Impact of imbedded wind generation on distribution systems 2

[ters]

Even though wind farms are becoming common almost everywhere, they are mostly large scale connecting to HV transmission lines and appear much less common as imbedded distributed generation (DG) connecting to medium voltage distribution lines. Such distribution scenarios appear not to be well researched yet and available references such as IEEE P-1547 not particularly informative.

So, I’m working on a 10MW wind DG facility project connecting to a rural radial distribution line 44kV. The line originates from a 230/44kV substation. The 44kV line is about 40km long and the DG facility about 25km from the substation. There are residential loads about 5-10km downstream from the DG facility.

The DG facility is served by a switching station with a breaker and transfer trip (TT) scheme which interacts with upstream utility protections located at the substation and also a line recloser, which is about 10km upstream.

The teleprotection scheme uses a radio link. Upon detecting a fault, the utility will trip its own interrupting device (breaker or recloser) and also send a TT signal to the DG facility protection. Once TT signal is received and the DG facility breaker is tripped, a Distributed Generation End Open (DGEO) signal will be sent back to the utility, which enables their auto-recloser to attempt to restore the line.

This seems to work; however, the problem is the islanding condition which occurs when the utility interrupter is already tripped while the TT signal is still being processed before it trips the DG breaker, which may take about 100ms.

During this time consumers on the line will continue to be fed from the islanded DG, while the fault may or may not be isolated. Islanding condition may also occur when the utility interrupter is inadvertently opened and no TT signal is sent, line conductor broken far upstream, etc.

The specific problem is that according to the WT supplier, if an islanding condition occurs, the overvoltage which lasts several cycles may be over 2.0 pu before the turbine controller reacts and readjusts inverter output. Apparently, utility consumers still left on the line may be severely affected by a large overvoltage surge.

The WT supplier also states that such islanding condition leading to severe overvoltage problem should somehow be cleared within half of cycle which is technically impossible, unless the TT scheme is such that utility protections trip the DG facility first and then trip utility breaker or recloser.

This, however, is not acceptable to the utility as the tripping time for their instantaneous protections would probably reach some 200ms total, by the time TT signal is received at DG end, DG breaker open, and DGEO signal sent back.

Would anyone be able to provide some comments please.

 

[ters]

Forgot to mention, there are also protections at the DG side looking at utility line faults. But due to the nature of the WT generator (inverter) the DG contribution to the line fault is low. So using of distance protection is not possible and in this case we are using a directional overcurrent with load encroachment element. This however is neither quick nor reliable to detect remote faults on the utility line, so the probability that DG protections will detect the utility line faults and trip DG facility before the utility breaker is open is not very high.

 

[davidbeach]

First, IEEE-1547 hasn't been a 'P' since it was issued in 2003. The 'P' indicates a standard in development.

Second, how on earth do you get a 2pu overvoltage? The inverters can adjust their output in a fraction of a cycle if they see that they are causing a voltage rise.

 

[Skogsgurra]

It *could* adjust the voltage in a fraction of a cycle. But it does not. All voltage regulators that I know of in grid tied generators and inverters are quite slow. Stable operation would not be possible if they weren't.

There is also the problem with inverse energy flow. The normal distribution grid takes care of voltage drop by having taps set for normal load. If load goes to zero, the voltage increases. And more the further away from the infeed.

Now, if you connect wind turbines to the far ends and reverse energy flow, there will be unexpectedly large voltage swings. The grid has to be fundamentally different when you start connecting generators to it on a large scale. We have had the problem already but politicians are slow to understand and encourage wind turbines also where they do not belong.

Reality and theory differ sometimes.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[rbulsara]

That is why 'stability' studies for such analysis cost a few hundred thousand dollars!

Rafiq Bulsara

http://www.srengineersct.com

 

[ters]

"Second, how on earth do you get a 2pu overvoltage? "

I'm just quoting what the WT manufacturer indicated. If they used a coservative number, then I'm wrong. Actually, they didn’t say 2pu, but potentially exceeding 2pu, which is worse...

As for the cost of studies, in this renewable craziness, where the show is mostly driven by accountants and lawyers, I haven’t heard that anyone has done any such serious study for connecting to a distribution system, especially not one which did cost several hundred $.

“The inverters can adjust their output in a fraction of a cycle if they see that they are causing a voltage rise.”

May be so. But can that really be generalized to any sort of inverter or converter, regardless what technology and topology it employees and regardless of collector system configuration, number of units running, nature of the line load etc.?

 

[ters]

Gunnar, thank you for your comments. It is interesting that the utilities, probably under political pressure due huge importance of green energy, do allow connecting sometimes even 3 x 10MW to a long remote rural line. As long as the voltage remains within the allowable limits and the transformer at the substation is large enough to export that energy to the transmission side when there is no local load on the line, and if the line conductors are thick enough to curry the current, the connection is likely to be granted. Dynamic studies are not required.

