Transformers in series for step-up over distance

[SteveO NE]

Hi,

I've searched the topic (first in NEC 2017, and then scoured the web) but haven't come to a satisfactory understanding of my options for a project we are currently designing out. Sorry I am probably telling you more than you need to know but I figured thorough is better than you getting frustrated with only part of what you need to help me out here.

It is a PV installation from a 3000A service, 480V, 3PH, 4WY fed from a 1500kVA transformer, we are installing a 1MW AC system. The interconnecting breaker is shunt trip enabled and we will have a relay on-site monitoring for islanding conditions, as well as neutral current from zig-zag transformer to ensure effective grounding (as required by the EDC), and tripping the interconnection breaker if any of these conditions exist.

Here is area I have some questions as to the best value engineering (not trying to be cheap just trying to not be frivolous as MV gear adds up quick and this system size isn't that big to justify not being budget conscious): we have a long distance to run and in order to combat losses and v-drop we are stepping up to 4800V delta and stepping back down to 480V wye at the array combiner gear (1600A CB or fuse) (additionally, the facility is has constraints that we must trench new conduit in concrete and the concrete has be remain glass smooth so all trenching is done with horizontal drilling, so also expensive and instead stepping up to 4800V we can use existing spare conduits). These transformers are in series with no load in the middle, technically article 690 (my 2017 model is in my office so not sure exact reference on the top of my head) tells me to examine the system with both sides of PV transformers as primary to see which way is the safest to run protection. The transformers will be manufactured with the primary winding towards the utility - they will be reverse fed by the solar but because they are soaked from the utility before the PV kicks on, inrush will be in the direction of utility so we want that direction to be the primary to minimize inrush. For what it's worth (I know for some here its not worth much in discussions I've seen) the site will be supervised and as I pointed out there will be SCADA and automated trip functions on the interconnected circuit. We haven't specified a %Z yet and I'm not worried about guidance on that but would like to know if your suggestions have any tie to a certain %Z of course.

I think I painted the picture but for summary we have, in order, utility main, utility metering cabinet, a split through a 4000A bus gutter to the existing 3000A main gear/loads, and a 1600A interconnection feeder on the other lug of the bus gutter on opposite end from utility (can be larger if need be to accommodate inrush). The main on the interconnection feeder is shunt trip enabled with supervisory control from an SEL utility grade relay. The main of that cabinet goes to a CT cabinet for production metering as well as CTs/PTs for the relay cabinet, then to a fused disconnect that is the EDS disconnect and also would function as the primary protection of the transformer (480v primary - could be delta or wye if that helps, 4.8kV delta secondary), also lugged off the bus is a 25kVA station service tap for PV related loads (for instance the relay, ST power, and communication equipment). Out of that transformer on the 4.8kV side there is a conductor that terminates solely at another 4.8kV transformer, but this time with a 4800V primary and 480V secondary which terminates solely in the PV combiner gear. Before anyone asks, yes, we asked the EDC to receive primary service on-site @ 13.8kV and they said no.

So the questions really revolve around primary and secondary protection here and what is required/desired. Mostly we aren't clear as to what is considered the primary or secondary side of each of the two transformers.

1)Inrush concerns feeding two transformers at the same time in series?
1a) one consideration is (even if no secondary protection on the first transformer) to have a disconnect near the main gear to be able to sequence soaking on black start to avoid nuisance tripping. If this is a silly thought then just place the disconnect by the transformer in the field as the service disconnect (it just occurred to me, I didn't is there any difference with MV service disconnects as far as the rule about line of site or placard plus lockable disconnect? this site will have lockable disconnect and electronic means of lock-out, tag-out on the entire PV feeder).
1a1) Any concerns of damage to the first transformer providing inrush to soak the second? Is there any merit to the thought that the first transformer while soaking would actually act somewhat as a soft-start for the second transformer actually reducing inrush? I can see this plausible even though its counter-intuitive in my brain because usually I think lower voltage, higher current...

