Parallel operation of VFDs on centrifugal pumps, 3

[GirlEE]

We're specifying a portable (moved around as we renovate 10 stations over the next 8 years) hydropneumatic pump station with 2 duty 1 standby centrifugal pumps with VFD drives. The pumping rate is proposed to be tied to the flow rate out of station. If the first pump can't keep up, we start the second pump. Do we have to program the VFD to match only half the flow signal once we turn on second pump? How is this handled? Pumps will all be one horsepower rating, which is not yet fixed but will be in the range of 15-50 HP.

 

[electricpete]

I think it's a tough question to answer.

IF your system is a "closed loop" control which senses flow to control pump speed AND IF you sense total flow downstream of both pumps, THEN you will need to leave the setpoints based on total flow.

IF your system is an "open loop" control which does not sense flow (just estimates from pump speed/characteristics) OR IF you sense individual pump flow THEN you can put your setpoint to half of the desired total flow.

 

[djs]

Hello;
In general where you have multiple variable speed pumps, you should set all running pumps to the same speed. This will distribute the flow evenly between them, prevent cavation, and ensure cooling in all pumps. This will also improve the overall efficency of the pump system.

 

[electricpete]

good point by djs. If speeds are different than one pump can be pushed into shutoff and may overheat quickly. Even if that extreme does not occur, there may be excessive vibrations from operating close to shutoff head or oscillations from control interactions.

 

[dydt]

hi..
my suggestion is to run the motors all at the
same time the rate of which is controlled by
a flow rate signal ( 4-20mA or 0-10V ) ...the drive
will automatically respond to the demand...
a "close loop" like electricpete mentioned above..

one thing you should consider too is that the
wear and tear of the pump/motor/drive combined
will be equal....

dydt

 

[busbar]

Write your specs around the desired performance of the control system. VFD reps should be able to answer the question {or you firmly request that they do,} including controller programming / interconnection diagrams.

Restating the previous point—"parallel" pump operation can be approached two ways...do you want one pump to wear out at a time, over a shorter interval, or have all wear evenly, and potentially take all off line at once for rebuild at a longer interval?

Part of the decision involves the mode of babysitting chosen for the equipment.

 

[electricpete]

One other item..... Pumps don't like to operate at low flow. Vibration increases, internal recirculation increases, and wear increases (worse than operating at full flow). Depending on expected flow patterns, running all three or even two pumps all the time may not be a good option.

 

[electricpete]

whoops, please ignore my last message. It was kind of a brain f*rt. I was thinking of fixed speed pumps.

 

[abeltio]

Usually what I've seen is to have the simplest arrangement possible.
If you have 2 pumps. select them to be each 110% capacity, the VFD will take care of the control. In eight years you'll probably not be there, moved to better, greater things...but the guys maintaining the skid will bless your heart for having full redundancy.
The control should be set so that if the lead pump is not performing it should shut itself down and the second (lag pump) should take over at full blast... do not worry yourself about running them in parallel, it will break your heart with VFD's (also see thread407-30121).
Saludos.
a.

 

[thowell]

Running variable speed pumps in parallel comes down to one issue. The head (discharge pressure) of both pumps must be the same in order for both pumps to move water through the pumps and into the discharge header. Since pump sizes are the same, many electrical engineers make the assumption that the pump curves are the same and they are not. You can get many pump curves for a particular size pump. Each pump must generate the same head as all the others. If a pump / drive combination cannot generate enough head to push its water into the header, the cavitation spoke of above will occur. For two pumps (with the same pump curves) and drive combinations that are the same, running them both at the same speed should work. It is like having a 6VDC battery in parallel with a 4 VDC battery. The 4VDC batter will not contribute any current to the circuit.

 

[jraef]

As mentioned by others, running both motors at the same output from the same flow signal is better than running one up, then adding #2 with 2 vfds. Even wear and tear is the prime benefit.

