Strange currents / phase angles at co-gen site 1

[BSEE1980]

Background:
350Kw co-gen in Indiana, mature radial network, local loads mostly residential.
Gen is 480Y, grounded at the GSU. GSU is 480Y to 12.47Y.

While running on load banks, resistive and reactive, the gen was fine, metering all looked very good. Voltages, currents, KVAR all good. The problems start when they parallel with the gird.

After running for a few minutes, they are getting phase current unbalance trips, so the utility brought out their metering dude and he took some data with the basic power analyzer in from the meter, see .jpg file.

A phase current lags by 47 degrees, B lags by 25 degrees, and C LEADS by 26 degrees. Yes, two phases lag and one leads.

The resultant VARS are 0, so our voltage regulator is happy trying to regulate PF=1. The currents are fairly well balanced on true RMS meters, but unstable as heck. PF meter was quickly jumping all over from lagging to leading, from .9 to .2.

The volts before sync were very close. (284, 283, 287 L-N, A, B, C respectively) so I don't suspect a voltage unbalance causing the problem.

The local utility load is highly residential, old, and only has over current protection, no VAR or PF metering. They have single phase L-N capacitor banks spread throughout the radial fed and they swear they are all OK. Radial feed at the substation was about 240A @ 12.47Kv, so our gen is a drop in the bucket.

They are running a 12.47KV Y solidly grounded system, most loads connected line to neutral. The xmfr feeding the gen is a Y-Y wound 12.47K to 480v. The gen is Y of course, solidly grounded. Yea, not a good situation, but the utility says they won’t change. I suggested to customer they float the neutral at the gen since they don’t really need it.

I am concerned about the harmonics, (Y-Y xmfrs have high third harmonics, as you know) so they are going to get a full analyzer out there soon. They had other problems (breaker tripping) so we could not load it up past about 60%. The unbalance seemed to level out (lower % unbalance) at the higher power, but it was still pretty unstable. I’ll bet we had some high reactive circulating currents in the gen.

I would greatly appreciate your feedback. Have you ever seen a system with 2 phases lagging and one leading? With the current phase angles that far off? Any thoughts on the PF stability?

I can understand how they have the different PF on the different phases: I suspect a cap bank got stuck with too much KVAR being added to the system. (Utility swears this is not the case.) But why the unstable readings and why the severe current unbalance? The genset CB tripped on % unbal showing 868A on B phase and 1680A on C. (it did not capture A phase). I suspect a bad trip unit (AC Pro retrofit kit) so they are replacing that today.

But I was standing right in front of the CB when it tripped, and there was NO WAY it opened under that kind of load. I have been doing this for 30 years and when a IEEE 1558 CB (an old GE AK-100) opens under that kind of load, you know it!

Below is some data from the meter captured well before the trip (at low load). The voltage angles are right on. CT's are 1500:5.


Meter Configuration
Meter Time:
2/19/2009 3:42:03 PM
3 Element 3 Phase 4-Wire WYE

Transformer Factor
1
Line Frequency
60
Hardware Options
NA



Toolbox Information


Measurements

Va Angle 0.000
Vb Angle 119.000
Vc Angle 240.300

Va 285.150 V
Vb 284.325 V
Vc 286.500 V

Ia Angle 47.800
Ib Angle 145.700
Ic Angle 214.900

Ia 0.240 A
Ib 0.372 A
Ic 0.268 A

kW 0.153
kVA 0.160
kVar 0.047
PF 0.956

Thanks, in advance, for your comments,
Steve

 

[marks1080]

Possible partial fault in the sync breaker?

 

[dpc]

Does this happen as soon as you synchronize?

Can you check for harmonics? This connection can have circulating triplen harmonics.

Cap banks on the utility side?



"An 'expert' is someone who has made every possible mistake in a very narrow field of study." -- Edward Teller

 

[BSEE1980]

We used the same CB for the load banks and got to 100% load just fine, so my first guess is the CB is OK.
But those are pure signals and the CB was tripping on % current unbalance, so we have replaced the CB trip unit. We have it replaced now and are waiting to start up again.

The unstability and unbalance start as soon as the gen CB is closed, yes. It gets a little worse over time and at light loads. At higher loads it seems to get better. The longest it has run is about 5 min.

Yes there are single phase cap banks throught the radial feeder. Just found out that their timers are set to 1 min: that is, they can change only at one min intervals.

We don't have equipment on site now for harmonics measurement, but plan on checking that as soon as we can get a power system anaylzer.

