cases, the tanks will try to equalize its levels.

[Andreas]

@Sam Chak @Steve Miller

It even gets crazier.

I build a model with two tanks (same but different initial liquid level), two partially filled pipes and no pumps. After the tanks equalizing their level, there is still a significant constant volume flow - perpetuum mobilé? When the partially filled pipes are replaced with normal pipes, the flow will converge to zero (normal behavior).

This seems to be wrong with the partially filled pipe - see attached model. Any idea anyone?

[Yifeng Tang]

Hi Andrea,

Looking at the diagram you have, especially the way you'll pump the liquid, I wonder if you really need the partially filled pipe. I put together a simple proof-of-concept model, where the tank outlets and the pipes in between have 2.5m of elevation and two pumps are moving the fluids. The initial liquid levels in the tanks are set at 2m so the A ports are always above the water, which will issue a warning (but that's OK).

Looking at the results of flow rate and water levels in the tank, things look quite reasonable. But I'll let you interrogate the model a bit harder and see if things make sense. I think the key here is that the total elevation needs to sum to zero around the loop, or you'll end up with a (super nice!) infinite source of energy from gravity.

This should work OK as long as the pump is running. If the pump is turned off, however, you may need to add a local restriction block to synchronize with the pump shaft speed to cut the flow. Otherwise I think you'll get some extra mass magically appear in the system.

Also, I used the thermal liquid domain as an example, since I saw heat exchangers in your diagram and I thought you may need this domain later. But the same idea should work with isothermal liquid domain too. Both domains have the tank block and a pipe block from Simscape Fluids library. Only the pipe blocks from the Simscape Fluids, not the foundamental library, have the A-B elevation options.

[Andreas]

@Yifeng Tang Thanks a lot for your effort.

" I think the key here is that the total elevation needs to sum to zero around the loop, or you'll end up with a (super nice!) infinite source of energy from gravity." :-D

I played a bit with your model and as you mentionend, it seems to go magic as without a local restriction block. The total volume of liquid in the loop will somehow rise and the liquid level in both tanks will level out when the pumps are shut off - see attached model.

My approach is to use the check valve, but i still hope there is a proper way to couple partioally filled pipes with thanks without creating infinite flow/suction/mass creation.

[Yifeng Tang]

Hi @Andreas,

I modified the model following the two models in this example (IL domain):

https://www.mathworks.com/help/releases/R2023a/hydro/ug/partially-filled-pipe.html

The resulting model in TL domain is attached. Could you please interrogate it and see if it's behaving as expected?

I am cycling one of pump to be on and off 50% of the time, while have the other pump controlled based on water level. Didn't observe any free power from gravity this time, lol.

[Andreas]

Hi @Yifeng Tang ,

thanks again - your model works and withstood all my testing so far. I'd like to mark your comment as solution - would you mind changing it to an answer?

[Yifeng Tang]

Good to know! Will do that.

Answer 1

[Yifeng Tang]

You can find here two examples on how to use a tank block and a partially filled pipe together in an isothermal liquid model. The corresponding components in the thermal liquid domain work in the same fashion. An example model, built to mimic your drawing of the design, is attached.

It appears to me that the key here are:

(1) make sure the total sum of height changes is zero for any loop.

(2) include an offset of the tank outlet height when calculating the AL signal for the partially filled pipe block, following the example in the Doc.