If you are going to operate as you described, better yet is to use one VFD and run #1 up to full speed, then switch it across-the-line and add-in pump #2 with the VFD as needed to maintain flow. Several VFDs on the market today have that capability built right into their software, even the control of the across-the-line starters. This is done all the time very successfully. Check it out. Subvert the dominant paradigm... Think first, then act!

 

[Macbeth]

I have extensive experience in this area from pumps as low as 2@1hp to 5@1500hp, and I agree exclusively with “thowell”. If more than one pump is running they must be all at the same speed. No exceptions exist for centrifugal pumps.

The best way to accomplish this is Parallel a 1-10vdc signal from a external PID controller to all drives.

Now for pump sequencing, many ways to go in this area. If you use small pumps then don’t worry about pump curve efficiencies. Just size the pumps and impellers so one can handle you low flow needs and two can handle your high flow needs.

However if your motors were let’s say 50hp each, you would get multiple benefits from sequencing them smartly. An example of sequencing them would be not to have a fixed order but start and stop them by runtimes, start the pump with the least amount of hours and stop the pump with the most amount of hours. Although some people will disagree with this and ague they need a fixed order to know if they are sequencing appropriately, the benefits would be even ware and tear and if the pumps shutdown for short period of time the motors will have time to cool.

Another idea would be not to consider the third pump as a spare but just another in the system that has 133% capacity. What I mean is don’t run one pump up 100% but start anther pump before you go to far past the B.E.P.(Best Efficiency Point.) However in order do this you will need to do home work on the pump curve to establish the best C.O.P. (Cost Of Performance) at various speeds vs. flow rates. If you look at the curves you will notice it is possible to have a lower C.O.P. with three pumps at partial percentage than two at 100%. Although some people will disagree with this, and ague they want a brand new spare pump waiting, when a primary fails. I know of many pump that sat so long when need they failed shortly afterwards due to things as simple as mice built their home in them. I have also decommissioned facilities with bran new 30-year old equipment.

Well That’s My $0.02

 

[rodger1]

Q: Do we have to program the VFD to match only half the flow signal once we turn on second pump?
A: No. The flow should be measured at a point common to the discharge of all the pumps... 1 setpoint (adjustable), 1 feedback. The single (adjustable) setpoint and feedback should be common to all the VFD's.

You mentioned 2 duty and 1 standby pump, all of the same HP. I would suggest that the 2 duty pumps be identical and the standby pump sized to handle the load of both the duty pumps. This would give you true backup.

A simple control scenario would be to parallel a 0-10vdc feedback (flow) signal to all the VFD's. A simple lead/lag switch could be incorporated into the control circuit to select which of the 2 duty pumps would start first. Using a programmable VFD, I prefer ABB, you could program the lead drive to turn on the lag drive based on a frequency that represented a fully loaded pump. The lag drive would then start and both drives would adjust to the common feedback signal. Inversely, as the load diminished, the lead pump could shut off the lag pump at a pre-selected frequency. In the event of duty pump failure, the standby pump could be started, OR used as the primary pumping source. This can all be accomplished without the use of external logic controllers. I would be happy to provide a list of the variables and their settings.

 

[jbartos]

Suggestion: When pumping and related piping system become more complex, then a software aid becomes appropriate. E.g.

http://www.eng-software.com/

http://www.engineering.usu.edu/bie/software.html#sprnksim

http://www.korf.co.uk/korf_links.html

etc. for more info

 

[Macbeth]

GirlEE,
I have a question. Are you going to control these by flow or pressure? It is more common to control by pressure, (constant pressure variable flow). In my area equipment needs a predetermined pressure drop to operate. As more equipment comes on line or increased demand, the lower the pressure goes and the faster the pumps pump.
If these are controled by flow and not pressure, can you elaborate on your process FMI please?

 

[jbartos]

Question: Do you have any specification on the flow rate out of the station, including the flow rate fluctuation? This could be the baseline document for set points of VFD and other associated controls.