Inrush problem

[StewG]

I have designed and built a single purpose burn in system for testing products at my plant. The only problem I have with the system is that the 3, 15kva 1 ph tranformers that provide the test voltages have a high curren in rush problem even with no load present, blowing a fuse regularly. The transformers do not get energized, through a contactor, until the "start test" button is pushed. What would be the downside of bypassing this contactor and having the transfomers energized all the time, so that there is no inrush? The only time inrush would exist then is if the main disconnect were disengaged/engaged or if there were a power outage requiring a restart. Usually a single fuse change gets it going, but it is too regular for production purposes. Max amp reading logged was 195 amps, no load, inrush only. Currently fused at 30 amps at service disconnect and 60 amps per phase internally. What say ye? Bypassing contactor and leaving tranformers hot?

 

[Smart $]

... Currently fused at 30 amps at service disconnect and 60 amps per phase internally. What say ye? Bypassing contactor and leaving tranformers hot?

I'm assuming you mean "primary disconnect" rather than "service disconnect".

Assuming the 30A fuses are blowing, I recommend upsizing them... but that is contingent upon disconnect rating being greater than or equal to whatever you upsize to, and also the primary's wire size, type, required derating, etc.

Bypassing the contactor and leaving the transformers hot is not going to cure the inrush problem for day-to-day re-energization (if any), power outage recovery, or after maintenance shutdowns.

 

[gar]

100901-1002 EST

Add an appropriate series resistor in each transformer primary. Add a second contactor that shorts the series resistors after a short time delay. A delay of maybe 20 to 60 milliseconds.

.

 

[winnie]

Inrush into a transformer is caused by the changing flux in the core exceeding the saturation level. Normally the transformer flux will cycle from -max to +max; but at startup the flux can go from 0 to 2xmax, causing severe inrush.

The 'resistive precharge' that gar describes essentially synchronizes the changing flux to the supply voltage. You still get that initial inrush, but the inrush current is limited by the resistor.

A slightly different approach to the resistive precharge is to have two parallel contactors, one suppling via resistors, one direct.

An entirely different approach is to use a system which synchronizes the closing of the contactors to the AC cycle. With correct synchronization, inrush on a single phase transformer can be reduced to zero.

-Jon

 

[StewG]

Thanks for the thoughts. I wonder now if a "soft start" unit would work on transformers. I have a couple from dismantled equipment that were used in the start up of high power motors.

 

[Jraef]

Generally, the magnetizing current inrush of a transformer is too fast to be effectively controlled by a soft starter. By the time the electronics wake up and start trying to do anything about it, the party is already over. But for sure, zero-cross SCR switching of large transformers is done all the time and the basic technology is the same. So at the least, you would have that capability and that can help avoid a number of problems, including the highest possible peaks of magnetizing inrush.

 

[gar]

100901-1850 EST

Jraef:

How does zero-crossing turn on reduce peak in-rush current?

The residual flux in the transformer core, and the turn-on volt time integral determines what happens to the transformer core flux after turn-on.

Suppose the residual flux is near a maximum positive value, and turn-on of the input voltage is at a zero-crossing such that the turn-on volt-time integral is in the direction to increase the already high flux level, then the transformer core is driven into saturation and the input current is high. If the turn-on at the zero-crossing was of the opposite slope, then the flux change would have been in the opposite direction, and no large peak inrush. But this requires knowledge of the residual flux state.

If the residual flux can be near zero, then turn-on at a zero crossing of either slope is good. Only a small turn-on current will occur.

How the transformer is turned off and/or knowledge of the residual flux in the transformer core is necessary to create a system that produces a small inrush current without external current limiting.

If the volt-time integral can be gradually increased in both + and - directions, then inrush can be small. This could be done with SCR control starting with small spikes and gradually decreasing the the turn on phase angle from near 180 (near 0 volt-time integral) toward 0 (maximum volt-time integral).

.

 

[dbuckley]

How the transformer is turned off and/or knowledge of the residual flux in the transformer core is necessary to create a system that produces a small inrush current without external current limiting.

The better quality lighting dimmers that can control inductive loads well do exactly this; they manage the off as well as the on to ensure that the dimmer knows which way the core is residually magnetized so the the inrush current is minimized.

 

[winnie]

100901-1850 EST
If the residual flux can be near zero, then turn-on at a zero crossing of either slope is good. Only a small turn-on current will occur.

I beg to differ.

