SMA Tripower connects with lost phase

[Coop]

Hi everyone,

I have an installation in CT where the utility requires a phase loss test. The installation consists of 4-SMA Tripower 20000TL inverters with a load side connection via a 480/208 Y-Y 100kW transformer to the MDP. The inverters are 3 phase 277/480 and the grid is 3 phase 120/208. The utility had us turn off the power and disconnect a phase from the utility side of the transformer. After re-energizing, to our surprise, the inverters started and connected as normal, even though the phase was disconnected. SMAs response was that the inverters have met UL1741 and we should just turn up the reconnect voltage to prevent this from happening. Turning up the reconnect voltage did allow the inverters to pass the test but I am concerned that they may not shut down if a phase is lost on the fly. The other problem is they may not reconnect after a grid failure if the utility voltage is low, like in the summer with AC loads. Also wondering how the grid power was passed by the inverters when there was a phase missing. It is a common core three phase transformer so there is voltage present on the leg that was disconnected (about 268 to ground), but it can't be at the correct phase angle and possibly is the wrong frequency. Anyone have any experience or ideas on this?

Thanks,

[SunEagle]

Hi everyone,

I have an installation in CT where the utility requires a phase loss test. The installation consists of 4-SMA Tripower 20000TL inverters with a load side connection via a 480/208 Y-Y 100kW transformer to the MDP. The inverters are 3 phase 277/480 and the grid is 3 phase 120/208. The utility had us turn off the power and disconnect a phase from the utility side of the transformer. After re-energizing, to our surprise, the inverters started and connected as normal, even though the phase was disconnected. SMAs response was that the inverters have met UL1741 and we should just turn up the reconnect voltage to prevent this from happening. Turning up the reconnect voltage did allow the inverters to pass the test but I am concerned that they may not shut down if a phase is lost on the fly. The other problem is they may not reconnect after a grid failure if the utility voltage is low, like in the summer with AC loads. Also wondering how the grid power was passed by the inverters when there was a phase missing. It is a common core three phase transformer so there is voltage present on the leg that was disconnected (about 268 to ground), but it can't be at the correct phase angle and possibly is the wrong frequency. Anyone have any experience or ideas on this?

Thanks,

I also don't understand how the inverters met UL1741. I believe it clearly states that with a "loss of phase" on the Utility side the inverter should disconnect within 2 sec.

I can see why they wouldn't trip for a "Utility low voltage" where the trip point is 88% (measured phase to neutral) so 268v would be > 95% and does not meet that condition.

But the loss of a phase is totally different condition, unless the SMA inverters use the criteria of the measured "phase to ground" voltage instead of "phase to phase" to determine if a phase is lost.

I don't think it would affect the frequency but you are right the vector angle for the lost phase wouldn't be correct.

[Coop]

Hi SunEagle,

Thanks for looking into this. I think I may have found the problem. When the phase is pulled from the utility side, the transformer is then connected like an open wye, which uses the neutral connection as a reference to the disconnected leg on the utility side. This is why the inverter is qualifying the connection. The 3rd leg on the 480 side is likely the correct phase angle and frequency, just a little low on voltage level. This could be fixed by simply removing the neutral on the utility side, but I don't think a wye is ok to run without it, unless you are very certain there will not be an unbalance in voltage. The YY transformer was recommended for this design by SMA. Hopefully they will cover our expenses to replace it. I think we have this setup on 3 different sites. I will post any updates I have on the out come.

Thanks,

[SunEagle]

Hi SunEagle,

Thanks for looking into this. I think I may have found the problem. When the phase is pulled from the utility side, the transformer is then connected like an open wye, which uses the neutral connection as a reference to the disconnected leg on the utility side. This is why the inverter is qualifying the connection. The 3rd leg on the 480 side is likely the correct phase angle and frequency, just a little low on voltage level. This could be fixed by simply removing the neutral on the utility side, but I don't think a wye is ok to run without it, unless you are very certain there will not be an unbalance in voltage. The YY transformer was recommended for this design by SMA. Hopefully they will cover our expenses to replace it. I think we have this setup on 3 different sites. I will post any updates I have on the out come.

Thanks,

I am curious as to why you installed inverters that deliver 480/277 when the grid is 208/120. Do you use that 277v for lighting or other equipment?

[Coop]

The SMA Tripower inverters we are using only come in 277/480V. They are transformer-less 1000V ungrounded string inverters.

[inetdog]

Hi SunEagle,

Thanks for looking into this. I think I may have found the problem. When the phase is pulled from the utility side, the transformer is then connected like an open wye, which uses the neutral connection as a reference to the disconnected leg on the utility side. This is why the inverter is qualifying the connection. The 3rd leg on the 480 side is likely the correct phase angle and frequency, just a little low on voltage level. This could be fixed by simply removing the neutral on the utility side, but I don't think a wye is ok to run without it, unless you are very certain there will not be an unbalance in voltage. The YY transformer was recommended for this design by SMA. Hopefully they will cover our expenses to replace it. I think we have this setup on 3 different sites. I will post any updates I have on the out come.

