Water Distribution Pressure and Flow Control

J

Thread Starter

Jason Alsin

I have been tasked with designing a control strategy to control both the pressure and flow of a USP water loop. USP requirements demand a minimum flow rate in order to insure water quality. Use point requirements demand a minimum pressure.

Basic System Description:
A storage tank supplies the distribution loop with USP water. A VFD pump is used to pump water through the use points. A flow meter and pressure transmitter monitor the flow and pressure of the loop as the water returns to the storage tank. There is a back-pressure control valve at the end of the distribution loop just before returning into the storage tank.

If I implement pressure control using the pressure transmitter and control valve and flow control using the flow meter and the VFD, the control loops will constantly fight.
If I fix the control valve (eliminating it from the equation), I can control either pressure or flow using the VFD. But, what is the best way to implement this option?

If I use two separate PID control loops (one for pressure and one for flow) and select the greater of the two CV to run the VFD at, then one of the PIDs will not be calculating correctly (and, therefore, not control correctly) since it is not really controlling the PV.
If I use a single PID control loop and select between pressure and flow as the PV, pressure and flow have different tuning parameters, how do I tune the loop?

Has anyone implemented an effective control strategy on a water distribution system like this?
 
B

Bob Peterson

With a centrifugal pump, the pressure is whatever it is at a set speed. Maybe a more or less fixed speed is the simplest approach.

I would just set the pressure with the VFD and use the control valve to tweak the flow in the loop.

 
B

Bruce Durdle

You can quite safely operate a selective control scheme with both pressure and flow controllers. As only one of the loops will be in service at a time, tune each independently. Anti-reset windup should be a feature of almost any control hardware, and will ensure that there is minimal loop interaction at the changeover.

If it's a high-flow system, power requirements can be kept low by say using the control valve to provide the initial control action, then trimming the VSD to maintain a set valve position.
 
M
I stumbled on the below post from 2012. It is the exact same as my situation. We planned to run pumps on VFDs (currently fixed speed) with control signal from the end of loop pressure transmitter (located just before the distribution dumps into the buffer tank). The current backpressure regulator gets its control signal from the same end of loop pressure transmitter we plan to used for VFD control. So they don't fight each other we planned to replace the backpressure regulator with a fixed position/manual one. I figure once there is a high demand from one of the users the pump would ramp to full speed to try and maintain the end of loop pressure and the backpressure valve would close some degree if the pressure cannot be maintained.

Questions: Do we in fact have to revert to a non-controlled / manual back pressure regulator. And assuming we do, I would think we would need to experiment with various end of loop pressure settings in order characterize the system with the goal of using the lowest set point in order to conserve energy but not so low that we do not get adequate flow to users that have spike demands.

>I have been tasked with designing a
>control strategy to control both the
>pressure and flow of a USP water loop.
>USP requirements demand a minimum flow
>rate in order to insure water quality.
>Use point requirements demand a minimum
>pressure.
>
>Basic System Description:
>A storage tank supplies the
>distribution loop with USP water. A VFD
>pump is used to pump water through the
>use points. A flow meter and pressure
>transmitter monitor the flow and
>pressure of the loop as the water
>returns to the storage tank. There is a
>back-pressure control valve at the end
>of the distribution loop just before
>returning into the storage tank.
>
>If I implement pressure control using
>the pressure transmitter and control
>valve and flow control using the flow
>meter and the VFD, the control loops
>will constantly fight.
>If I fix the control valve (eliminating
>it from the equation), I can control
>either pressure or flow using the VFD.
>But, what is the best way to implement
>this option?
>
>If I use two separate PID control loops
>(one for pressure and one for flow) and
>select the greater of the two CV to run
>the VFD at, then one of the PIDs will
>not be calculating correctly (and,
>therefore, not control correctly) since
>it is not really controlling the PV.
>If I use a single PID control loop and
>select between pressure and flow as the
>PV, pressure and flow have different
>tuning parameters, how do I tune the
>loop?
>
>Has anyone implemented an effective
>control strategy on a water distribution
>system like this?
 
D

David W Spitzer

Mike,

My understanding is that you want to use a VFD on the supply pump for better energy efficiency and want to know if you need to control the return back-pressure.

The answer to this question could be yes or no depending on the process. For example, the purpose of the existing back-pressure valve could be to keep the piping full or to maintain cooling water pressure above process pressure, or maybe it is there for no reason at all because the return piping can be allowed to drain freely back to the tank.

I suggest that you check with the process engineer to determine the process constraints before addressing the control issues.

Best,

David W Spitzer
ISA Life Fellow
Spitzer and Boyes, LLC
 
H

Heinrich Baumann

> Has anyone implemented an effective control strategy on a water distribution system like this?

