Help needed to model inertial effects to my vertical water pumping circuit

Hi!!

I am simulating a pumping circuit that have three pumps whose outputs are combined together in a single pipe which is vertical in position. The inputs signals for these pumps are as follows

Due to this varied signal, there will be inertial effects in the water column that exerts pressure at the bottom of the vertical pipe, which I want to find out through simulink simulation. Could you help me do that? I tried with "Fluid Inertia" model. But that doesn't show any pressure spikes in the pressure plot? Pipe blocks available in simulink don't account for inertial effects. But surely there will be pressure spikes in the actual pumping. Could anyone help to model this phenomena in Matlab simulink? This is quite urgent for my research.

Following is the circuit

Pressure plot

Thing is how to get that pressure spikes? That's my only concern in this problem. Please help me out!!!

You may contact me on [e-mail]

Thanks

Immanuel Keba

[Yifeng Tang]

The way the model is put together looks fine to me, and I would expect you see pressure spikes even without the inertia effect. Varying flow through the pipe should give you varying pressure drop but you are not seeing that. This leads me to wonder whether the pressure relieve valves you have at the outlet of the pumps are triggered so the outlet pressure at all pumps remains a constant. Could you take a look? Is the flow being released through the relieve valves?

[immanuelkeba]

Hi Yifeng Tang!

Grateful for your valuable inputs. The reason why I put pressure relief valves is that I want to control the total outlet pressure of Pump to a certain limit. And I removed the PRVs from the circuit and I didn't get spikes either. So I can confirm that pressure relief valves don't play a role there. Now I am trying with "Segmented Pipe LP" block instead of Hydraulic Pipe LP block which accounts for Frictional losses, Fluid Inertia & Fluid compressibility together. I hope I am close to my solution.

But this block is named under low pressure block in Foundation liberary whereas my problem involves high pressure. Does this block still applicable to my problem?

I will greatly appreciate if I could contact you for further clarifications. Email : [e-mail]

Could you send a message so that we can discuss further for the betterment of the solution please?

Thanks

Immanuel Keba

Chennai, India

[Yifeng Tang]

Hi Immanuel,

If it's possible for you to share the model on MATLAB Answer, I can take a look. If not, I suggest that you contact your sales rep and get connected with an application engineer in our India office. You can try tech support option, too.

[immanuelkeba]

Hi Yifeng Tang,

Glad that you are ready to help. But I cannot share my file here. Could you please share your mail ID or send me a startup msg to [e-mail] so that I can share the file over there for solution please!!!.

Thanks

Immanuel Keba J

[immanuelkeba]

Hi Yifeng Tang,

I went on to try something new. Instead of using a "Segmented Pipe LP" block which accounts for Fluid inertia, Resistive losses and Fluid compressibility, I tried using a "Hydraulic pipe LP" and "Fluid Inertia" blocks in series. The reason is I didn't get meaningful results by using a Segmented Pipe LP block (I got meaningless pressure undulations and flow rates at the outlet). I couldn't find any problom with my circuit for that.

Now the problem is, in which way should I connect them in series and what is the significance of it?

1. Hydraulic pipe LP at first followed by Fluid Inertia as shown

2. Fluid inertia block followed by Hydrauilc Pipe LP as shown

Flow happens from bottom to top up

Does the order make any significance?

Because I received meaningful pressure outputs but not the flow rates for configuration 1 and I received meaningful pressure as well as flow rate outputs for the configuration 2. I wish I get adviced and taught by you regarding the Order of the aforesaid connection.

Your response is highly appreciated.

Thanks

Immanuel Keba J

[Yifeng Tang]

Hi Immanuel,

Is there any reason that you have to use R2018a and the Hydraulic domain? The physical effect you are interested in (inertia, compressibility, etc.) can be better handled using the newer Isothermal Liquid domain released in R2020a. See the doc page: https://www.mathworks.com/help/simscape/upgrading-hydraulic-models-to-use-isothermal-liquid-blocks.html

Does your license allow you to upgrade?

Again, I suggest you reach out to the local sales and support team for further help with your model, since you can't share it via MATLAB Answer.

[immanuelkeba]

Hi Yifeng,

Thanks for your contribution. As you've suggested, I am upgrading to Isothermal Liquids.

I just need one more clarification form you. I used "Segmented Pipe LP" block which accounts for fluid inertia, compressibility & frictional resistance. For the aforementioned signal in this question, I received pressure outputs. In which, at the interval (at null period) the pressure goes down to -500bar (negative) which is not practically possible. such phenomena doesn't occur in actual situation. Could you please suggest a solution for this? The circuit is same as shown in this question.

Thanks

Immanuel Keba

[Yifeng Tang]

Is the pressure measured from the sensor over the pipe? If so, it's measuring the DIFFERENCE in pressure over the two ends, so it's possible to get a negative value, meaning the pressure at A is lower than that at B. The spiky behavior is likely due to the inertia.

At the same time, you do want to check the absolute pressure at those ports and make sure those don't go negative. I don't think it would, because otherwise you'll likely see an error message.

[immanuelkeba]

Hi yifeng,

Thanks for the reply.

I have three problems here.

1. As you've said, the pressure curve is the difference of pressure between end A & B and not absolute pressure. But the value goes till -500 bar which says that the difference in pressure is -500bar which is not the actual case in this problem. That spike is expected due to fluid inertia, but the negative spike is my only concern.
2. how to measure absolute pressure in simhydraulics? I don't find any sensors in the library? Is there any other way we can take it from simulation log?
3. The flow rate becomes like the image shown below if i use "segmented pipe LP' (to consider inertia). What could be the reason for the negative values in flow rate? The unit in Y axis is m^3/hr. The fluid is filled initially right? if so, this curve doesn't justify anything.

Please give your inputs

Thanks

Immanuel Keba

[Yifeng Tang]

Usually the unphysical results are caused by unreasonable parameters or driving boundary conditions. But since you can't share the model here, I won't be able to tell. Given that, my suggestion is to complete your conversion to the Isothermal Liquid domain. As you go through the blocks, you may notice something to pay attention to.

To measure the actual pressure, connect the B port of the pressure sensor to a hydraulic reference. You'll then measure the pressure difference between Port A and the atomsphere, which will give you the gauge pressure. There are plenty of such examples in the Documentation. In more recent releases, the pressure sensor in the Isothermal Liquid domain contains an option to measure pressure difference (2 ports) or absolute pressure (1 port).

[immanuelkeba]

Hi Yifeng,

Thank you so much for your guidance.

Grateful!!

Immanule Keba