Standard VFD and transformer sizing practices?

[paul_arc]

What I have here is one of our 100 hp VFD's went out. Currently it is powering a 108 hp motor. I would like to get a larger VFD for this configuration but it is on a 125kva transformer.

My question here is what is the standard practice in sizing the two? I'm sure there is nothing wrong with having a larger transformer than the VFD, but is there anything wrong with having a larger VFD than transformer? I am thinking that I can just set the overloads on the VFD at around 150A but always think that the next guy in there can raise it up some time down the road for what ever reason and blow the transformer if the pump and motor gets stuck.

So is it standard to just size the transformer larger than the drive?

Here is how the power goes:

480V MCC > 100 hp VFD > 125kva isolation step up transformer > 108 hp, 1775v, 39A motor.

Transformer ratio 1910/480= 3.98* 39A= 155A@ 480v* 120%= 186A

 

[Ingenieur]

Weird voltages
motor 1775
xfmr 1910

For this application I would size the xfmr based on vfd output v and i
Assume 125 A at 480 vac, kva = 104 kva
a 125 kva should be fine

what is the actual motor voltage
assume 3.98 x 480 ~ 1910
kva at 39 A = 129 kva
but since v is higher than rated i should be lower
xfmr is marginal

 

[paul_arc]

Weird voltages
motor 1775
xfmr 1910

For this application I would size the xfmr based on vfd output v and i
Assume 125 A at 480 vac, kva = 104 kva
a 125 kva should be fine

what is the actual motor voltage
assume 3.98 x 480 ~ 1910
kva at 39 A = 129 kva
but since v is higher than rated i should be lower
xfmr is marginal

The motor voltage is 1775v, but I don't know what the actual voltage is by the time it reaches the motor since it isn't accessible.

We plan to keep the existing 125kva xfmr, at least until it fails. And also keep the motor but I would like a larger drive. The question i'm wondering is should the drive always be smaller than the transformer?

 

[Jraef]

1775V? Never heard of it. I'm more inclined to believe that 1775 is the RPM, the transformer is likely just a 1:1 Delta-Wye drive isolation transformer because the 480V source is Delta and the VFD mfr required Wye. But hey, I'm not there to look at it, I'm just going by the fact that 1775V would be classified as Medium Voltage and would require a MUCH different VFD that would NOT be anywhere near 'standard" in any way, shape or form.

A 100HP 125A output VFD will have an input of around 97.6kVA (pulling this off of an Allen Bradley VFD chart). So if that transformer is used ONLY for that drive, it's fine. Sizing a transformer that close to the VFD size WILL increase the harmonic distortion quite a bit though, so hopefully that transformer is not feeding anything else.

Unless... the transformer is on the OUTPUT side of the VFD and the motor is on something like a deep sea submersible robot, using a custom wound motor because of the extreme voltage drop on a miles-long cable. In that case, that transformer may be too small.

Oh wait, I just read your post again, that's exactly what you said. Sorry...
The transformer will end up absorbing a lot of the output harmonics and if operated at less than design frequency, will experience added eddy current heating in the core. The usualy practice is to either design the transformer with a special core, or over size it by 50%.

 

[paul_arc]

1775V? Never heard of it. I'm more inclined to believe that 1775 is the RPM, the transformer is likely just a 1:1 Delta-Wye drive isolation transformer because the 480V source is Delta and the VFD mfr required Wye. But hey, I'm not there to look at it, I'm just going by the fact that 1775V would be classified as Medium Voltage and would require a MUCH different VFD that would NOT be anywhere near 'standard" in any way, shape or form.

A 100HP 125A output VFD will have an input of around 97.6kVA (pulling this off of an Allen Bradley VFD chart). So if that transformer is used ONLY for that drive, it's fine. Sizing a transformer that close to the VFD size WILL increase the harmonic distortion quite a bit though, so hopefully that transformer is not feeding anything else.

It is a step up isolation transformer. The motor is 1775v, 3470 RPM.
The vfd is only 480v, which powers the step up xfmr, which powers the motor.

The transformer is only used to power just this 1 motor, nothing else. I know the motor and transformer combination isn't ideal but its what I have to work with on this one.

