Switching VFD 3

[BRIS]

We have a potable water pumping station which will be equipped with 13 identical pumps. For the majority of the time only 3 of the pumps operate. The remaining pumps are for emergency transfer of water. The 3 pumps used in normal operation are required to meet variable flow conditions and will have VFD. In emergency conditions all 13 pumps must operate at maximum flow and hence do not need to have VFD.

For O&M reasons it is intended that the 3 pumps are selected from the 13 installed pumps on rotation.

My question is do we need to install 13 VFDs. Must each pumps set have its own dedicated VFD or can we install only 3 VFDs and connect the VFDs to any 3 of the 13 pumps.

 

[BigInch]

With so many pumps a possibility, I'd think that you could easily develop a configuration where no vfd are required to meet any flowrate you could possibly imagine and do so very efficiently as well.

Independent events are seldomly independent.

 

[Artisi]

I'm with BigInch on this - for me VFD's are over stated and another expensive operating problem (in the majority of cases). I would carefully analyse operating 1, 2, 3 -- 13 as individual units to meet any change to the requirement, this could be based on demand monitored by flow or pressure to bring each pumps on line as demand dictates.


It is a capital mistake to theorise before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts. (Sherlock Holmes - A Scandal in Bohemia.)

 

[BRIS]

I don't have this option. Basically I have two completely different pumping systems one is pumping directly into the water distribution system and is operating on variable speed to limit heads under and maintain best efficiency under variable daily flows. The other system is a mass transfer system that transfers water to another region in an emergency. This system can use fixed speed pumps. We have configured the hydraulic system so that both systems are using the same pump. To meet normal conditions we have two or three pumps operating on VFD pumping into distribution and 1 or 2 pumps on fixed speed pumping into the transfer system (these are operating mainly to maintain water quality and to keep the transmission pipeline pressurised).

In emergency conditions we may run up to 5 pumps at full speed into the distribution system or 10 pumps at full speed into the transmission system. In my worst case scenario I have 13 pumps operating with 10 at full speed into transmission and at variable 3 into distribution. I have many other scenarios. The client insists that the distribution pumps operate at variable speed. Thus I need 3 pumps with VFD into distribution and 10 fixed speed pumps into transfer. Because the transfer pumps operate very rarely (may be 2 to 3 days per year - they are there for a catastrophic event) whilst the distribution pumps operate 24/7 we wish to utilise all 13 pumps on rotation on the distribution system. What I do not want to do is to provide 13 VFDs when only 3 pumps ever need to work on VFD at any one time. The hydraulics are good it is the electrical end that I have problems with.

 

[bimr]

When a system includes multiple motors, it is the usual practise to provide a dedicated VFD for each motor. A motor starter is typically dedicated to a pump. You do not mention the size of the pumps, but it is not usually feasible to switch VFD's and motor starters.

If it is necessary to rotate the operation of the 13 pumps, then you can equip each VFD with a bypass. A bypass allows a motor to be run directly from the incoming line, bypassing the VFD. Have an electrical engineer evaluate the cost for this option.

 

[stanier]

Do away with the VFDs and use flow control valves

“The beautiful thing about learning is that no one can take it away from you.”
---B.B. King

http://waterhammer.hopout.com.au/

 

[Artisi]

From the description now given, you might be better to tabulate the numerous operating conditions and switching sequences and post in the electrical engineering forum - it sounds to me like a fairly major switchboard / logic system needed.

It is a capital mistake to theorise before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts. (Sherlock Holmes - A Scandal in Bohemia.)

 

[BRIS]

I agree with Stanier, and that was my recommendation, unfortunately it did not convince the client. I also agree with Artisi, I should have posted in the electrical forum. I posted on here because you guys talk language that I understand.

Thanks for the responses, I will let you know if we come up with any interesting solutions. By the way the motors are each rated at 2 MW.

 

[Artisi]

2MW, now you tell us.

It is a capital mistake to theorise before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts. (Sherlock Holmes - A Scandal in Bohemia.)

 

[bimr]

A good thing that your client did not listen to you or the pundits above for that matter. There is no single control strategy that is optimal for all pumping systems.

One generic best practice is to minimize the use of valve throttling losses in system control. Throttled valves convert hydraulic energy that the pump has imparted to the fluid into frictional heat, thus wasting a portion of the pump's energy.

In addition, most power companies will not allow a motor of that size to start across the line. A soft start of some kind is usually required because of the large starting power demand of an on/off motor on the power grid. A VFD by the way is considered to be a soft start. It might also be noted that the transient protection considerations would need to factor into the nature of the system, including the static head as well as the general piping distribution.

For a 2 MW motor, you may be talking about a power cost of over $500,000 per year. Energy usage is a critical factor in determining the lifecycle costs of pumps and their motors. Energy accounts for over 85% of the ownership cost for the pump and motor.

There is no single control strategy that is optimal for all pumping systems.

Here are links for further information:

http://w3.usa.siemens.com/us/intern...erfect Harmony Air-Cooled LABR-00400-1008.pdf

http://www.ciras.iastate.edu/publications/EnergyBP-ChemicalIndustry/Sourcebook_Chapter7.pdf

However, there are many systems for which the choice is not so clear or in which two or more different control schemes would work equally well. And there are some systems that merit a combination of controls, such as multiple parallel pumps with adjustable speed drives for each pump.