Assuming zero residual flux, the proper turn-on point to minimize inrush is at the _peak_ of the voltage sine wave.

Before turn on, the transformer current _must_ be zero. In order to minimize the transient, you want to turn on at the point where in steady state the magnetizing current would be zero anyway.

Another way of looking at it:
In normal steady state operation, the ET product of a half cycle takes the core flux from one maximum to the other. So the ET product of a quarter cycle is the right amount to take the core flux from zero to maximum.

You can improve the transient reduction if you know the residual flux in the core, but in a good transformer the residual should be a small fraction of saturation, so you can assume zero residual as the first order solution.

-Jon

 

[gar]

100902-1208 EST

winnie:

I was wrong and you are correct that with 0 initial flux one wants to turn on at a voltage peak. To go from the mid point of flux (0) to maximum only takes 1/2 the volt-time integral of a half sine wave. The full half sine wave is a flux change from -max to +max under steady-state conditions,

On initial residual flux I do not think it is clear cut that it is near 0. As an indirect measure with a typical transformer and using zero voltage turn on I can get at least a 5 to 1 ratio in peak inrush current resulting from different turn off conditions.

.

 

[iceworm]

The only problem I have with the system is that the 3, 15kva 1 ph tranformers that provide the test voltages have a high curren in rush problem even with no load present, blowing a fuse regularly.

Max amp reading logged was 195 amps, no load, inrush only.

A couple of questions:

The primaries are 480V?

The three separate transformer primaries are connected delta?

W

 

[robbietan]

I have designed and built a single purpose burn in system for testing products at my plant. The only problem I have with the system is that the 3, 15kva 1 ph tranformers that provide the test voltages have a high curren in rush problem even with no load present, blowing a fuse regularly. The transformers do not get energized, through a contactor, until the "start test" button is pushed. What would be the downside of bypassing this contactor and having the transfomers energized all the time, so that there is no inrush? The only time inrush would exist then is if the main disconnect were disengaged/engaged or if there were a power outage requiring a restart. Usually a single fuse change gets it going, but it is too regular for production purposes. Max amp reading logged was 195 amps, no load, inrush only. Currently fused at 30 amps at service disconnect and 60 amps per phase internally. What say ye? Bypassing contactor and leaving tranformers hot?

in our poco, the fuses we use are fast enough to detect faults but slow enough not to blow when transformers are energized. use a slower acting fuse for your system

 

[don_resqcapt19]

I have been told, but not sure that it is correct, that a secondary load can reduce the primary inrush current.

 

[gar]

100907-0717 EST

don:

I suspect that the observation has less to do with the load or no load at turn on vs the kind and amount of load at turn off.

The residual flux state of the core, determined at turn off, and the point and direction of voltage at turn on determines the inrush current.

.

 

[mxslick]

While I do enjoy the theoretical discussion going on, we have still not answered the OP's question.

Ok, here's how I see this getting solved:

1: Replace primary contactor with a manual disconnect switch with fuses sized to handle the transformer inrush current (Or use a non-fused disconnect and size the feed breaker.)

2: Install the contactor on the secondary side to apply test power. (I would take the precaution of using a key switch, in addition to any other switch(es) controlling the contactor, to act as an extra safety lockout.)

Problem solved.

Only drawback I see is the power consumed by the transformers when idle.

 

[gar]

100907-1245 EST

mxslick:

In the first three responses solutions were provided.

.

 

[mxslick]

gar:

Perhaps, but they were talking about resistors, extra contactors, et al.....

Why overthink this? It is not rocket science and sometimes the simplest solution is the best. :grin:

 

[Open Neutral]

100907-0717 EST

don:

I suspect that the observation has less to do with the load or no load at turn on vs the kind and amount of load at turn off.

The residual flux state of the core, determined at turn off, and the point and direction of voltage at turn on determines the inrush current.

.

I consulted my buddy the magnetics guru; the only person I know who really understands3-phase ferroresonant & ordinary transformer design. This is how he explained it to me.

Further, he suggested the previous approach; in effect a reduced voltage starter -- resistor based. Of course no one makes same any longer but it's simple: 3 big honking Milwaukee Resistors, and 2 contactors; the shunt one closing a few ms. after the main. I'll be using that scheme.

I also looked at two approaches from Germany; they in effect tickled the transformer back into phase before applying the full line. I have a bunch of links on same if anyone is interested. More German than I recall would likely help...