Lead-lag vfd pump with overlap?

Two identical sized vertical turbine pumps in a water pit. Both on VFDs.

My plan is to use 1 control loop targeting water pressure. The control loop will scale 0-100%.

I was planning to scale the lead/lag 0-100% command the following way:
lead pump 0-100 = CV 0-60
lag pump 0-100 = CV 40-100

Basic equations
lead pump command = (1.667 * CV)
lag pump command = (1.667 * (CV - 40))

My thought was to have some overlap so that when the lead pump is at 100%, the lag pump is spooled up to some speed that the output should be meaningful. I thought that if the lead/lag was based on 50% of the CV, there would be more delay waiting for the lag pump to actually start pumping meaning flow.

Thoughts?

That makes sense to me, seems like it will work. The only thing you might want to add is some sort of hysteresis for the 2nd pump so that it doesn't constantly turn on and off if the CV bounces above and below 40%.

Your idea is good, that how it's done. It's called split range output.

The only thing is that PID loops are designed for controlling symmetrical and linear processes. You should arrange your output so that a 10% increase will give the same increase in flow/pressure regardless if you go from 0-10%, from 35-45% from 50-60% etc.


So you should run the pumps manually at different outputs and check the flow/pressure or whatever you are controlling to see how much each pump contributes to the flow. Then calculate backwards and decide at what percentages each pump should start.

Perhaps there are also data available from the pumps on how much flow you get at different speeds. Someone has at some point decided on what size pumps and how they should be connected, what diameter and length of piping etc. You should be able to calculate how and when the both pumps should run get get your system linear.

Read this for instance:
https://chemicaljobs.files.wordpress.com/2012/09/pumps-in-parallel-and-series.pdf

g.mccormick said:

Two identical sized vertical turbine pumps in a water pit. Both on VFDs.

My plan is to use 1 control loop targeting water pressure. The control loop will scale 0-100%.

I was planning to scale the lead/lag 0-100% command the following way:
lead pump 0-100 = CV 0-60
lag pump 0-100 = CV 40-100

Basic equations
lead pump command = (1.667 * CV)
lag pump command = (1.667 * (CV - 40))

My thought was to have some overlap so that when the lead pump is at 100%, the lag pump is spooled up to some speed that the output should be meaningful. I thought that if the lead/lag was based on 50% of the CV, there would be more delay waiting for the lag pump to actually start pumping meaning flow.

Thoughts?

Click to expand...

I have almost the exact same set up at me plant. Two identical pumps running on VFDs controlled by water pressure. But we have a different control strategy.

We have one PID control loop, that controls both lead and lag pumps. We always have one pump running. When the pressure gets below the lag pump turn on of 85 psi, the lag pump turns on. The problem is at 85 psi the PID is putting out a speed command of 100% for both the lead and lag pumps. The lag pump runs at 100% until the high pressure set point of 110 psi is hit. Then both pumps shut off.

I am in maintenance and am taking classes to learn how to program. However, the class I am taking does not cover PIDs.

What does the "lead pump 0-100 = CV 0-60" mean in plain english.

Where does the 1.6667 come from?

Thank you

g.mccormick said:

Two identical sized vertical turbine pumps in a water pit...

Thoughts?

Click to expand...

I would think that both pumps should be operated at the same speed during 2-pump operation from both an efficiency and pump-health perspective. I would try to contact an 'application specialist' or someone in 'technical marketing' from the pump manufacturer to discuss what your application is, what safeguards to incorporate (such as minimum speed), and what operating regimes to avoid.

If both are to run at the same speed, some transition strategy would be required: when to go from one to two pump operation, and when to go from two to one. To avoid Saturn_Europa1's dilemma, one might choose a strategy such as this:
When adding the second pump: The primary pump's VFD operates normally to control pressure. The second pump is started at minimum speed and then ramped up at a reasonable rate (fairly quickly but not so fast as to disturb the pressure control too much); once the speed of second pump nearly matches the first, both pumps operate at the same speed (speed determined by the pressure controller)
When stopping the second pump: The primary pump continues to control pressure. The second pump's VFD is slowly ramped down to minimum and then stopped.

Let us know what happens.

