One Drive, two pumps

[Marke]

I have become involved in an application where the end user wants to use one drive with two x 850KW 400V pumps.

The plan is to start one pump and get it to full speed and then switch it onto the supply and connect the drive to the second pump and use that to control the flow.

The pumps will probably be coupled by a long shaft.
The plan is to use open transition switching from the output of the drive to the supply. I am concerned that this will result in a high curent and torque transient and this could cause problems with the long drive shaft, plus potentially compromise the drive.

One option is to use the drive on one pump and a reduced voltage starter on the other, or to use a drive capable of shunt switching. The end user is trying to minimise the cost, but I am concerned they will compromise the reliability of the installation due to penny pinching at the design stage.

Has anyone had any expereicne with this type of application?

Best regards,

Mark Empson

http://www.lmphotonics.com

 

[DReimer]

If the one pump will always run at full speed I would recommend using a soft-start on it, then use the drive on the pump that will be speed-controlled. What you said, in other words. I don't understand how these pumps are to run at different speeds, however, if the shafts are coupled together. Or are you talking about the motor coupling?

I've been that end user before, doing project engineering work. Often you can only make a financial case for a project by cutting corners on the scope of work, or by trying to come up with schemes like the one above.

 

[rbulsara]

Sounds like right out of Rube Goldberg's (not sure about spelling his name!!!) book.

Are you the supplier? or designer? Ask the Owner to hire a good engineer, before you start helping this client.

If he insists, stay out of this project! I would love to see the one line and switching arranggement of this one.

 

[Marke]

Hello DReimer

The plan is to switch the first motor from the output of the drive onto the incomming supply so that it will run at a constant full speed, with the speed of the second motor being controlled by the drive.
I believe that the threory behind this proposal is to minimize the start current that would otherwise be required.

Best regards,

Mark Empson

http://www.lmphotonics.com

 

[Marke]

Hello rbulsara

I am working with the supplier and questioning the proposal put forward by the consulting engineers.
My preference is to use a drive on one pump and a soft starter on the other. I believe that it will work out cheaper and provide some redundancy in case of a failure.

Best regards,

Mark Empson

http://www.lmphotonics.com

 

[wschel]

Multiple pumps on a single drive was fairly back in the days drives cost an arm & a leg. The drive is set up with output contacts and interlock to switch between the 2 motors and uses closed transition transfer to accross the line after reaching full speed.

They normally would not use open transition because of flow transients (water hammer & such).

I once replaced a system like this that used current source drives that never worked because the drive output could not synchronize with line voltage.

 

[quark]

I saw couple of installations(don't ask me for electrical details) with single vfd and two pumps but the control scheme was not the one you mentioned. The prerequisite for parallel pumping is that both the pumps should work at same discharge head. The idea of supplementing the flow with one pump running at full speed and the other one at reduced speeds will result in varying discharge heads and this may lead you into trouble.

The problems will be less if 1) the system head is purely dynamic or 2) the supplementary variable speed pump is carefully chosen to match the discharge head with that of full speed pump.

The general logic will be either to run one pump at full speed or run the both pumps at some itermedeiate speeds, so that overall efficiency of the pumping system will be optimum. Danfoss has some controllers which take inputs about pump characteristics and will allow parallel pumps to run at optimum efficiency. You can also speak to ITT Fluid Systems.

I wonder why electricpete is not participating in this discussion.

Regards,

 

[mc5w]

If both pumps have the same suction and discharge pipes then trying to run them at different speeds will create more problems than it solves.

If he needs to more closely match the pumps to the load your customer or better off to use 4 smaller pumps, say six 300 KW pumps or 1 850 KW pump plus three 300 KW pumps. This would allow the use of more economical soft starters.

Also, in large pump applications of soft starters and VFDs there also has to be soft STOPPING so that check valves do not slam shut creating a water hammer that will break the check valve, break the pipes, blow gaskets, or blow seals.

 

[jraef]

Marke,
I have done this several times with different mfgrs, and very different results. This is a common application for MV drive manufacturers who can provide true closed transition transfer (Robicon is somewhat famous for that), but those are few and far between in low voltage because of the associated costs of the VFD with that capability.

I have built two LV systems at the insistance of consultants who saw this done but didn't realize the transition aspects. In one it worked fine, but it was a very low head / high flow axial pump application where the motors didn't seem to slow down when breifly disconnected. I used a macro routine for this provided by the VFD mfgr. (ABB ACS500) and it worked well, but the transition was open and I felt our success was specific to that application.

The other system was on vertical turbines and there wer catastrophic consequences on the pumps. Both shafts twisted and one coupling snapped in the third week. We solved that problem by using a soft starter on the second pump. When the flow from pump #1 was maximum with the VFD, we started the second pump with the soft starter as slowly as possible. To avoid pressure surges we tuned the PID loop of the VFD of pump #1 to decrease speed as fast as the flow of pump #2 increased (well, as close as we could get). Then as requirements increased, we brought pump 1 back in slowly as needed.

The only drawback was that we lost the ability to alternate the pumps to equalize wear and tear, but having soft starting on both pumps reduced the importance of that loss. It could have been done, but the end user didn't like the cost and complexity of the output contactor arrangement that would have been necessary to facilitate that.

"Our virtues and our failings are inseparable, like force and matter. When they separate, man is no more."
Nikola Tesla

 

[itsmoked]

jraef; Is there no black box that can hook to the line and the VFD output to put out a "synchronized" signal allowing the closure of a parallel contactor?

 

[rbulsara]

Marke:

While stil standing by my original response,

I am curious as who how the motor will be 'switched' to the utility, this will need some serious 'closed transition and "synchronized" transfer between the VFD output and the utiltiy. Is the transfer to the 'utility' manual or automatic?

And after that what protects the motor from overload? Is there a separate overload protection? or thermal protection?

Then you will need switching (auto or manual) to connect the VFD to the other motor.

More importantly, what if the drive fails or need serice? What will 'start' either of the motors?..I just can't fathom this setup. There are two motors and you need two starters. It is as simple as that.

By far cost of a soft starter would be much less than all the rat nest of switcing, controls and cabling you would have.

 

[electricpete]

good point by quark and mc5w. If these are two similar pumps in parallel with one at full spped you cannot reduce speed of the other one very much without putting it far to the left on it's curve. Not effective control and not good for the pump.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[DickDV]

Marke, this is not such an unusually situation and is handled somewhat differently than assumed above.

In a two pump configuration with common suction and discharge headers with check valves, the inverter initially operated only the first pump with the second pump being off. As water demand builds, the first pump reaches full (60Hz) speed and, at that point, closes a mag starter on the second pump. This pump proceeds up to speed rapidly and, as it picks up the flow, the first pump slows down to maintain the setpoint pressure.

The first pump now continues up in speed as demand increases until it also is at full speed. There are no sychronizing or surge problems with this arrangement.

ABB's system for this also reconfigures periodically so the second pump becomes the variable speed unit and the first pump becomes the across-the-line unit. That helps to even out the wear on the two units.

I've seen this used successfully with up to four pumps---three on mag starters and one variable speed.

I'd recommend it for your example.