Each system must be evaluated on its own terms. The nature of the system curve, the
performance characteristics of the installed pumps, the nature of the load variability, and other factors influence the decision process.

Note that you have not provided information on the system head. Understanding of the system head requirements is also a critical best practice. Note that it is not an equipment choice or simple rule of thumb. It is recognized that knowledge and understanding are essential to proper equipment selection.

You need to contract with an electrical engineer with experience in control systems and medium voltage switchgear. This should be part of the project since the switchgear will be valued at hundreds of thousands of dollars.

 

[BigInch]

Fine, however with VFD, you can toss at least 5% of that power cost out the window before you even get started.

Independent events are seldomly independent.

 

[ScottyUK]

An MV VFD should be better than 95% efficient, even if it has a matching transformer. 98% is claimed by some of the big manufacturers, although that is probably peak efficiency and is not across the full range.

Adding an MV drive is not a trivial task. Usually the motor will be designed for inverter duty, which basically means reinforced insulation in vulnerable areas. Adding one to an existing drive is possible, but the additional protection required for the motor will knock the efficiency down.

In terms of switching you are looking at a fairly complex arrangement to say the least. I would expect a switching scheme of this complexity to be interlocked via Castell keys or similar. Suggest you post a link to this thread in forum237 and see what the opinions are from a couple of the guys there: two or three of them work for the major drives manufacturers, so their input would be worthwhile. I must say that my instinct is to dedicate a small number of drives to VFD control if your client insists and to use automation via a PLC / DCS / SCADA to equalise usage on the remaining drives. The switching plan as-proposed sounds like an operational nightmare.

Or just use an FCV on the pump discharge.

 

[bimr]

As stated above, installing a control valve may or may not be a better alternative to other control methods, such as using adjustable speed drives.

Without information on the head requirements, proposed pumping equipment, equipment run times, utility reuirements, etc., nothing can be evaluated. It is not that simple.

 

[stanier]

Bimr,

A control valve will have a pressure drop but it could well push the operating point of the pump to its BEP.

The VFD has a great deal of energy losses not only including that in the VFD itself. This has been covered in many postings on this site. Refer attached. Also check the notes in the presentation and download the reference papers.

This whole topic is way beyond a simple question posted here. A design review by specialists is required. As BIMR has said the final design is dependent upon a number of issues not annunciated in the OP.

“The beautiful thing about learning is that no one can take it away from you.”
---B.B. King

http://waterhammer.hopout.com.au/

 

[Artisi]

"A good thing that your client did not listen to you or the pundits above for that matter." If they had listened they might be a very worried client by now and seeking alternate arrangements.

It is a capital mistake to theorise before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts. (Sherlock Holmes - A Scandal in Bohemia.)

 

[itdepends]

Why not have 4 pumps with VSD (3 online + 1 rotating spare) and make the others fixed speed. For maintenance purposes you should be able to get away with 1 spare for 3 online pumps.

As a chem eng/metallurgist the first part of any answer I give starts with "It Depends"

 

[BigInch]

I think itdepends is on to it. I'm sure there is a way, if there is a will.

Independent events are seldomly independent.

 

[BRIS]

In the case we 10 of the pumps are only for emergency use and may never operate. There are only capital costs, there are no energy costs to include in the equation. I am looking for a solution that minimises capital cost and allows the pumps to be operated on rotation. A major capital costs is for 13 VFD's when no more than 3 VFDs are required to be in operation as VFD at any one time. It is not a problem to demonstrate that the NPV of fixed speed pumps with FCVs is substantially less than providing VFD to all 13 pumps. However the client wants VFD on the 3 pumps that are operating 24/7 to supply into distribution. So we need to provide at least 3 VFDs. Obviously we have to start the pumps. Our power supply is from a 66KV +40MVA primary substation. I can see a possible solution of starting the first 10 pumps DOL and the last 3 pumps on VFD. If I only have 3 VFD's then I need someway of switching the VFDs and DOLs between the 10 standby and 3 duty pumps. Interesting problem but energy costs are not a factor in the equation.

 

[BRIS]

To clarify I work for the client, our internal client is the end user (operations department). Our task is to evaluate compliance of designs with the client's requirement and to ensure that they provide reliability and value for money. We have an in depth hydraulic analysis and optimisation. We cannot get away from the fact that need to invest in 10 pumps that (unless we have a Katrina situation) we never need to operate . What I have difficulty in accepting is that we also need to invest in 10 VFDs that are not required for operation and will also never operate but must remain 99.99% reliable. The alternative of a lot of complicated switching arrangements and soft starts may save some capital costs but if this is at the expense of reliability then it is not acceptable. As I said an interesting problem greatly appreciate all of the posts. .

 

[ScottyUK]

Depending on operating voltage and current you may be able to use a high voltage connector of the type made by Cavotec for the mining / rail / dockside industries to connect your VFD to the motor of choice. These connectors are far removed from your average plug & socket, being massive bronze castings and rated to about 25kV.