Thanks,

The details of wye-wye versus delta-wye or delta-delta transformers for a utility connection are complicated and have some subtleties that are not at all obvious at first glance.
An electrician or electrical engineer with experience in this area would be very helpful.

Part of the problem is the ambiguity of the "loss of a phase".
If the utility source is a wye generator or wye transformer, then loss of a supply phase would either be the opening of one transmission line or the grounding of that line, depending on how the "loss" happened. The more interesting of the two if the event of one transmission line going open.
In that specific case, the voltage on your end of that line, with respect to ground and to the neutral, will be zero unless there are other loads on that stretch of transmission line. If there are other loads, especially transformers with a delta secondary and three phase motor loads, then the voltage on the transmission line can be very close in phase and amplitude to the normal voltage. That sort of phase loss is hard to detect, period.
The UL standard would have to closely define, in a way suitable for testing, just what phase loss actually was!

If the phase loss consists of the utility opening one hot line feeding your wye-wye transformer and there are no other loads on the secondary of the transformer except the inverter, I would expect the voltage on that wye leg to go to zero (or at least for the inverter to see that phase line as islanded) because the stiffness of that voltage with respect to changes in inverter output would be very small.

In your case I would not recommend opening the connection between the wye point and the utility neutral. Even though if a normal transformer feeding a load has a delta secondary and a wye primary, it is reasonable and in some cases necessary to disconnect the wye point on the primary side.

[sensij]

@inetdog
Welcome back!

[SunEagle]

The details of wye-wye versus delta-wye or delta-delta transformers for a utility connection are complicated and have some subtleties that are not at all obvious at first glance.
An electrician or electrical engineer with experience in this area would be very helpful.

Part of the problem is the ambiguity of the "loss of a phase".
If the utility source is a wye generator or wye transformer, then loss of a supply phase would either be the opening of one transmission line or the grounding of that line, depending on how the "loss" happened. The more interesting of the two if the event of one transmission line going open.
In that specific case, the voltage on your end of that line, with respect to ground and to the neutral, will be zero unless there are other loads on that stretch of transmission line. If there are other loads, especially transformers with a delta secondary and three phase motor loads, then the voltage on the transmission line can be very close in phase and amplitude to the normal voltage. That sort of phase loss is hard to detect, period.
The UL standard would have to closely define, in a way suitable for testing, just what phase loss actually was!

If the phase loss consists of the utility opening one hot line feeding your wye-wye transformer and there are no other loads on the secondary of the transformer except the inverter, I would expect the voltage on that wye leg to go to zero (or at least for the inverter to see that phase line as islanded) because the stiffness of that voltage with respect to changes in inverter output would be very small.

In your case I would not recommend opening the connection between the wye point and the utility neutral. Even though if a normal transformer feeding a load has a delta secondary and a wye primary, it is reasonable and in some cases necessary to disconnect the wye point on the primary side.

First off welcome back and hope all is well.

One area that can happen is if that Y-Y transformer is fed by 3 single fuses. It doesn't happen often but I have seen the primary of a transformer lose a "phase" when one of the 3 fuses blew.

The issue is with the inverter software and what it determines a "lost phase" and how it is measured. In the case of the OP's inverter it looks like SMA may be measuring the phase to neutral voltage and if it didn't go below 88% of the Vnom trip point it may still thought all is ok. I am guessing this was the issue since SMA suggested "turning up the reconnect voltage" which in turn kept the inverter from reconnecting.

I am still doing some research on what happened but from others are telling they think the transformer being a "common core" Y-Y allowed enough voltage to get through to the 480/277 side even though a phase was lost on the 208/120 side. If the inverter generated only 480v 3 phase delta then a Y-Y would not have been used.

[Coop]

Update

Thank you to everyone looking into this

I think my idea that when the phase was pulled on the utility side the GroundedY/GroundedY became open wye/wye was not correct. It may be partially correct in that the utility side becomes connected as an open wye when a single phase is pulled, but this doesn't explain the behavior of the voltages on the secondary side. The secondary side would have to be connected as an open wye secondary or open delta, which it is not. Is it possible the common core of the transformer is acting like additional windings connected in delta and that is where the mystery phase is coming from? With one phase pulled on the utility side, the voltages on the secondary side were 277, 276, 268. They are actually at the correct phase angle according to my colleague who tested with a power analyzer. So essentially the utility open wye is coupled to the delta derived from the cores, which in turn is coupled with the wye secondary. All loads were disconnected at the time the test was done. The 268V phase is probably very power limited? Any thoughts on this? The scary part is it is close enough for the inverter to connect.

This has been very educational, I never had a reason the think about all the complex interactions in a 3 phase transformer before this.

Thank you,