IET
Decrease font size
Increase font size
Topic Title: Overvoltage issues "due to PV system"?
Topic Summary: Overloaded network or is there something I have missed?
Created On: 18 February 2013 06:58 PM
Status: Read Only
Linear : Threading : Single : Branch
Search Topic Search Topic
Topic Tools Topic Tools
View similar topics View similar topics
View topic in raw text format. Print this topic.
 18 February 2013 06:58 PM
User is offline View Users Profile Print this message



jammyc

Posts: 10
Joined: 25 November 2009

Hi all,

I was passed this project to resolve by one of our local teams, and after much going back and forth I'm not stuck so any guidance or suggestions appreciated.

The client has a 17kW (ac) PV system installed on a single phase in a rural network. This is odd enough to start with, but the team did get permission to connect from the DNO without any network upgrades or even a G59 relay.

All of the works were carried out in accordance with MIS 3002 (v2.1, being last year), with, crucially, the design voltage-rise calculated to less than 1%. The inverters are type-tested (Fronius) with settings based on a nominal voltage of 240V, so the lower overvoltage cutoff is at 264V.

Since then, the site has been suffering from overvoltage issues, with the inverter LoM protection nuisance tripping regularly. The DNO has been to site twice and taken recordings of the voltage at the cutout, each of a week with PV system both on and off. These showed that the voltage is just below ESQCR limits (253V) without the PV, and well over when it is on.

We have been back to site and checked and retested our installation, no problems found. The inverter is operating at pf=1.0. The DNO has been performed a voltage test at the local pole transformer, which showed L1=242-252V, L2=233-247V, and L3=236-246V.

I have been told that the site originally had a 50A cut-out fuse which was subsequently upgraded (prior to our works). Looking at the DNO's network plan the lines seem a tad small, though I've not got experience on their side of the fence (I can post details if neccessary). Furthermore, there are essentially no daytime loads between the PV and the transformer.

My assessment is that frankly the connection shouldn't have been allowed in the first place, without upgrading a line and/or converting to three-phase. But the proper procedure was followed and the DNO did give permission, and now it seems like the network impedance is too high, and, since there is minimal load between generator and transformer, there is no "headroom" left for the resulting voltage rise.

I am loathe to request permission from the DNO to increase the inverter overvoltage cutoffs beyond 240V as a) It obviously comes with the long-term risk of causing problems with other parts of the installation, b) It's not actually fixing the problem, and c) The customer would need to "sign-out" of the ESQCR limits, which should always be avoided.

My questions are:
1) Is there anything I may have missed that could be causing this problem from within the installation?
2) Do the required upgrades have to be undertaken at the client's cost, given that permission was originally given?

Thanks in advance,

James
 18 February 2013 11:55 PM
User is offline View Users Profile Print this message


Avatar for JonathanHill.
JonathanHill

Posts: 225
Joined: 09 September 2002

I'd agree that it's unusual for a DNO to accept 17kW (74A) as a G83 connection onto a single phase. To compound this, it sounds that you may be on L1, and therefore might get a better operating scenario if you could get the DNO to swing the premises onto L2 or L3.

One difficulty with PV is that, dependant on occupancy, the load during peak PV generation may be at a minimum. Are there any useful loads that could be switched on during peak PV generation - triggered either by monitoring generation currents or voltage?

I'd be interested to see further details if you want to PM me. I think the options of changing to 3-phase may provide a resolution, but could be costly (and you'd need a new inverter). As the voltage range measured is in the upper band (230V +10%, -6%), then the DNO could be asked to consider reducing the local transformer tap, but this may not be an option for them, dependant on the length of their LV circuits to other customers, load variations, etc, and would also necessitate a shutdown (as these tap switches can only be operated when the transformer is off-circuit).

Challenging, but not necessarily insoluble.

-------------------------
Jonno
 19 February 2013 05:58 PM
User is offline View Users Profile Print this message



jammyc

Posts: 10
Joined: 25 November 2009

Hi Jonno, thanks for the reply.