 

[paul_arc]

Unless... the transformer is on the OUTPUT side of the VFD and the motor is on something like a deep sea submersible robot, using a custom wound motor because of the extreme voltage drop on a miles-long cable. In that case, that transformer may be too small.

Oh wait, I just read your post again, that's exactly what you said. Sorry...
The transformer will end up absorbing a lot of the output harmonics and if operated at less than design frequency, will experience added eddy current heating in the core. The usualy practice is to either design the transformer with a special core, or over size it by 50%.

Yep, the motor is under the ocean floor with a pump on it.

Over size the transformer by 50% of the motor is standard? What about sizing the drive? Just keep it smaller than the transformer does anyone go larger drive than transformer?

 

[Ingenieur]

Is the motor actually operating at 100% load?
often it is less, 80-90%

how is the speed modulated?
continuously by a pid?
constant speed?
etc

 

[Besoeker]

What I have here is one of our 100 hp VFD's went out. Currently it is powering a 108 hp motor. I would like to get a larger VFD for this configuration but it is on a 125kva transformer.

My question here is what is the standard practice in sizing the two? I'm sure there is nothing wrong with having a larger transformer than the VFD, but is there anything wrong with having a larger VFD than transformer? I am thinking that I can just set the overloads on the VFD at around 150A but always think that the next guy in there can raise it up some time down the road for what ever reason and blow the transformer if the pump and motor gets stuck.

So is it standard to just size the transformer larger than the drive?

Here is how the power goes:

480V MCC > 100 hp VFD > 125kva isolation step up transformer > 108 hp, 1775v, 39A motor.

Transformer ratio 1910/480= 3.98* 39A= 155A@ 480v* 120%= 186A

A couple of points if I may?
You can't use the motor current to calculate the VSD input kVA. The input power factor is usually significantly higher than the motor power factor.
Powering a step up transformer from a VFD needs special considerations. Any asymmetrical component in the drive PWM output can cause transformer saturation.
I have done a few VFDs with a transformer on the output but they were custom built with an air gap in the core. Just be aware that this is not for the faint hearted.

 

[Jraef]

Yep, the motor is under the ocean floor with a pump on it.

Over size the transformer by 50% of the motor is standard? What about sizing the drive? Just keep it smaller than the transformer does anyone go larger drive than transformer?

Over sizing by 50% is a "poor man's" work around to the problem of the complexity of running a motor through a transformer on the load side of a VFD rather than spend the money (and lead time) to have a custom transformer built to handle it properly. Besoeker's mention of building it with a special air gap is a common solution, I've seen others, but none of them cheap or fast.

The problem from the VFD's standpoint is that the drive must be sized not only for the motor, but also for the losses between the drive and the load, which will now include the transformer and cable losses. So "fudging it" by over sizing the transformer to handle the added heat means paying for it in the VFD sizing anyway.

In your case you already have this transformer and for all we know, it was custom made for this application, I mean it's not as though a bizarre ratio like that exists off the shelf somewhere anyway. So assuming someone knew what they were doing with that, and assuming that part of it was working fine, to my mind all you need do is increase the size of the VFD. The fact that the first one eventually failed was possibly related to it being constantly operated at maximum capacity. Over sizing would be almost imperative in this situation. Over sizing a VFD for an application hurts nothing but the budget, up to a point. Many VFDs will not like being connected to a load smaller than 50% of its rating just because of sensor range tolerance issues. So I would not go larger than a 200HP drive. But how big, I don't know.

If you have any empirical data on how the last drive was operating, than might be helpful in deciding, but if it was in a constant state of current limit at its maximum output, you may not know what it could have seen if it were possible to operate at full speed. But one possible indirect (albeit crude) way of estimating would be that if for example it was in a constant state of current limit at 125A, and to achieve that the output frequency was limited to, let's say 48Hz, that's 80% speed. If it's a simpler constant torque load, like it would be with a positive displacement pump, you can simply see it as a 125% increase (the inverse of 80%) in power (kW) to get 100% speed from it. If the load is a centrifugal (aka quadratic) pump it gets a little more complicated as to the load at 100% speed because the power requirement will increase at the cube of the speed change, so 195% power would be inferred (125% cubed). Then from whatever power increase you see, to get the drive output current required you will have to know (or reasonably guess) the power factor of the motor at those speeds.

