Split and Blend - Controlling Volumetric Flow via Branching Pipelines

[qubitsized]

Hello Everyone!

Suppose I have 3+1 VFD pumps abstracting 18000 m^3/day of water from a filtered water tank. The pumps discharge at a common header. The header then branches into two pipes, one goes to disinfection tank, one goes to the RO then to the disinfection tank. See figure below:

Consider 3 general cases:

Case 1-Water is good: All the water goes to disinfection tank directly.
Case 2-Water is bad: All the water goes to RO, then to disinfection.
Case 3-Water is average: Water is split and blended, say 70% goes to RO and 30% goes to disinfection.

For case 3, will the above configuration work? The valves at the header going to RO and disinfection are both modulating.

Thank you for any opinion/advise!

 

[ashtree]

In principle yes it should work, but i say that with some caution.
The two modulating valves would have to have different control loops and controlled from different elements, but would interact as well. Any change on one valve will cause a change to occur in the other or for the pump speed to change.
The pumps would have to be able to supply enough flow and pressure to make it all work.
The valves would have to have enough sensitivity and control to achieve the necessary flow etc.

I would wonder whether it might be easier to have a couple of pumps dedicated to each of the outputs but have some interconnects and manual valving to allow some cross overs if needed to provide duty/standby type arrangements. Although this will still have some control complexity it would still be easier than the above control scenario.

Regards
Ashtree
"Any water can be made potable if you filter it through enough money"

 

[qubitsized]

Thank you ashtree!

 

[MartinLe]

You can probably simplify the control scheme by having one three way (diverting valve) valve as the 'splitter'.

If you go with two valves, I would look for a control scheme where one valve is alway 100% open.

 

[BigInch]

One valve always 100% open would imply that you would be OK with letting the process with the active control valve also control the other process. You would not be able to control each process individually. That one valve would effectively control both processes. Flow T * x, where 0<=x<=1, sent to one process would probably send something like T*(1-x) to the other. If T*(1-x) would be OK for that process, given the first had flow T*x, then, and only then, would one valve stand a chance of working.

 

[QualityTime]

Lets make it simple. I think you are missing flow meters to control your rate control valves

 

[bimr]

Agree with QualityTime. You need flow meters in order to monitor the flow. Without flow meters, you will never know if the flow splitting system is working.