I put the option of swapping phases to the DNO a few weeks back, but they're not interested in doing it. Similarly with stepping down the 11kV transformer: Apparently their policy is to always leave MV trafos at the top tap setting and make adjustments at the 33kV primary substation, and obviously they're not going to go down a step for one small customer.

I'm not aware of loads that could be switched, but it's a working farm so definitely worth a look.

I'll PM you the local network map to see what you think.

I didn't ask yesterday, but I wonder if anyone has had any success in offering to the DNO to set the inverter output power factor away from unity to give a bit more voltage control. You lose a bit of yield, but I understand it's common on the Continent, albeit normally on utility-scale sites, and could possibly be an option.

Otherwise, I think you're right that a switch to three-phase will solve it... Who foots the bill given that it was permitted on one phase is another question!

James
 19 February 2013 11:02 PM
User is offline View Users Profile Print this message


Avatar for JonathanHill.
JonathanHill

Posts: 225
Joined: 09 September 2002

James, in my experience, UKPN are the good guys with wider vision than a number I could mention. I can see the docs you've sent but am finding them difficult to view on my phone. I'll have a better view on laptop tomorrow.

If changing to 3-phase my guess would be that you might negotiate a goodwill discount but your client will probably end up paying.

I agree that it's worth exploring your suggestion of looking at sucking VArs if the inverter can do this. The effectiveness of this will depends on the X/R ratio of the network. You might then have to consider getting 4 quadrant metering installed to demonstrate that the VAr demand is associated with generation and not demand, for which your customer may be penalised in their energy bills.

-------------------------
Jonno
 22 February 2013 02:12 PM
User is offline View Users Profile Print this message



944SE

Posts: 86
Joined: 17 May 2007

I'll go along with what has been said so far. I 'm amazed at them allowing the connection under G83. A couple of years ago I think the DNO would have looked VERY carefully at such a connection because of possible affects on adjacent consumers.
I think recently the DNO's have been overwhelmed with generation type connection applications and dare I suggest....may not be as thorough as they once were. The proposed relaxation on Inverter driven generation when G59 /2 is revised is further evidence that they cannot look at connection applications with the same scrutiny .
944SE
 22 February 2013 04:22 PM
User is offline View Users Profile Print this message



jammyc

Posts: 10
Joined: 25 November 2009

@ Jonno - Thanks for the advice. Did you get a chance to view the voltage graphs and network map I sent you?

@ 944SE - I know - I did a double-take when I was told it was single-phase too!
Going slightly off-topic, but where you mention G59/2 revisions, are you referring to the consultation on CENELEC CLC/FprTS 50549? I've got a public draft on my reading list but not had the time to even skim it yet. Or is there another revision in the pipeline?

James
 23 February 2013 09:25 AM
User is offline View Users Profile Print this message



944SE

Posts: 86
Joined: 17 May 2007

James, there is an ENA committee sitting at present to bring G59/2 settings into line with the new G83. As far as I can gleen they are a also looking at the use of "certified" inverters and removing? altogether the 50Kw limit whereby a separate G59 relay/site witnessed test is required....if certified inverters are used.
Recently I had an open letter from ENA . This mainly talks about need to change ROCOF settings from 0.125 Hz/sec to something like 1-2 Hz/sec and in some cases it may be desirable to do this retrospectively. [For many years we have avoided using ROCOF as we felt the existing setting were leading to nuisance tripping]
Regards 944SE
 24 February 2013 11:27 PM
User is offline View Users Profile Print this message


Avatar for JonathanHill.
JonathanHill

Posts: 225
Joined: 09 September 2002

James, sorry for going quiet on you - I've been full on these past few days. I also needed to get prints to get a better view.

Comments:
- significant periods during max generation with voltages >253 (stautory upper limit)

- Max generation only indicating 30 - 40A, with perhaps ~10A local load, whereas with 17kW, I'd expect ~74A -> 64A allowing for the load. Is the 17kW correct?

