Water control issue

[one2go]

Hello everyone,

i have problem with water control, let me explain the process
Customer has pump connected to vfd and controlled by plc, informations are gathered from pressure and flow transmitter. Pump has to be run at minimum of 60% (30 Hz) power because of overheating. They have water concession contract with distributer for maximum flow of 29 l/s. What they want is to maintain pressure at set value and limited flow. Water consuption varies from min to min, depends on production(its not possible to predict). I succeed to keep wanted pressure but flow is going over limit. I would like to hear your suggestions how would you stop flow from going over limit and sometimed allow to go a bit higher?

Currently Im using PID control for wanted pressure and im blocking flow by reseting PID's output, but when consumption is high, output is a bit "dancing".

Thanks forward.

 

[jraef]

With an unpredictable useage you can control one aspect or the other, but not both without a reservoir of some sort.

 

[one2go]

yes I thought the same :/

I have already told that to customer, but he wants sheep and money in same time

 

[cardosocea]

If the plant demands more than 29 l/s and you limit it, the pressure will always be lower. You can always present two quotes to fix the problem...

Quote 1 - install a reservoir (which may not be feasible depending on location)
Quote 2 - bend the laws of physics.

I'll let you put a price on both.

I have a similar problem, although we have a massive reservoir, and the solution was to install a control valve that diverts flow back to the reservoir. What we are trying to achieve is to maintain the pump working in its ideal condition to avoid long term damage to it as happened in the past.

 

[RussB]

A classic example of a pooly engineerd project attempting to be "fixed" by the programmer.

This should be a "sticky" and labeled as such.

 

[cardosocea]

I think it's really self inflicted... we achieve so many miracles (from the perspective of the people counting pennies) that to them this is just another thing you can fix.

 

[one2go]

A classic example of a pooly engineerd project attempting to be "fixed" by the programmer.

This should be a "sticky" and labeled as such.

Yes Sir, totally agree with you, there are 2 reservoirs but wrong calculated (too small to feed system when pump has to be powered off)

 

[HJTRBO]

Have a search for cross limiting control. usually it is applied to boilers, i.e air / fuel but for your application it could be very applicable

 

[sparksy@gage]

Install a Sampson valve after the pump and control flow via 4-20 in proportion to the flow.
We use same setup to control trade waste water to the sewer. We are limited by authorities to 6L / sec.

 

[Gene Bond]

If you put a limit on KW, the motor would limit output, thus the pressure would drop at full flow. Will this satisfy the customer?

 

[one2go]

Install a Sampson valve after the pump and control flow via 4-20 in proportion to the flow.
We use same setup to control trade waste water to the sewer. We are limited by authorities to 6L / sec.

Yes I know, but do you need to keep pressure at set point or pressure doesnt matter? I guess you dont need to bother with pressure

 

[mort81]

I agree with the valve. To prevent the two loops from fighting each other, I would run the psi loop in a step control and the flow with the valve in a PID control. This will need to be field tested and tuned for both, but should give you the desired effect without running the pump up and down all the time. Depending on how the loops respond, you might get away with both in a PID, but again typically the loops fight each other when done that way.

 

[one2go]

yes that might be a good idea, will think about that and probably go infront of customer

thanks for help!

 

[one2go]

another thing that might be problem is actuator's acting speed since they have 4 big machines that uses cold water from ground to cool thier products. Time or how many macines will be powered on at same time is unknown and as I said cannot be predicted or calculated while they have other consumers too. Consumption can be 5 sec aprox. 15l/s and after that can go over 25l/s. Its kinda fast system so I would need fast actuator for my valve.. hm hm

Now I have to check pipe diameter and how fast actuator can open/close valve.

 

[Saffa]

Throttling with a valve just wastes energy in your scenario unless the pump is running off the end of its curve due to the high flow/ low pressure scenario. As others have said you can't bend the laws of physics... If they try and draw more than the allowed flow then the pressure must drop.

If these high flow demands are quite short duration you may be able to smooth things out with a large pressure vessel / accumulator on the discharge of the pump. These have an internal air bladder that is compressed as you fill them with water. Only good for very short periods of high flow though.

For longer duration they could install a plastic water tank and fill that at a regulated rate from their utility supply via control valve or ballcock / manual throttle. The pump then draws from the tank at whatever rate is required to maintain pressure.

To size the tank you need to know peak flow rate and duration. The tank needs to be able to store at least that volume without dropping below whatever level is needed by the pump to meet its NPSH requirements.

Looking at your original post it sounds like they have tanks. Where do these sit in the system?