Yours is not a simple challenge to be sure...

 

[mike_kilroy]

I agree that you can put in a larger VFD so it does not have to work as hard and should then last longer than the last one did (how long DID it last?).

We typically use an AUTO transformer rather than iso - get same output voltage conversion, added L filtering, and non of the additional transformer heating due to running at lower than its design frequency; but in our uses, we typically need to run from about 10hz to 100hz. Use them for both increase and decrease output voltages for various reasons. Not as common as it used to be though. All were designed for the application, or oversized if standard transformers as Jraef stated. Easy analogy for oversizing: is you would not expect a 60hz transformer to run loaded at 50hz, let alone the added harmonic load.

Since yours is a pump, I agree you probably run at near base speed; and as commented previously, the transformer still works fine. I doubt you would run a new 25-50% oversized VFD at higher load than present. If it made you feel better, slide thermal switches between core and each coil, put them in series, and feed to your control or VFD to shut down if it gets too hot. Then just move on to the next project.

 

[Besoeker]

We typically use an AUTO transformer rather than iso

I've done that too but, I think that in this case, the ratio is too great to make that a practical proposition.

 

[mike_kilroy]

I've done that too but, I think that in this case, the ratio is too great to make that a practical proposition.

True, the autofmr would not save any money as typical compared to iso xfmr, but we don't do it for that reason; we use auto on output so it works just as well at 10hz as 60 or 100hz to pass thru the VFD output.

 

[Besoeker]

True, the autofmr would not save any money as typical compared to iso xfmr, but we don't do it for that reason; we use auto on output so it works just as well at 10hz as 60 or 100hz to pass thru the VFD output.

A double wound will do that too. It's a matter of a constant V/Hz ratio.

 

[Ingenieur]

How long has it been operating in its current configuration?
has the xfmr temp ever been measured?
drive output i measured?
xfmr output i measured?
drive input i measured?
xfmr config: y:y, d:d, etc? pu Z?

for all we know this set-up may run to an expected service time without issues

 

[paul_arc]

Over sizing by 50% is a "poor man's" work around to the problem of the complexity of running a motor through a transformer on the load side of a VFD rather than spend the money (and lead time) to have a custom transformer built to handle it properly. Besoeker's mention of building it with a special air gap is a common solution, I've seen others, but none of them cheap or fast.

The problem from the VFD's standpoint is that the drive must be sized not only for the motor, but also for the losses between the drive and the load, which will now include the transformer and cable losses. So "fudging it" by over sizing the transformer to handle the added heat means paying for it in the VFD sizing anyway.

In your case you already have this transformer and for all we know, it was custom made for this application, I mean it's not as though a bizarre ratio like that exists off the shelf somewhere anyway. So assuming someone knew what they were doing with that, and assuming that part of it was working fine, to my mind all you need do is increase the size of the VFD. The fact that the first one eventually failed was possibly related to it being constantly operated at maximum capacity. Over sizing would be almost imperative in this situation. Over sizing a VFD for an application hurts nothing but the budget, up to a point. Many VFDs will not like being connected to a load smaller than 50% of its rating just because of sensor range tolerance issues. So I would not go larger than a 200HP drive. But how big, I don't know.

If you have any empirical data on how the last drive was operating, than might be helpful in deciding, but if it was in a constant state of current limit at its maximum output, you may not know what it could have seen if it were possible to operate at full speed. But one possible indirect (albeit crude) way of estimating would be that if for example it was in a constant state of current limit at 125A, and to achieve that the output frequency was limited to, let's say 48Hz, that's 80% speed. If it's a simpler constant torque load, like it would be with a positive displacement pump, you can simply see it as a 125% increase (the inverse of 80%) in power (kW) to get 100% speed from it. If the load is a centrifugal (aka quadratic) pump it gets a little more complicated as to the load at 100% speed because the power requirement will increase at the cube of the speed change, so 195% power would be inferred (125% cubed). Then from whatever power increase you see, to get the drive output current required you will have to know (or reasonably guess) the power factor of the motor at those speeds.