2)What protection would you say is required in between the two transformers (secondary on the step up from the utility and primary on the step down from the solar? I'd venture to say that from a fundamental perspective you have generation on one side and the utility on the other side of the MV system therefore the entire MV system is secondary and both transformers are technically step up (utility is a voltage source primary in the direction from the utility to the transformers, the solar is a current source primarily flowing from the inverter to the transformer).

3) This is maybe a wacky thought, could CT's on the MV line function as secondary protection if overcurrent on the MV conductor was to shunt trip the interconnection breaker, via the relay on site. (we probably need to monitor the MV side anyway to ensure no islanding on loss of phase with a Delta-Wye Transformer)
3a) Same question but if shunt trip was moved to the primary protection device of the transformer does it make a difference
3b) if yes would you worry about an islanding condition even though there is no load between the transformers...would your concern be different if we used WYE configured transformers.
3c) any reason we'd want a specific location for the zig-zag relative to the primary OCPD of the first transformer (utility side)?

I think that gets it all. Sorry for the long winded post, just trying to give all the details I can to get some guidance on this. Usually we just go straight to utility voltage for projects large enough to warrant it and get primary metering on-site but this project, being relatively small, is a unique situation to deal with. Hopefully, I am not asking something that should have been right in front of my face somewhere, but as one would expect we need the project yesterday so many of these longer term things to think about and research has to be figured out relatively quickly...winter is coming.

Thanks in advance,

Steve


P.S. Not sure if I mentioned in but the MV run from xfmr to xfmr is about 1200 conductor feet.

 

[GoldDigger]

Just a quick comment on wording.
I do not feel good referring to this as two transformers in series, because neither the primary windings nor the secondary windings are in series with one another. They are "in series" only in the sense that power must pass through both of them to get from one end of the overall circuit to the other.

But I am hard pressed to find a much better word to use here. Maybe in sequence or in cascade (which pretty well implies the output of one going to the input of the next.)

 

[electrofelon]

1. I dont think you will have a problem with nuisance tripping from inrush from starting, just size your protection near the upper end.
2. Your transformer protection code reference is 690.9(F).
3c. I would do what your utility wants as far as the zig zag goes. Sounds like it is just a silly utility thing that doesnt accomplish anything anyway. Do you have Nat grid in CT? I had a 2.5 MW PV system last year and I posted about the zig zag on here and no one could figure out what the purpose of it was. It has to do with the utility's "effective grounding" requirement. I have done multi MW PV system for a bunch of other utilities and no others required a zig zag for a grounding wye PV service. Sorry to rant, just trying to say that its hard to give advice on something that doent make sense in the first place!

Have you thought about using wue wye transformers? Just seems like that would simpler as you could use the MGN provision in the NEC, have the simplicity of a grounded system between you transformers, and use more common transformers. In case you werent aware, typical padmounts have fusing integral so you shouldnt need a separate piece of MV gear for your 450.3 protection for the second transformer.

 

[SteveO NE]

Regarding the series vs cascading I think it points to the same issue we have with PV and AHJ's as well as the table referring to primary and secondary protection. Typically I hear people refer to Primary and Secondary being synonymous to HV and LV and in fact 450.3A seems to back that by not having a primary option under 1000V and a secondary over 1000V; however, we have to consider functionally how the system is (as specified in 690.9(F), not just how the transformer is wound or what the industry wants to call something. Its along the same lines of 2017 adding the recognition of "functionally grounded systems" to finally clear up that silly issue with "ungrounded" PV and tranformerless installs (way to have to be different from the international community recognizing that 10+ years ago). To avoid sounding like I want to dispute semantics, let me make sure its clear what's happening instead of assigning a title to it (coming from the utility industry I wouldn't call this cascading transformers either and I didn't know what else to call it). From a transformer winding perspective we have Utility-Pri(1)-Sec(1)-Pri(2)-Sec(2) -PV. From a​

functional

perspective during nominal operation we have Utility-Pri(1)-Sec(1)-Sec(2)-Pri(1)-PV.

