Step Up Transformer Issue

[MattSW]

Hello,

I wanted to post a question on an issue we recently saw. Unfortunately I wasn't at the site so I have to take all the information third hand but here is what I know.

We sold a 875 KVA step up transformer to a customer to put on a SWD running a 1000 hp motor and yaskawa drive. The motor is 4160 volts and 122 amps, thus the step up transformer. They added a sine wave filter to the yaskawa drive. Yaskawa was rated at 1200 amps.

This is where I am losing info though. They said they were only able to get 840 KVA out of the transformer. I'm not sure where they were measuring, if they were reading that off the drive or taking measurements, I'm not sure on that. I found a similar motor online and used the numbers from it to get that it would take 788 kilowatts to attain full load. Now, the transformer is rated somewhere around a K-4 and with 100% being electronic load, it should probably be somewhere in the k-20 range. My question is this, has anyone seen a transformer not able to take additional loading like that? I don't believe they were increasing voltage to cause saturation. I've never seen it before, usually it just runs over temp. I've never seen it before and was wondering if anyone here had seen a similar experience. Either way, I believe they need a new transformer. But I wanted to get some information from anyone that could rationally explain the failure.

 

[Ingenieur]

not the xfmr
imagine a short ckt on the sec, you would get many x rated kva
either the drive or more likely the load is limiting it
you will only draw what the load requires
can't force it higher unless you increase the mechanical load (or mess with motor eff and pf which can be considered fixed at a stable load)
95% of full load

I assume the drive and filter are on the primary (480?)

btw
if the motor started even with drive limiting you likely got 3000 kva

what does your factory performance test data show?

 

[Besoeker]

Hello,

I wanted to post a question on an issue we recently saw. Unfortunately I wasn't at the site so I have to take all the information third hand but here is what I know.

We sold a 875 KVA step up transformer to a customer to put on a SWD running a 1000 hp motor and yaskawa drive. The motor is 4160 volts and 122 amps, thus the step up transformer. They added a sine wave filter to the yaskawa drive. Yaskawa was rated at 1200 amps.

This is where I am losing info though. They said they were only able to get 840 KVA out of the transformer. I'm not sure where they were measuring, if they were reading that off the drive or taking measurements, I'm not sure on that. I found a similar motor online and used the numbers from it to get that it would take 788 kilowatts to attain full load. Now, the transformer is rated somewhere around a K-4 and with 100% being electronic load, it should probably be somewhere in the k-20 range. My question is this, has anyone seen a transformer not able to take additional loading like that? I don't believe they were increasing voltage to cause saturation. I've never seen it before, usually it just runs over temp. I've never seen it before and was wondering if anyone here had seen a similar experience. Either way, I believe they need a new transformer. But I wanted to get some information from anyone that could rationally explain the failure.

A couple of points if I may. The 875kVA is a bit marginal. And, if the motor power factor is say 0.85 that equates to over 900kVA for the motor. The drive, at 1200A and assuming it's 480V is nearly 1,000kVA. All a bit tight so I think you are right about getting a new transformer.

One more point. The last time I did this we used a gapped transformer core to avoid core saturation resulting in any asymmetry from the output PWM waveform.

 

[Ingenieur]

rated motor kva = 122 x 4160 (likely rated 4000) x sqrt3 = 879 kva
likely more like 845 using 4000 vac
agree marginal

eff x pf = (1000 x 746) / 845000 = 0.88
or (1000 x 746) / 879000 = 0.85
actually pretty good either way

 

[Besoeker]

rated motor kva = 122 x 4160 (likely rated 4000) x sqrt3 = 879 kva
likely more like 845 using 4000 vac
agree marginal

eff x pf = (1000 x 746) / 845000 = 0.88
or (1000 x 746) / 879000 = 0.85
actually pretty good either way

But it failed according to the OP. So, not so good.

 

[Ingenieur]

But it failed according to the OP. So, not so good.

the motor failed?
an eff x pf of 0.85-0.88 is good
assuming a pf of 0.9 eff is > 95%

the xfmr did not 'fail' per se
he stated he got sketchy info that its output was limited to 840 kva
he's not even sure of that though
no measurements of v or i
no location of measurements
no idea of accuracy, it is only 3-4% off
etc

afaik nothing 'failed'
what he said it appeared to not make rated output
he called that a failure, to me failure implies device destruction in this context

this could be due to the drive or mechanical load
I've never seen a large xfmr hit a wall below its rating
usually fine at 125 or even 150% based on cooling and duration
then they overheat and possible short/burn (assuming no protection)

I doubt it was on a dyno where he could increase load to draw 875 kva?
for all we know the power source could be the limitation, but I doubt it

 

[steve66]

rated motor kva = 122 x 4160 (likely rated 4000) x sqrt3 = 879 kva
likely more like 845 using 4000 vac
agree marginal

eff x pf = (1000 x 746) / 845000 = 0.88
or (1000 x 746) / 879000 = 0.85
actually pretty good either way

But wouldn't the input to the drive have a much better PF than the motor itself?