- the DNO's LV distribution seems quite extensive - do you have any voltage data from the extremeties. This might allow you to construct a case to pusj them on tapping down the transformer. However, from the traces, you'll need a 5% reduction to keep peak voltages <253V.

- Regular V & I spikes are indicative of motor starts - possibly chillers? Could be opportunity to keep these running during peak generation taking the temperature lower than "normal" thermostat, if the "produce" can tolerate the resultant temperature excursions. However, I don't think this alone will be enough to keep the volts <253V at the cut-out.

- Does the customer have any domestic/ commercial heating requirements (eg space heating or hot water) that may be currently fed from LPG or oil that could be transferred to operate at times of peak generation?

- It's not clear what DNO wiring mods would be equired to convert to 3-phase. Also, this will probably need a new PV inverter. You could make an application to UKPN to quantify.

- Option perhaps for a dedicated 11kV/ 400V 3-ph transformer dedicated to supply the premises. Again you could enquire of UKPN.

- There seems to be some grounds for getting the DNO to help out with changes to 3-phase if an application was properly submitted in the first instance. Bear in mind that the Connection Agreement will probably contain a clause that forbids the customer taking the network voltage outside statutory limits despite the fact that the DNO "allowed" this connection.

Also, to pursue your idea of getting the inverter to suck VArs, you could either try this and monitor, or model it. For the model, you'll need X/R + Fault Level data from the DNO and some load-flow software (and operator?) to do the modelling. It's possibe to do quick fag-packet calcs, or you could ask the DNO if they could run this for mutual benefit. If you decide to try it, then the DNO could get upset if the resultant power factor at the cut-out were to be worse than 0.95.

Sorry not to offer anything more definitive.

Do let us all know how you get on.

-------------------------
Jonno
 05 March 2013 12:01 PM
User is offline View Users Profile Print this message



jammyc

Posts: 10
Joined: 25 November 2009

Hi all. In case you're wondering, this hasn't gone dead yet: The DNO has asked to have (another) site visit before we can have further discussions. Once that's done we can talk about the suggestions made above.

I'll let you know how it goes.

Thanks again,
James
 08 March 2013 10:18 AM
User is offline View Users Profile Print this message



jammyc

Posts: 10
Joined: 25 November 2009

I called the DNO again yesteday and they've decided it's probably aggravated by an undersized transformer: As well as our issue, they've got low voltages at the far end too. As a result, the trafo is to be replaced in the next couple of months. Hopefully that will stabilise the voltage and give us the headroom we need.

They did also say they would be happy for us to fiddle with the power factor settings as a stopgap until it is in place. Bizarrely the operative I spoke to seemed to think it wouldn't have any effect on their network... We're going to speak to our client and if they want us to pursue it I'll get that in writing as no-one wants a surprise charge for PF coming down the line.

Thanks for the assistance,
James
 10 March 2013 06:22 PM
User is offline View Users Profile Print this message


Avatar for JonathanHill.
JonathanHill

Posts: 225
Joined: 09 September 2002

As well as the transformer, my guess is that a long LV feeder, which was adequate at the time of installation, is now causing an excessive voltage drop to the extreme customer(s) due to increases in load. It confirms the reasons for the DNO's reluctance to reduce taps

Increasing the transformer size will alleviate this to a degree. The improved voltage regulation at the transformer LV tails may help the voltage swing from the PV, but I think it'll be marginal.

The voltage reductions you may get from setting the PV inverter to suck VArs will depend on the X and R charachteristics of the network at that point of connection (the higher the X/R ratio, the better). It's the only possible solution I can see considering the other limiting factors (long LV feeder), unless the DNO consider splitting the LV network and installing an additional transformer.

There's probably a moral argument that, having accepted the level of generation in the first place, the DNO should give your customer the same consideration as they appear to be giving the customers suffering low volts. It would be an overall cheaper solution for that area of network if the DNO could consider both issues with a view to solving both together.

Please let us know how you get on.

-------------------------
Jonno
Statistics

See Also:



FuseTalk Standard Edition v3.2 - © 1999-2014 FuseTalk Inc. All rights reserved.