Yours is not a simple challenge to be sure...

The transformer is a K-rated transformer. The last running data before the drive went out was the motor running at 51hz, 140A. So the transformer and drive are both pretty much maxed out.
It runs at a constant speed, doesn't go up or down automatically, only manually.

I can go with a 125hp drive to be a close match to the transformer but if it doesn't hurt anything, then i'd rather go with a 150hp drive.

It looks like this 100hp drive and transformer was installed in 2012. So a 5 year run time pretty much maxed out before having any problems.

 

[paul_arc]

How long has it been operating in its current configuration?
has the xfmr temp ever been measured?
drive output i measured?
xfmr output i measured?
drive input i measured?
xfmr config: y:y, d:d, etc? pu Z?

for all we know this set-up may run to an expected service time without issues

The only data I can find on this set up is that its been going since 2012 trouble free and running at 51hz, 140A @ 480v. So not that much data but some at least.

 

[paul_arc]

I agree that you can put in a larger VFD so it does not have to work as hard and should then last longer than the last one did (how long DID it last?).

We typically use an AUTO transformer rather than iso - get same output voltage conversion, added L filtering, and non of the additional transformer heating due to running at lower than its design frequency; but in our uses, we typically need to run from about 10hz to 100hz. Use them for both increase and decrease output voltages for various reasons. Not as common as it used to be though. All were designed for the application, or oversized if standard transformers as Jraef stated. Easy analogy for oversizing: is you would not expect a 60hz transformer to run loaded at 50hz, let alone the added harmonic load.

Since yours is a pump, I agree you probably run at near base speed; and as commented previously, the transformer still works fine. I doubt you would run a new 25-50% oversized VFD at higher load than present. If it made you feel better, slide thermal switches between core and each coil, put them in series, and feed to your control or VFD to shut down if it gets too hot. Then just move on to the next project.

That is a good idea about the thermal switch, but unfortunately they would rather the equipment burn up than causing a shutdown because of heat. There are however pressure switches installed to shut the drive down if there is an unsafe condition down the hole.

 

[Ingenieur]

The only data I can find on this set up is that its been going since 2012 trouble free and running at 51hz, 140A @ 480v. So not that much data but some at least.

you are at the ragged edge
upsizing would be prudent
the cost delta 125 hp to 150 is small (imo depending on i out rating 125 is ok)
the xfmr will the real cost, if designed for this app 150 kva should do it
if designed for this application the existing 125 kva may be ok
125000/(480 x sqrt3) = 150 A > 140 A, tight but ok
measure the xfmr operaring temp

 

[Besoeker]

the xfmr will the real cost, if designed for this app 150 kva should do it

The existing transformer survived unless I'm mistaken. And 125kVA would seem more than adequate.
The motor is 100hp/75kW. Again, it survived unless I'm mistaken.
So, I don't see any justification for upgrading either.

The drive worked for five years before it failed (from what cause we don't know) so one might assume it was fit for purpose.
So again, I don't see any need for an increase in rating.

Unless something has changed on the process side that we don't know about - then all bets are off.

 

[Ingenieur]

I agree that you can put in a larger VFD so it does not have to work as hard and should then last longer than the last one did (how long DID it last?).

We typically use an AUTO transformer rather than iso - get same output voltage conversion, added L filtering, and non of the additional transformer heating due to running at lower than its design frequency; but in our uses, we typically need to run from about 10hz to 100hz. Use them for both increase and decrease output voltages for various reasons. Not as common as it used to be though. All were designed for the application, or oversized if standard transformers as Jraef stated. Easy analogy for oversizing: is you would not expect a 60hz transformer to run loaded at 50hz, let alone the added harmonic load.

Since yours is a pump, I agree you probably run at near base speed; and as commented previously, the transformer still works fine. I doubt you would run a new 25-50% oversized VFD at higher load than present. If it made you feel better, slide thermal switches between core and each coil, put them in series, and feed to your control or VFD to shut down if it gets too hot. Then just move on to the next project.

......