And drive has a filter on the input. I'm not sure what kind of filter it is, but that should reduce the harmonics enough that I doubt you need a K-20 xformer.

With a drive that large, it may have a nicer input circuit (what's it called again? An 18 pulse input?) so the harmonics may not be that bad to start with.

 

[Ingenieur]

But wouldn't the input to the drive have a much better PF than the motor itself?

And drive has a filter on the input. I'm not sure what kind of filter it is, but that should reduce the harmonics enough that I doubt you need a K-20 xformer.

With a drive that large, it may have a nicer input circuit (what's it called again? An 18 pulse input?) so the harmonics may not be that bad to start with.

I assume the drive is on the xfmr primary
480 filter drive xfmre 4160 motor?
yes it will have likely a higher pf, but the motors pf will determine it's power/kva

we don't have enough info
measuring point
actual motor load

I do not think there is an issue
other than the xfmr size may be marginal at 95%
but as long as its temp is within spec let her rip lol

 

[Besoeker]

afaik nothing 'failed'

Well, it seems to have failed to achieve the performance he was expecting. Not altogether surprising.

 

[Ingenieur]

Well, it seems to have failed to achieve the performance he was expecting. Not altogether surprising.

actually very surprising
xfmr 875
motor 845 (or 880) depending on rated voltage
if rated at at 4000 v and 122 A
if v is 4160 i decreases
so load ~ 845


or, more likely, the load only required 840
never saw a transformer hit a brick wall at 0.96 x rating
unless rated 0.3 or so of pu unit load, not the case here
when we KNOW it put out more since it started the motor which ran at almost full load (95%)

imo the transformer was not 'asked' to deliver more

we need more info
that is an issue around here
partial/incomplete info is given and an answer is requested for complex issues

imo no xfmr issue
it delivered sufficient kva for the motor to run under the connected load
both xfmr and motor ~95%
an 'understanding of the system' issue

 

[Sahib]

They said they were only able to get 840 KVA out of the transformer. I wanted to get some information from anyone that could rationally explain the failure.

They might have said so based on transformer temperature exceeding design full load temperature. Check it also.

 

[Jraef]

First off, I would guess that the transformer is the primary suspect. I never recommend doing this without a deep consultation with the transformer mfr, because putting a standard transformer on the output of a VFD is problematic. The gapped core, as Besoeker stated, is a common way of customizing it to avoid saturation. If you didn't account for that in the design, that transformer is likely saturating. It's the only way a transformer can limit the output to a given load. Transformers are not smart. That said, there IS a "smart" device in this situation; the VFD, and although less likely, here's how it may be the culprit.

The transformer impedance, plus the sine wave filter impedance, will drop the available voltage getting to the motor. The drive will not know this however, because the transformer and sine wave filter are on the load side of the drive; the drive cannot distinguish the transformer and filter losses from the motor current. So at any amount of the DRIVE's load current, the motor current, and thus torque, will be reduced compared to what the drive THINKS it is delivering to it. Reduced torque at the motor will result in higher slip, thus more current for the same flow, or less flow for the same current.

If the drive is then set up for a current limit (also sometimes called stall protection), it can only base that on all of the current it sees, so it will back off the output frequency and voltage to maintain that current limit. If in their calculating the transformer kVA at 100% of motor rated current, they are assuming the drive output is at 100% voltage, that may then appear, to the untrained eye, as if the transformer cannot deliver more kVA. In other words it may be a secondary affect of a perfectly normal phenomenon associated with the way they are attempting to get around using an MV drive. When using a step up transformer or a sine wave filter, you cannot expect to get full output from the motor. They have both; double whammy.

Last possibility; many VFDs now come with the factory default programming to use Sensorless Vector Control (SVC). SVC requires that the drive be tuned to the motor equivalent circuit, typically done via an auto-tune function in the drive. In this situation, you cannot use that because of the transformer and sine wave filter. The circuit information that the drive gathers will again not be for the motor, it will be for everything. So when the drive expects a certain response from the motor and doesn't get it, it's SVC PID Loop will keep adjusting the output wave form to compensate. That will exacerbate the transformer saturation. When using a step-up transformer you must run the drive in V/Hz mode only.

 

[Besoeker]

actually very surprising
xfmr 875
motor 845 (or 880) depending on rated voltage

So you think it's surprising that the transformer overheated when its kVA rating is LOWER than that of the motor nameplate rating?
Why would anyone design it that way? With a negative margin.

You're right about one thing though. We need more detail. But I won't be holding my breath.