 

[Marmite]

It seems to me that the windfarm protection as described is inadequate.Is the minimum load greater than the maximum generation? If it is then undervoltage and underfrequency protection at the wind farm should be sufficient. If the minimum load is less than the maximum generation then you should have additional loss of mains protection at the windfarm, rather than relying solely on an intertrip signal. Protection such as rate of change of frequency, change of voltage vector, reverse VAR, neutral displacement are all in widespread use on similar windfarm connections in the UK.
I've attached a paper which you may find of interest.
Regards
Marmite

 

[ters]

Mermite, thank you for comments and useful paper.

The minimum load is not greater than the maximum generation.

As for the DG protections, it is not only the inter-tripping scheme, but we are also using a multifunction relay which has many features including change of voltage vector.

However, none of the protections + breaker opening time (which is 5 cycles in the 60Hz world) is quick enough to detect islanding and open the DG breaker in no time, which seems to be the recommendation of the WT manufacturer (1/2 cycle is practically no time...).

The only solution seems to be that islanding never happens, i.e., that utility never trips before the DG facility does. Which in theory would be possible (excluding incidents such as open conductor), but in that case utility instantaneous protections at the remote end would likely mis-coordinate.

The only interrupting device I’m aware of which exists on the marker and which may be sufficiently quick is ABB IS limiter. But I’m not sure if it could be set to detect and operate on a rapid voltage rise, as it is an extremly fact current limiting device used in applications where breakers or fuses are too slow.

As Gunnar indicated, distribution systems don't seem to be designed to easily accommodate dispersed generation of substantial size (similar or larger in size than the line minimum load).

 

[davidbeach]

What ever. It's up to the folks proposing the interconnect to solve their own operational problems. I'd never approve an interconnection application where the applicant would ever force any voltage outside of ANSI range A under any condition. Throw it back in their lap.

 

[desrod]

We performed a number of "network impact studies" for utilities and I fully support David's comment.

As various technologies connect to the network it will be impossible to adapt the network to solve all their problems. There are a number of solutions on the generation side...might cost money but they should solve it.

Daniel

 

[ters]

The political reality (and pressure) called green energy craziness combined WT supply and demand seems to be different from what such studies may indicate and what manufactures can or will possibly do in a particular case. It generally seems, however, that in N. America experience with connecting large WTs to radial distribution lines is not widespread.

 

[rbulsara]

Yeah, it is frustrating when wind and electricity does not obey political demands!

Rafiq Bulsara

http://www.srengineersct.com

 

[rbulsara]

do not..

Rafiq Bulsara

http://www.srengineersct.com

 

[ters]

Thank you for constructive comments .

 

[rasevskii]

for Ters:

Can`t resist adding this here:

How does the WT supplier get a result of more than 2 PU voltage rise on load rejection? Is that some peak waveform value from the inverter output or an actual RMS amount? In a conventional synchronous generator with the best AVR there would be only a couple percent rise, but here we are dealing with some very fancy software based system that is proprietary and they wont tell you about it. Likely the WT agent doesent know himself, only what the factory gurus say over the phone.

Depending on the contract and your authority in the situation, they should be refused a connection until more is known, and the WT supplier has put in place some fast acting OV protection at their own expense, and properly type tested also.

rasevskii

 

[ters]

Rassevskii, nice to hear from you after quite some time.

WT manufacturer indicated that overvoltage of 2pu would be RMS lasting several cycles. They didn't elaborate much as why it may occur. Seemed to have something to do with Wind Turbine Ride Through which will attempt to boost the voltage, but if the fault is cleared by the action of utility remote protections only, and depending if the generation is higher than load the overvoltage may occur.

The utility has fairly comprehensive guidelines for the DG connections, and elaborate islanding as well, but seems that the case of dynamic (overvoltage or any other) conditions, when the utility is lost and the DG continues to island, are not considered in details.

 

[ters]

I posted this question more as an ilustration of unusual problems which are likely going to just get worse in the years to come, rather than that I expected some magic solution.

Don’t know where other posters are reporting from, but in the jurisdiction where I live things seem to be a bit different - the electricity grid is becoming more and more impacted by decisions made by politicians, accountant and lawyers, and there is an obvious pressure leading to compromised situations like this one.

On the other side, private investors building renewable DG facilities definitely don't chose equipment based on the best performance under various dynamic conditions (and supported by comprehensive studies) but rather price and availability.

Sometimes, the best wind turbine is one which you can get on time to meet your in-service date (which is in the contract and there are financial consequences in case of a delay).

Similarly, large WT manufactures, and there is only few out there, do not appear particularly accommodative, they all have enough business for now, so their logic seems to be more of: take it or leave it...

At the same time some larger utilities are overwhelmed with renewable applications and may not know or notice everything when it comes to assessment and their own studies they do for such connection.

Consultants, in most cases, also don’t seem to know everything about anything when it comes to various power converter technology connecting to medium voltage lines, but even if they did know everything, somebody usually needs to pay them to do comprehensive studies which in some cases simply doesn't happen.