Primary fuses failing due to transformer inrush

[Magic Gorge]

Obviously the fuses should be sized to handle the inrush, but why would only one phase blow during the startup of a 3-phase transformer?

 

[xptpcrewx]

Obviously the fuses should be sized to handle the inrush, but why would only one phase blow during the startup of a 3-phase transformer?

Inrush is usually higher on one phase depending on when the switching occurred relative to the voltage waveform. Switching at a zero crossing will always give you the worst case inrush.


Sent from my iPhone using Tapatalk

 

[kwired]

Inrush is usually higher on one phase depending on when the switching occurred relative to the voltage waveform. Switching at a zero crossing will always give you the worst case inrush.


Sent from my iPhone using Tapatalk

And once one fuse blows, only possible current left is between the other two phases which have different time for zero crossing point.

 

[Besoeker3]

Obviously the fuses should be sized to handle the inrush, but why would only one phase blow during the startup of a 3-phase transformer?

There are different types of fuses depending on the application,,,,,,,,,,,,,,

1. Fuse type
Fuses are identified by 2 letters, according to their application category. In low voltage installations gG and aM fuses are mainly used.


gG cartridges
gG cartridges are for general use and they protect the circuits against low and high overloads and, of course, against short circuits. gG cartridges are marked in black.


aM cartridges
aM cartridges are used with electric motors and they protect against high overloads and short circuits. They are calculated to resist certain temporary overloads (starting a motor).

 

[gar]

210505-1710 EDT

Xtptcrewx:

Why?

See my scope plots at:

BETATRONICS&reg, Electrical photos, communication, Long Cable, High Baud Rate, RS-232, fast, I232, vendor, supplier, manufacturer, industrial

Electrical photos.

beta-a2.com

See P6 and P7.

Whether the core is driven more into saturation, or not, at application of voltage, is a function of the direction of residual flux in the core at the turn on time, and the direction in change of flux as a result of the applied volt time integral.

.

 

[xptpcrewx]

210505-1710 EDT

Xtptcrewx:

Why?

See my scope plots at:

BETATRONICS&reg, Electrical photos, communication, Long Cable, High Baud Rate, RS-232, fast, I232, vendor, supplier, manufacturer, industrial

Electrical photos.

beta-a2.com

See P6 and P7.

Whether the core is driven more into saturation, or not, at application of voltage, is a function of the direction of residual flux in the core at the turn on time, and the direction in change of flux as a result of the applied volt time integral.

.

The part about residual flux is true. As for why you get maximum DC offset for switching at a zero crossing, its math and RL circuit physics. As you probably already know, a transformer is basically an RL circuit approximated by a differential equation. Based on the physics of inductive circuits and the initial conditions (at the time of switching), the solution will contain an exponential transient term that decays with time (DC offset/decay). The DC offset can be thought of the forcing condition that reconciles two conflicting phenomena: (i) current lags the voltage, so at zero volts, the current is either maximum positive or maximum negative, and (ii) current cannot change instantaneously from a zero value before switching to a maximum value at the time of switching.

Said in a practical manner, the circuit can only accept energy (an increase in current) at a certain rate. The DC offset current charges the inductance at this acceptable rate (which is the same thing as the circuit time constant). Why it’s predominant in one phase of a three-phase systems has to do with the fact that the other phases are not seeing a zero crossing at the time of switching.

Note: Everything I’ve mentioned thus far ignores the non-linear behavior of transformers cores, which is yet another mechanism for inrush.


Sent from my iPhone using Tapatalk

 

[gar]

210512-2252 EDT

Xptpcrewx:

The linear transient current (no core saturation) resulting from applying voltage to an unloaded transformer at a voltage zero crossing is relatively small compared to the transient current from the core being driven into saturation. The transient current plot that I made and referenced is the real problem in most cases of tripping an over-current device on the primary side.

.

 

[xptpcrewx]

210512-2252 EDT

Xptpcrewx:

The linear transient current (no core saturation) resulting from applying voltage to an unloaded transformer at a voltage zero crossing is relatively small compared to the transient current from the core being driven into saturation. The transient current plot that I made and referenced is the real problem in most cases of tripping an over-current device on the primary side.