 

[Besoeker]

First off, I would guess that the transformer is the primary suspect. I never recommend doing this without a deep consultation with the transformer mfr, because putting a standard transformer on the output of a VFD is problematic. The gapped core, as Besoeker stated, is a common way of customizing it to avoid saturation. If you didn't account for that in the design, that transformer is likely saturating. It's the only way a transformer can limit the output to a given load. Transformers are not smart. That said, there IS a "smart" device in this situation; the VFD, and although less likely, here's how it may be the culprit.

The transformer impedance, plus the sine wave filter impedance, will drop the available voltage getting to the motor. The drive will not know this however, because the transformer and sine wave filter are on the load side of the drive; the drive cannot distinguish the transformer and filter losses from the motor current. So at any amount of the DRIVE's load current, the motor current, and thus torque, will be reduced compared to what the drive THINKS it is delivering to it. Reduced torque at the motor will result in higher slip, thus more current for the same flow, or less flow for the same current.

Good points if I may say so, young fellow....

 

[Ingenieur]

First off, I would guess that the transformer is the primary suspect. I never recommend doing this without a deep consultation with the transformer mfr, because putting a standard transformer on the output of a VFD is problematic. The gapped core, as Besoeker stated, is a common way of customizing it to avoid saturation. If you didn't account for that in the design, that transformer is likely saturating. It's the only way a transformer can limit the output to a given load. Transformers are not smart. That said, there IS a "smart" device in this situation; the VFD, and although less likely, here's how it may be the culprit.

The transformer impedance, plus the sine wave filter impedance, will drop the available voltage getting to the motor. The drive will not know this however, because the transformer and sine wave filter are on the load side of the drive; the drive cannot distinguish the transformer and filter losses from the motor current. So at any amount of the DRIVE's load current, the motor current, and thus torque, will be reduced compared to what the drive THINKS it is delivering to it. Reduced torque at the motor will result in higher slip, thus more current for the same flow, or less flow for the same current.

If the drive is then set up for a current limit (also sometimes called stall protection), it can only base that on all of the current it sees, so it will back off the output frequency and voltage to maintain that current limit. If in their calculating the transformer kVA at 100% of motor rated current, they are assuming the drive output is at 100% voltage, that may then appear, to the untrained eye, as if the transformer cannot deliver more kVA. In other words it may be a secondary affect of a perfectly normal phenomenon associated with the way they are attempting to get around using an MV drive. When using a step up transformer or a sine wave filter, you cannot expect to get full output from the motor. They have both; double whammy.

Last possibility; many VFDs now come with the factory default programming to use Sensorless Vector Control (SVC). SVC requires that the drive be tuned to the motor equivalent circuit, typically done via an auto-tune function in the drive. In this situation, you cannot use that because of the transformer and sine wave filter. The circuit information that the drive gathers will again not be for the motor, it will be for everything. So when the drive expects a certain response from the motor and doesn't get it, it's SVC PID Loop will keep adjusting the output wave form to compensate. That will exacerbate the transformer saturation. When using a step-up transformer you must run the drive in V/Hz mode only.

according to his profile he is the xfmr mfg?
How are you related to the electrical industry?:Remanufacture electric motors and manufacture new transformers.

it would appear he is an engineer familiar with motors/xfmrs?

 

[Ingenieur]

So you think it's surprising that the transformer overheated when its kVA rating is LOWER than that of the motor nameplate rating?
Why would anyone design it that way? With a negative margin.

You're right about one thing though. We need more detail. But I won't be holding my breath.

not over heated, he never stated that
he said output was limited to 840 kva for some reason
I surmise that is all the load required
basically 99%+ of motor rating

the motor name plate 4000 v/122 A/1000 hp
4160 system motors are rate 4000 similar to 480 system motors being rated 460
motor 845 kva
xmr rating 875 kva
measured load 840 kva
nothing was overloaded
besides, running a xfmr at 115% seldom causes issues other than decreased life cycle

 

[Besoeker]

not over heated,

From the op:
"I've never seen it before, usually it just runs over temp.

 

[Besoeker]

not over heated, he never stated that
he said output was limited to 840 kva for some reason
I surmise that is all the load required
basically 99%+ of motor rating

the motor name plate 4000 v/122 A/1000 hp
4160 system motors are rate 4000 similar to 480 system motors being rated 460
motor 845 kva
xmr rating 875 kva
measured load 840 kva
nothing was overloaded
besides, running a xfmr at 115% seldom causes issues other than decreased life cycle

Doesn't matter how you try to excuse it, it's just bad engineering.

 

[Ingenieur]

From the op:
"I've never seen it before, usually it just runs over temp.

'usually'
in this case it is not doing so

 

[Ingenieur]

Doesn't matter how you try to excuse it, it's just bad engineering.

I am not ready to make that leap
the numbers indicate otherwise
motor at ~100%
xfmr at 95%
drive ~ 85-90%

we need to know how it was determined that output was limited
mechanical load increased and kva plateaued?
etc