Thanks for your feedback...keep you posted.
Steve

 

[rbulsara]

A couple of more thoughts:

Fistly what does the gen field tech says? (may be you are the one?)

Need to check the basics first. Is the AVR getting proper signal (contact closure) to say the gen is paralleled to the utility?

Could be circulating current due to wrong pitch? Most US utility supplies don't like anything other than 2/3rd pitch windings.

Or the gen is trying to motor? (less likely)

CT and PT wiring to load share module? Any CT backward??

 

[BSEE1980]

Yea, we are helping the end user (our customer) so you could say we are acting as the gen field tech. Acting is the important word here!

AVR is getting the correct analog signal. It drives the volts correct in order to sync (util volts higher than nominal setting, so it raises correctly) plus we can change the PF setting from .95 (under excited) to .95 (over excited) it drives the volts.

Our 32 setting is pretty tight, so I doubt it is trying to motor, but we will check that. Thank you.

We have 3 sets of metering (1 at gen, 1 at tie CB, 1 in utility metering cabinet) and they all agree. Plus that was the first thing I checked. We spent 6 hours ringing out every CT and PT wire, checking polarities, etc and everything sure looked OK.

Thanks
Steve

 

[waross]

What are the generator voltages and the utility voltages BEFORE you close the breaker?
Does the utility use line to neutral voltage regulators on the system?
Single phase regulators may cause neutral displacement and unbalance phase to phase voltages.
Can you check the utility phase angles before you connect?
If the utility voltages are unbalanced you may end up drawing VARs from one phase while delivering VARs to the other phases. That will give you the strange phase angles and the unbalanced currents.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[BSEE1980]

I think you are right on track, Bill.

The voltages are very close (+/- 2 volts) before we close.

Utility voltage regulation is strickly by cap banks, we think. They were not sure if the 12.47Kv xmfr has a step V regulator (load tap changer) or not.

We don't know the system PF/KVAR at the tie point, no, and no way to check. (no CT's there) Of course, there is no gen PF until we sync.

I'm leaning toward a cap bank problem. They regulate volts only with caps on the assumption that they need to add VARS to raise the volts. But if they are already leading, adding VARS will only reduce the volts. So one of the volt regs got upside down, I think.

Thanks for your feedback,
Steve

 

[racobb]

I know you said you have already checked the instrument transformer connections, but...I was asked to do what you seem to be going through just last summer. Client was having almost identical issues on parallel. I was assured that the metering transformers had been checked, so I had the utility lock down the voltage regulators, even took cap banks off line, but with no help. Voltage balance was good. Started looking at the Volts and Currents sequence data and realized there was something up! Had them revisit the CT connections (third time I was told) and low and behold, they had not only connected to the wrong CT's but they were mis-wired as well. Since they had connected to the wrong units, the ratio was also wrong in the switchgear PLC(check the controller as well).

Also make sure that one or more of the CT's is not mounted backwards and a PT is not connected backwards. Sounds as though you have, but just wanted to share what I had to go through to get the answer.

My units were also going out on reverse power and sounds like yours are not.

And as someone here has already said, make sure the generator knows when it needs to be in droop mode and when to be isochronous.

Will be interested to hear what you find to be the problem!

Alan Cobb

 

[alehman]

I suppose it's possible leading load pf is pushing your generator outside of it's capability curve and loosing voltage regulation. Any loss of excitation alarms?

 

[jghrist]

Generators have a very low zero-sequence impedance relative to the utility system. Because of this, the generators will carry more than their share of system unbalanced load when connected with wye-wye transformers. Load drop compensation on 1Ø feeder voltage regulators may create voltage unbalance in response to the unbalanced load sharing.

 

[rbulsara]

plus we can change the PF setting from .95 (under excited) to .95 (over excited) it drives the volts.

Why overecxite??? What you may want to try is leave the power factor mode (on AVR) to fixed VAR mode so gen is always exporting some VARs. I have seen more issues leaving it on so called "auto" mode and some times even in PF mode that is trying keep PF constant because a slight unstable load will drive the gen controls crazy to keep PF constant.

 

[BSEE1980]

Great feedback, all, thank you!
I'll try to answer your questions.

Alan: I think I will ask them to get a phase angle meter and re-check PT's and CT's. Couldn't hurt. Electrical contractor was not experienced in switchgear work, either.

Aleman: No loss of excitation alarms. We could be outside the capability curve on one phase, if we trust the current unbalance readings.

Jghrist: I did not know that about unbalanced load on Y-Y xmfrs, thank you. Opening the neutral at the gen might help. That is on our list to try.