.

These are interrelated. The time-varying inductance associated with the saturation of the core is just another variable to the problem I just described.


Sent from my iPhone using Tapatalk

 

[GoldDigger]

This explains why only one fuse blows each time.
If a different fuse blows each time, this is exactly what would be expected.
But if the same one fuse blows each time there would have to be a different reason.

 

[winnie]

gar, I believe you are focusing on one factor of an inter-related pair of factors.

Consider steady state operation of the transformer. Core flux and magnetizing current lag the terminal voltage by 90 degrees, thus when voltage is at a zero crossing flux is at a + or - maximum. The volt-second integral of a half cycle of the applied voltage determines the change in flux over the course of that half cycle. The transformer will be designed so that at maximum design volts/Hz the flux change in steady state will be from near - saturation to near + saturation.

For any given amount of residual flux there is a corresponding phase of applied voltage that matches the steady state condition; if you close the switch on the transformer right at that point then the system will just drop into steady state operation with no inrush.

In the case where the residual flux in the core is 0, and the voltage is applied right at zero cross, then flux will try to climb to 2x its normal maximum value. This pushes the transformer hard into saturation causing worst case inrush. Take the transformer in the same 0 residual flux state and apply voltage at peak (with only 1/4 cycle to get to zero cross) and you have no inrush.

For any other residual flux there will be a corresponding worst case and best case phase angle to apply the voltage.

Both residual flux _and_ phase of applied voltage alter the inrush current magnitude.

Because transformers are built with 'soft' magnetic materials, one would expect flux to decay away to zero over time. But these materials are not perfectly soft and always retain some residual flux unless intentionally degaussed. So there will be some random flux in the core related to when the supply breaker was opened.

I believe that if the breaker is closed at exactly the same phase angle as opened, inrush current will be minimized. However I don't know if breaker operating time makes it practicable to phase synchronize opening and closing time.

-Jon

 

[synchro]

Obviously the fuses should be sized to handle the inrush, but why would only one phase blow during the startup of a 3-phase transformer?

So is this a true 3-phase transformer with a delta or wye primary, or is it a V-V (open delta) or T-T connection of two single phase transformers (sometimes T-T's are in a single housing, not sure about V-V's) ?
The asymmetry of the latter arrangements might cause the worst case inrush current to be localized on one particular phase (for example, where the two open delta primary windings are connected together).

 

[topgone]

I usually refer to published maximum practical inrush current factors of transformers. On one Eaton data, I have found out that the bigger the kVA, the smaller the maximum practical inrush current factor! That only shows that the bulkier cores will give you a better and manageable inrush than a transformer of the same capacity with a higher designed operating flux (smaller core)! Hope this helps on others' decision-making!

 

[gar]

210517-1457 EDT

winnie:

What I am focused on is a controlled experiment on transformer inrush current. This really has nothing to do with steady state current. My experiment has to do with the initial flux state of the transformer core at the time that voltage is applied, applied source voltage (assumed very low impedance relative to the transformer load that will develop upon application of voltage), voltage is a sine wave of known voltage and frequency, and that voltage is turned on at a zero crossing. Two zero crossing phasings are possible. One increases the flux and the other decreases it. I pick the increasing one to force the core further into saturation. Note turning on voltage at a voltage zero crossing produces the greatest amount of flux change.

I do not look at this problem as a linear circuit, but rather as a nonlinear transient problem.

The hysteresis curve shape, and transformer load for the particular core material will determine what the residual flux state is at turn off. Residual flux state is relatively stable for very long periods if not subjected to demagntizing forces. I still have magnetic recordings from the early days that are in good condition, meaning late 1940s. I have a compass from the 1930s that still works.

It is true that some magnetic materials may demagnetize easier than other materials. Mechanical shocking may do demaging.

.

 

[synchro]

210517-1457 EDT
... I still have magnetic recordings from the early days that are in good condition, meaning late 1940s.

Are those on wire instead of tape? Ampex had their first production audio tape recorder in 1948.