@electrofelon

1 ) To my point above, 450.3A vs. B, is this table referring to HV/LV and not actually Pri/Sec because there is no configuration that has a primary less than 1000V and a secondary above. Would you say that max primary protection is 125%? That is how I would understand it but this transformer is a "transformer over 1000V" as the title of the table references but there is no column that is applicable in the table...this is really the fundamental issue why my question arises in the first place...do I use table B for the LV Primary and table A for the HV Secondary? It should be noted that AHJ's around here do in fact argue that HV is always the "primary" as far as NEC is concerned and the LV is always the secondary and they dispute the relevance of 690.9(F) (thanks for the reference) because (even though 690 states is supersedes any conflicts, they say 450.3 exists and must be followed).

2) thank you

3c) The utility doesn't know what they are looking at anyway to your point...there is nothing to calculate because the solar is a balanced load with the neutral as a voltage reference only...it is impossible for it to generate neutral current. Generally they argue amongst themselves that it should be between the utility transformer and the main switch, but then there metering department argues that customer equipment can't be ahead of metering:slaphead:. In this case they wouldn't particularly care because of how this site is configured.

Of note, this is Eversource and yes I'm also in N-Grid territory. I used to be on the other side of the table and argued against it while I was there too. I did have one engineer able to prove one thing through math, though I still think its functionally a stretch because of the grid should be effectively grounded; he pointed out that its not the PV plant itself that is the issue its the interaction of the grid during a trip of the line at the substation with generation still coming in and creating neutral currents because of customer loads being unbalanced. He was able to prove mathematically that the system would cause v-rise otherwise and a single leg could get above their +3% (I know most places it is +5% but for this EDC its +3%). However, it did completely ignore the grounding, resistance, and capacitance of the grid and grid conductors...also if I am thinking correctly that would mean we are looking for neutral current inbound from the utility, not outbound to the utility so we'd have to program our relay in that direction which is not what they are having us do ever...so I take it for what it is, a pay to play provision set up by New England utilities to keep the rates high and reduce the risk of losing revenue from their largest customers...what a shame the utility company has competition now.

As for using a wye transformer, that's OK for me, I think a wye is 4.6kV instead if my memory serves me well; this was an option by the installer's supplier and it saves the cost of the copper on the neutral. Could you clarify what you mean by the MGN provision...I understand you are talking about Multigrounded neutral, but is there something I am missing about something potentially justifying the cost of the neutral conductor?

Thanks for pointing out the obvious:slaphead:...I did forget that they would have primary side fuses. Can I get them with secondary fuses as well though? We still have the issue of secondary protection unless we are saying this is table A, Secondary Column, Over 1000V, and Supervised Row. Again that's what makes sense except table B doesn't make reference to a Primary on that same transformer being less than 1000V so I'm trying to make sure I'm using the correct table.
​

Thanks again,

Steve

 

[SteveO NE]

It won't let me edit so I'll put it here:

… ​

From a​

functional

perspective during nominal operation we have Utility-Pri(1)-Sec(1)-Sec(2)-Pri(1)-PV. -> This is really the crux of my question here, 690.9(D) says to look at it functionally, so do I consider this series from a protection standpoint because steady state operation has this secondary to secondary its only pri to secondary during transient state. In addition to what was already mentioned about the high side of the transformer in the field, if its agreed upon that its series during steady state does that mean that I could come into (or out of depending on your perspective) the transformer from the PV through a disconnect that goes out to the CB for each of 4 combined circuits instead of a single main out at the field?