Rbulsara: Thanks for the suggestion and I forgot to say that we did try it in a manual mode and we still had the unbalance current problem.
Gens almost always run over-excited, what you (and I) call it exporting vars. As you know, if you run under excited you risk slipping a pole and damaging the gen. (consuming vars, you might call it)

I use the terms over-excited and under-excited because one's view of lagging and leading varies if you are a utility or an industrial. Utilities say they are supplying lagging vars (over-excited) to compensate for the lagging vars in the system while industrial plants say they supply leading vars (when over-excited) to make up for the lagging vars inherent in the system.

When I used to work for GE, I would go to power plants and their PF meters had "LAG" to the right of 1 and that is where they ran all the time. Raising excitation current increased lagging vars.
In industrial plants or on motors, the PF meters had "LEAD" to the right of 1. Raising the excitation current increased the leading vars. (By the way, they were the same meter with different face plates, but we charged utilities more!! )
Thus started my education into leading vs lagging. But the one thing that is consistent in an motor or generator, is that if you raise excitation you increase vars. So I started calling it over-excited and under-excited. Now if you call those lagging or leading vars is up to you!

Technically we are not over-excited because the gen is rated at .8 leading, (see, I call it leading too!) but when our excitation is increased over 1 PF I call it over-excited.


Thanks again, all, and I will keep you posted on what we find.

Steve

 

[dpc]

This issue of leading/lagging for generators has come up several times in the past. For the record:

A generator that is overexcited (producing vars) is lagging. Underexcited (consuming vars) is leading. This is IEEE standard terminology.

A look at the voltage and current phasors will confirm this.

A rule of thumb: When power and vars are flowing in the same direction: lagging. When power and vars are flowing in opposite directions: leading.





"An 'expert' is someone who has made every possible mistake in a very narrow field of study." -- Edward Teller

 

[racobb]

I went back to look at your phase angles and don't see how you have a 96% PF with Ia and Ib at 40+ deg. with respect to the voltage, unless the kvar is calculated incorrectly due to Ic being approx. 180 deg. opposite from Ia? Haven't done the math but that might be why is looks wierd.

Alan

 

[rbulsara]

dpc/BSEE:

By Bad. Yes, what I meant was exporting vars, lagging of for gen.

My prime suspects:
1. CT wiring- The metering CTs may not be the CTs that take part in paralleling functions.
2. Winding Pitch

 

[BSEE1980]

DSB: Technically, you are absolutely correct on var flow. But there are a lot of technical folks out there working with sync motors that think increasing excitation moves from lagging to leading because they are used to motors.

Alan: You hit the nail on the head! I tried vector math on the phasor diagram and the result is about zero KVAR, the defination of 1 PF.

We agree on the assumption of the cause. Now if I can only make it stable and get rid of the unbalance...

thanks
Steve

 

[waross]

If the generator voltage is 480V, 480V and 480V, and the line voltage is 478V, 478V and 482V, you may be simultaneously exporting and importing VARs. The lines importing VARs will be lagging and the line exporting VARs will be leading.
That would explain your strange phase angle.
Check your line voltages line to neutral and line to line.
I have seen interesting issues with wye distribution systems. Heavy unbalanced loads may cause local voltage drops in the neutral. A voltage drop in the neutral has to go somewhere and you end up with a neutral conductor that is electrically displaced from the geometric neutral. That gives phase shifts.


Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[subtech]

I'm not trying to over-simplify, but your numbers look to me like you've a cap (or two) stuck on C phase. Remember that even if the phase currents are equal, but the angles differ significantly from the normal 120 deg, you will have neutral current.
Overall, your pf may be near 1, but with the leading C phase, neutral current will be considerable.
I'd look down the line(s) for a bad oil or vacuum switch and get the leading pf on C phase to relent and you should be o.k.

 

[BSEE1980]

Good ideas, much appreciated.
I did not know that about unbalanced loads and displaced neutrals, waross, thank you. That makes sense. One of the things we need to pass on to the utility to check.

Subtech: yes, I agree with you. We have asked the utility to check, but I am not sure they would admit it if they found a problem. So we may never know for sure.

Update: the gen ran OK today at about 90%. Loads still a little unbalanced. Next step will be to bring the meter anaylzer guy back for another vector snapshot. And we are waiting to hear from the utility to see if their investigation turned up anything. I wonder if they fixed anything or it is just the way the system is today.

Keep you posted and many thanks for the technical discussion.
Steve