 

[gar]

210517-1712 EDT

synchro:

I have both wire recordings and tape. The wire recordings were made in 1946, but I don't presently have a way to play them, These were made on a mechanism made in Detroit, and the audio stuff I built. The the wire recorder had a capstan drive mechanism. The tape recordings started just before 1950.

The the tape and disk recordings I can still play. I also have a disk recorder from the same time frame (phonograph type disks, lacquer on aluminum) (these records are still playable, as are disks from the 1920s) (I also have radio components from the 20s).

.

 

[MTW]

I usually refer to published maximum practical inrush current factors of transformers. On one Eaton data, I have found out that the bigger the kVA, the smaller the maximum practical inrush current factor! That only shows that the bulkier cores will give you a better and manageable inrush than a transformer of the same capacity with a higher designed operating flux (smaller core)! Hope this helps on others' decision-making!

Do you happen to have a link for that Eaton data sheet, or a title to search for?

 

[winnie]

winnie:

What I am focused on is a controlled experiment on transformer inrush current. This really has nothing to do with steady state current. My experiment has to do with the initial flux state of the transformer core at the time that voltage is applied, applied source voltage (assumed very low impedance relative to the transformer load that will develop upon application of voltage), voltage is a sine wave of known voltage and frequency, and that voltage is turned on at a zero crossing. Two zero crossing phasings are possible. One increases the flux and the other decreases it. I pick the increasing one to force the core further into saturation. Note turning on voltage at a voltage zero crossing produces the greatest amount of flux change.

Gar,

Perhaps I am misunderstanding what you meant by:

The linear transient current (no core saturation) resulting from applying voltage to an unloaded transformer at a voltage zero crossing is relatively small compared to the transient current from the core being driven into saturation. The transient current plot that I made and referenced is the real problem in most cases of tripping an over-current device on the primary side.

As I read the above, you are stating that the phase of the voltage when applied to the transformer is a small factor in peak inrush, and that the residual core magnetization dominates the startup transient effect. I am claiming that the two factors are intertwined and neither dominates.

Steady state operation is relevant to understanding the startup transient for several reasons. The steady state flux swing from peak- to peak+ flux is the flux change which must happen when voltage is applied at a zero cross. If voltage is applied at zero cross and there is no residual magnetization, then this flux swing drives the core into hard saturation. We both agree (and your oscillographs show) that saturation is the root of transformer inrush; the inrush current appears to start climbing 90 degrees into the voltage waveform.

So the _key_ is that when the initial volt-second product plus the residual flux drives the core into saturation we will see inrush current. The converse of this statement is that if the initial volt-second product does not drive the core into saturation, we won't see inrush current. There are two plausible situations for this: 1) A transformer with such low steady state peak flux that no combination of residual flux and applied phase drives it into saturation or 2) A transformer where the phase of the applied voltage is matched to the residual flux such that the transformer is 'instantly' in its steady state operating condition.

Thanks for any clarification.

Jon

 

[topgone]

Do you happen to have a link for that Eaton data sheet, or a title to search for?

Just type "transformer inrush, Eaton" as your keyword in searching the web.
I have found one, here.

 

[gar]

210517-2341 EDT

winnie:

For driving into maximum saturation from a given source you want to apply excitation of the voltage to the magnetic coil and core at a voltage zero crossing. This provides the greatest volt-time integral to the core, and you want the direction of this flux change to be in the direction of increasing the flux from that of the residual flux in the core. This takes you to the maximum saturation point. The starting residual flux in the core is a function of the load on the secondary, and the way source voltage was reduced at last turn off. If you gradually reduced voltage at turn off you could essentially demagnetize the core, and have nearly zero residual flux. But this not the usual way to turn off a load.

At turn off you follow the hysteresis curve back from the peak flux point to where the excitation current is zero. Rather near zero voltage when an SCR or Triac is the switch.

.

 

[winnie]

gar:

I think you are describing the head of the elephant and I am describing the tail. We agree on the physics

You've described the condition that gives maximum saturation and would likely cause a very large inrush current. What condition would give _minimum_ saturation and thus prevents large inrush current?

-Jon