Sorry for not answering back to this. I've been busy trying to get everything up and running. Hopefully tomorrow morning I will be able to actually turn on the pumps. Attached is the pump curve for the two pumps. I believe that target pressure is going to be 60-65psi.

I developed a 3 pump air/gas blower system in Lead/Lag and fallback configuration for a city.
I had PumpB running at constant speed and PumpA was variable speed/PID controlled.
Initially off, PumpB had preset speeds of 30%, 60%, and 100%.
The load on PumpA triggered the next preset speed of PumpB.
For example if PumpA reached over 70% for a certain amount of time then PumpB would kick in at 30%. (85% would start PumpB immediately)

Once this was in place I mapped the control and speed signals to whichever physical pump I wanted to be PumpA or PumpB. It was easy to do wear balancing, failure fallback, and however else you wanted to select which pumps ran.

This worked out very well. The system had to deal with some aggressive demand spikes and lulls and it handled them nicely.

The logic I am developing and I think this will work:
1. Control loop (PI control) 0-100% targeting pressure.
2. The lead pump will scale 0-100% of its speed to 0-60% of the Control loop CV.
3. The lag pump will scal 0-100% of its speed to 40-60% of the control loop CV.
4. The lag pump minimum speed will be set to 10% when Control loop CV goes over 40%. the lag pump minimum will stay at 10% until CV goes below 30%. My thought here is that when control loop first goes abome 40%, the lag pump will come on and as the CV integrates down, taking the lead pump with it, the lag pump will stay at steady state. This should hopefully ease the transition from 1pump to 2

Saturn_Europa1 said:

What does the "lead pump 0-100 = CV 0-60" mean in plain english.

Where does the 1.6667 come from?

Thank you

Click to expand...

In plain english, "lead pump 0-100 = CV 0-60" means that the lead pump will be runnng 100% (60hz) when the control loop is at 60%, the lead pump will be at 50%(30hz) when the control loop output is at 30%.

The 1.667 comes from 100/60. The span of the lead pump 0-100% is scaled as 1.667 X span of control loop.

Look at the pump curves....are the pumps on a common header? What is the minimum pressure required on the header to get flow?

Well the two pumps control is working well. I did put some logic in place to set the min for the lag pump speed to 20% once the control loop gets to 40% and stay there until the control loop integrates below 30%. This was after noticing that the pumps do not really move any water (and the brakes are still ticking) until the speed gets to about 15hz.


We have had both of the pumps trip on overcurrent. The nameplate data says 37amps, and actually 60hz the pumps are pulling about 38.8amps. To get around this, I am scaling my output so that the vfd runs max of about 57hz which puts the motors pulling about 35amps.


Something to ask/think about: For the staging of the lead/lag pump what does everyone feel is better scenario.

Scenario 1.
A is lead, B is lag until:
A has 500hrs
B has 350 hrs.
Then B is lead, A is lag until:
B has 500hrs
A has 600 hrs
Then A is lead, B is lag until:
A has 1000 hrs
B has 850 hrs
etc, etc, etc.

Scenario 2:
A is lead, B is lag until:
A has 500hrs
B has 350 hrs.
Then B is lead, A is lag until:
B has 850hrs
A has 700 hrs
Then A is lead, B is lag until:
A has 1200hrs
B has 1100hrs
etc etc.

(note hours numbers are just examples made up on the fly)

I'm thinking leap frogging (scenario 2) is the better approach. This way the pumps will have the higher possible delta in run time. If/when one pump dies, the other pump should have some more usable life in it.

g.mccormick said:

We have had both of the pumps trip on overcurrent. The nameplate data says 37amps, and actually 60hz the pumps are pulling about 38.8amps.

Click to expand...

What is the service factor (S.F.) of the motors?

cow said:

what is the service factor (s.f.) of the motors?

Click to expand...

1.15

Then you should be OK to let them run at 60Hz. Just set the trip/OL current setting to FLA * S.F.

As far as your lead/lag switchover, I think the best option is to have both pumps wear as evenly as possible. It's not like they're going to fail at the exact same time anyways. And they're unlikely to fail at all really, more likely to just lose efficiency gradually. They should both be replaced as a pair when you see them consistently needing to run both near full speed.

37 FLA x 1.15 S.F. = 42.5 amps

It's not an issue running the motors at 38.8 amps.