Controlling Water pressure with two pumps?

[bobwithdana]

I have been tasked to maintain 80 PSI water pressure in a 6 ingh pipe using two 40 HP pumps with two Powerflex 700 VFDs in and existing Controligix PLC system.

This needs to be a lead lag system with one pump controlling until the demand reaches a certain point when the second pump is needed.

There are challenges as we dont want the second pump to run below a certain frequency.

I know I can Hack some code to work but am asking for some ideas that will be clean and easy to troubleshoot. Kicking around some ideas in my head about using a cascaded PID loop but wanted to throw it out there in case any of you have run across this before.

Thanks in advance
Bob

 

[Oakley]

Run a pump at a continuous speed ... determine where this is (where the min speed of the additional pump needs to be). Then use PID for the other pump.

 

[TheStarr]

In aircraft hydraulic systems there is a problem called "paralleling pump control"

These pumps are self-regulating and their pressure controls go crazy if they both try to maintain the exact same pressure.

There always needs to be a master and a slave. As with Oakley's description, you need to know what one pump can do, flat out.

Then use the other pump to help the master pump when needed.

To even out wear-and-tear you would want to pass the "master" role back and forth between the pumps based on run time.

You will also need to know what effects your load and how fast it reacts to change to get a good starting point on your PID.

A quick Google on "parallel pumps" brings up tons of information.

 

[Mickey]

Similar to Oakley’s and TheStarr's control is one called “inching control" ( or trim).
The first pump is controlled by a PID. When it's output gets to 80 to 85% you start the second pump and slowly ramp it up. This will cause the first pump's PID output to go down. At about 55 to 60% stop ramping the second pump and leave it running at whatever speed its at. If the first pump's PID increases again to 80 to 85% ramp up the second pump again. Continue this cycle as needed. If the PID reduces back down to 20 to 25 % then ramp down the second pump ,stop ramping down when the PID increases to 55 to 60%. Continue this cycle as needed. You decide at what speed to shut off the second pump.

 

[daba]

I second Mickey's approach as the most sensible, and very similar to a 3-pump system I worked on for a HTHW (High Temperature Hot-Water) system in a brewery. In that application the pumps simply had to pressurise a recirculating main, i.e. the pumps did not have to produce a flow, only pressure.

Oakley's "Run a pump at a continuous speed" is counter-productive, as this will not produce a constant 80 psi over the range of demands below that at which the second pump must run. Your first pump must be under PID control to meet the specification.

As an addendum to Mickey's solution, I would determine the minimum speed for the second pump, i.e. the speed at which it produces no additional pressure, and use that as the starting speed. Without it, the wait for the ramp up to that speed may be too long, as the pressure increase by pump 1 above 80-85% will be minimal.

If you can get the curves for the pumps you will get a better idea of when to cut-in and cut-out the second pump.

Also make tags for your cut-in and cut-out setpoints, not literals in the code, and any other things you may need to tweak, you don't want to be editing code to get this tuned up.

 

[DickDV]

I would suggest contacting Allen Bradley about the availability of "Pump/Fan Lead/Lag" software for these drives.

More than likely they already have it all worked out with sequential control of multiple pumps and maybe even rotation of the lead pump after a designated period of time.

I know that many VFD manufacturers have this software available.

 

[John Gaunt]

Another way to run two loops like these is to have one control loop as a pressure control (PV = Pressure) and the other loop as a control using the output of the first loop as its PV (PV = Loop1 Output %). The output of each loop controls its own pump speed.

If you try it you will be amazed at how well it works. There is no nasty interaction between the loops. Just tune the Pressure control loop a little more agressively that the its output control loop.

The response and precision of control will generally be better than that achievable with just pressure control.

 

[Mickey]

Another way to run two loops like these is to have one control loop as a pressure control (PV = Pressure) and the other loop as a control using the output of the first loop as its PV (PV = Loop1 Output %). The output of each loop controls its own pump speed.

If you try it you will be amazed at how well it works. There is no nasty interaction between the loops. Just tune the Pressure control loop a little more agressively that the its output control loop.

The response and precision of control will generally be better than that achievable with just pressure control.

What is the set-point of the second loop?

 

[Nick B]

You can go another crude method that works. Use the same output from the PID for both. Then decide on a threshold for each to come on or off. I would suggest Ramping off the drive that is cutting off slower just to allow other pump(s) to take up the slack etc. Ramp drive in PLC and bring VFD ramp rates to something small (1 secondish depending on pump size). I would build a ramping block that I can switch between two ramp rates depending on whether I'm on or coming off.

Simple and effective. Which comes off first? The one with the most operating hours. A couple of simple tests and some specifications from pump vendor will tell you what values make sense.

This method works for multiple pumps as well.

 

[John Gaunt]

Hi Mickey,
The setpoint of the second loop is the desired output of the first loop. So to give the first loop plenty of room to move the SP of the second loop could be set to 50% as a good place to start.
The second loop will adjust its output (second pump speed) to try to keep the first loop (first pump speed) at 50%.

 

[Lare]

Little bit different than Mickey's approach, but same principle however.

First pump is controlled by PID. When PID's output gets over eg. 90% for certain time, start second pump at speed 90%. At same time, when second pump is started, you have to manipulate PID's output to go minimum. (Pump1)

If PID's output goes under eg. 25% stop second pump and manipulate pid's output to start from value 90%.

Lare

 

[Tom Jenkins]

Before you start programming you had better find out the system curve and the pump characteristics. If you have less than 50% turndown on your pumps you will never get good control. You can't reliably overcome mechanical deficiencies with programming.

You need to plot the system curve and superimpose the pump curves for a single pump and two pumps in parallel. That will tell you the minimum speed and flow for each pump and the pair of pumps. It makes a big difference whether you have a high static head or friction head only.

The way I do similar applications is to have the output of the pressure control loop be a flow command. Then if you have one pump running you send the flow command as the setpoint of the pump flow control loop and pump speed as the output of the pump flow control loop. If you have two pumps running you send half the flow output of the pressure loop as the setpoint to each pump flow control loop.

Dick is right about most manufacturers having pump control logic built into their VFDs. You just need to examine their control logic closely to make sure it matches both your system needs and your hydraulics.

 

[bobwithdana]

Thanks a million

Hey guys, thanks. I have plenty of information to get started thanks to you all.

Bob