To solve this you shall forget the flow control and only use pressure control. This is the only parameter that can be influenced by the pump and the valve.

I would suggest a split range control. controller output from 0 - 50 % the valve shall control the pressure, from 50 - 100% the VFD shall control the pressure through the pump speed. (The 50% split shall be calculated by the size of the control valve compared to the pump power.)

It is important to set a high enough fixed minimum speed at the VFD which will assure the required pressure to be reached for this to work and allow the control valve some control.

Hope this helps.

Best Regards.
Heinrich Baumann
 
the most valuable equipment on your loop is the pump. you should have a vfd connected, it will protect your pump and let you better adjust the flow rate

the pressure of your loop is not uniform as it leaves the pump the pressure starts to drop, tube lenght, tees, elbows etc all affect the pressure

The flow is the most crucial element of you loop so as to respect turbulence values.

a valve pressure regulator to control the flow rate located pre/post-flow sensor would be one way to go but the pressure differenences are so low the valve would need a high degree of filtering to reduced fluctuations. and its another element to add and another future headache

I would stick with the vfd control, its simple to set and control ramp up and down, it protects the motor of your pump.

Pressure transmitters are important to monitor loop pressure and alarm functions to shut off the pump.

cip systems are critical skids that depend on critical flow elements (sprayballs etc), I never seen one controlled by a pressure transmitter.

bare in mind the automation with your tank level, loop temperature post pump and pre tank return, steam sterilization, Ozone injection, uv bypass and online usage, valve sample ports, automated valves with steam traps at all POU, rtds!! automation package control and monitoring!! you got a handful

i have seen wfi loops with pumps running at 100% with no vfd and the flow controlled by a modulating valve, dont like it, why ad more valves?
 
NOTE:

If you have sub-loops you would want to use modulative pressure control and run your pump at max.

I am currently revising an installation the there are 8 subloops cold operation with heat sanitization and 5 main supply loops. All is pressure regulated. In this case it makes more sense.
 
Flow rate and pressure are both affected by the pump speed.

The key here is the restriction which represents the end use points control valves and your own control valve.

To get a certain pressure and flow rate you need a combination of pump speed and restriction factors (restriction factor you can control by your dump valve)

If your end user enters a high use state you may begin changing the flow rate by closing your dump valve, however, it is likely that the capacity allotted to the end user far exceeds the minimum regulatory flow rate.

So..

First start closing the valve when flow rate increases beyond your minimum. I would still use a variable control for this. More about that later.

Once you reach a full closed state, or any minimum dump rate you have for practical purposes, you can only raise flow rate by increasing pump speed. This leads to your second issue which is that pressure must also be maintained. As the end user increases the flow rate by opening their control valve(s), the pressure will begin to drop. Closing your control valve will probably be to slow to correct for pressure drops when a large load is added to the system.

The pressure issue is not an issue but actually works in your favor. Simply use a PID control loop to maintain the correct pressure at all times using the variable speed capabilities of the Variable Frequency Drive running your pump. This will also have the effect of increasing flow rate when a greater load is placed on the system and decreasing your flow rate when the end user reduces consumption. Open the dump valve when the flow rate gets to within a rage of your minimum flow rate...say 125% of your minimum...more or less depending on how critical this limit is. A PID controller can also control the valve setting based on flow rate. Using a flow rate setpoint somewhat above the regulatory minimum will provide a buffer to ensure compliance as PID control can overshoot if not tuned perfectly. Above a certain flow rate, the flow controller can be tuned so that the valve is fully closed and automatically open to maintain minimum flow when necessary regardless of the pump speed at the other end.
 
F

FrancisRumsky

You should definitely use a cascaded loop for this (two PIDs running, one feeds a setpoint to the other).

Run the flow PID as your primary loop with its output directly to the drive speed. The setpoint for this loop will be based on the minimum flow setpoint but will also be affected by the pressure loop. You could multiply the flow setpoint by (1 + PressureLoopOutput) or have the pressure loop output as an adder to the flow setpoint. Either way would work.

The pressure loop will use the minimum pressure setpoint. If the minimum flow is met and the pressure is above the minimum then this loop will naturally ramp down to 0%. If the minimum flow is not meeting the pressure requirement then this loop will begin ramping up. As it ramps up it will then increase the flow setpoint, which ramps up the flow PID and the drive speed.

You want one of the PIDs to be much more reactive than the other so they don't oscillate. Make the more important of the two loops hotter.

I have done something very similar except that I had to worry about a critical maximum pressure. In my case the flow setpoint was what the operators wanted but was not critical. My pressure PID was very hot to make sure that they never overpressured the system and the flow PID was slower so the drive speeds were not bouncing around a lot. We had to put this in to work around a malfunctioning PRV. It worked like a charm.

Have fun!
 
Top