(Exception: A power transformer with a current rating on the side connected toward the
interactive inverter output, not less than the rated continuous output current of the inverter,
shall be permitted without overcurrent protection from the inverter. )

Also, I forgot to ask...so what's the verdict on a service disconnect of the transformer in the field. The above provision is only talking about considerations for OCPD, not for means of disconnect. Could it be eliminated with LOTO provisions on the disconnect to the entire feeder and a placard on both transformers to the location of the LOTO? Shutting that down would make the transformers cold. Just because I can doesn't mean I should...so the follow up to can I, is what's your opinion on should I...again not trying to be cheap just trying to be cost conscious and not put in redundant equipment if I don't need it. Does the SHOULD I question bring into question the use of wye instead of delta...in other words maybe something along the lines of the capacitance of the wires has nowhere to discharge without a grounded conductor when the system is deenergized? Though a disconnect on the MV side wouldn't fix that anyway.​

 

[electrofelon]

Steve,


Lots of stuff going on. Ill try and reply more later, but a few points for now:

Why not just go up to 15KV class transformers? My thinking with is that 15KV is so ubiquitous, not sure you would save anything by going lower. In fact the smaller wire size alone should be a big $$ saver. Note load break elbows start at 15kv class. I know there is lower cable ratings than 15kv, but concentric neutral 15 kv cable is so common. If you set it up like a MGN, you just run three concentric MV cables so you are not running any more conductors. That was my thinking on that.

Regarding the disconnect, I would try to avoid that and use the remote lockable provision. I am not very familiar with MV equipment prices, but I assume a Mv 200A disconnect is not going to be cheap.

Regarding 450.3(A) vs (B), I admit it is not totally clear, but I believe "over 1000 volts" means one or both windings is over 1kv, and if just one is, it doesnt matter which one that table A applies.

 

[SteveO NE]

Ah...that is a thought using concentric cable, great idea. I will check into the pricing and lead time going that route.

Thanks for the opinion on A vs. B that is what I was thinking.

So you think I can use the remote lockable provision through even though there is a transformer in the middle? I do believe functionally it does the same thing, I just wanted to make sure there wasn't any NEC reference that would prohibit devices in between your remote provision.

I think you hit everything (not specifically my thoughts on pri/sec protection at the array but your opinion on which table to follow effectively answers the question), the rest of that was mostly supplementing your rant on the utilities need for "effective grounding" :rant:.

Thanks so much. Big help on this. I will try and follow up to close the loop on if going 15kV will save the installer $$ in the big picture.

 

[electrofelon]

Ah...that is a thought using concentric cable, great idea. I will check into the pricing and lead time going that route.

Thanks for the opinion on A vs. B that is what I was thinking.

So you think I can use the remote lockable provision through even though there is a transformer in the middle? I do believe functionally it does the same thing, I just wanted to make sure there wasn't any NEC reference that would prohibit devices in between your remote provision.

I think you hit everything (not specifically my thoughts on pri/sec protection at the array but your opinion on which table to follow effectively answers the question), the rest of that was mostly supplementing your rant on the utilities need for "effective grounding" :rant:.

Thanks so much. Big help on this. I will try and follow up to close the loop on if going 15kV will save the installer $$ in the big picture.

450.14 clearly allows a remote lockable disconnect. Keep in mind however that is just for the transformer disconnect requirement and there may be other disconnect requirements like for a structure, but I think you should still be good since the NEC has recently clarified that "equipment" is not a structure. On the LV side of the second transformer you would have your switch/OCPd which will meet you tranny protection, secondary conductor protection, etc, and can be the kill switch for the whole system. The tranny would still be energized but I dont see any issue.

When I get a few spar minutes, I want to try to digest your explanation of the "effective grounding" purpose - Ill get back to you.

Let me know how the different MC cable options price out. One of the utilities here lets us run primary from the pole to their tranny on the customer premesis so some supply houses here stock 15kv concentric. IIRC its like 2.20 a foot, but not sure what size that is, probably #2. Say you used 7.2/13.62kv thats only 63 amps