Grid-tied to ESKOM (legally)

[phil.g00]

Does anyone know of anyone, when has successfully navigated the administrative sea to become legally grid-tied directly to ESKOM?

(With specific reference to small scale solar generation at LV level for own use (no-export), when the customer has his own dedicated MV/LV transformer connection).

 

 

[SilverNodashi]

For which municipality?

[phil.g00]

None, eskom direct

[anotherbrownbear]

Hi. To legally connect to Eskom you need to have a grid limiter to control the PV output so you don't feed energy back onto the grid. So self consumption only. Then you also need to do an application, attached is the application form and test sheet. You will also require a deadgrid safety lock, also see attached. 

 

The application document and testing document might be a little out of date but seems like it is still accepted.

 

The DGSL document is recently updated, so very current.

 

IPP_SMG_ApplFormRev4_20160718.pdf

vib. Eskom_SSEG_Dead Grid Safety Lock Specification_20181019.pdf

vib. Eskom_SSEG_Dead Grid Safety Lock Specification_20181019.pdf

[phil.g00]

@anotherbrownbear, Thank you very much for that documentation.

 

[phil.g00]

I find some very interesting issues raised by ESKOM's dead grid safety lock specification.

I want to compliment ESKOM ( Now, he's lost the plot!) and have a bit of a rant. Why?  Well, just because its the internet.

I want to rant about the bureaucrats in the city of Cape town members, and their  OTT requirement of a ROCOF relay.

The national laws have to be respected by ESKOM, so I think it is safe to assume their  dead grid safety lock specification is legal.

I want to say further the actual isolation required is reasonable. By isolation, I mean the quality of the physical separation of two electrical systems as required by the law.

My point is, this document considers over-voltage, under-voltage and frequency conditions, and at the end of the day it settles on the single requirement of under-voltage protection only as being the measure necessary.

The requirement of the quality of the physical separation of the two electrical systems is still required and being met.

This makes sense, and defines in essence what the CoCT is required to demand off you in terms of upholding the law. It is the difference between a pen pusher's interpretation of the law and an engineer's.  

It represents a valid legal challenge to the CoCT's requirement of the all bells and whistles specification of the very expensive ROCOF relay.

When I mention the ROCOF relay, I am actually generalizing on all equivalently featured relays. Not because they aren't an excellent products, but because they are.

ROCOF stands for "Rate of Change of Frequency", probably the most well-known one deployed is the Ziehl relay.

They are far too fully featured. It is like using a 100 tonne press to crack a nut. 

ESKOM, quite rightly takes ownership of over-voltage conditions and frequency events. Yes, just like the CoCT they have to face up to their responsibility of being the big fish to your minnow in this scenario, and they do.

Those events, ESKOM acknowledges are not caused by your piffling solar panels, but by far larger events on their system.

It is ESKOM's responsibility to disconnect you,  it is not for you to go out and buy, commission and maintain specialist state of the art protection at a sub 20kVA level, to disconnect them. It is their responsibility to deploy this form of protection at suitable points in their system.

(I want to just mention at this point that to test the df/dt function of a ROCOF relay takes a test kit that cost more than 1/2 a million rand and as good as your local sign-off electrician is, he didn't test that function. The ESKOM relay will be tested because they have (had?) the resources to do it).

This is an unreasonable expectation and demand by the CoCT. 

The difference is what ESKOM demands is once they have disconnected your generation, they have understandably presented you with an easily detectable under-voltage condition, and they specify a minimum standard of automatic isolation of that generation from their dead system to effect repairs safely. This is an entirely reasonable demand in these circumstances.

So with regards to this ESKOM specification, (and I know I am going to have to suffer slings and arrows here), I say "Bravo", for writing a well-considered specification that is considered and practical.

To the CoCT, I think you are going to be facing many a legal challenge in the future, with your dictatorial heavy-handed approach.

Your specifications have been written bureaucrats, it shows,  - legally, and under the scrutiny of a technical argument they will be found to be unreasonable and undesirable.

The CoCT will have accept responsibility, that it is in fact the custodian of over-voltage and frequency events.

They will have answer the contradictions like:

Why if this level of protection is necessary at a sub 20kVA level? (which it isn't).

 For every one of their own MV /LV transformers at street level, equally representing a similar point of generation, why does there own spec not apply to themselves?  

Simple answer is because its unnecessary, too expensive and the wrong application of this type of protection deployed in the wrong place in the power system.

The power system stability relies on customers not disconnecting en masse willy-nilly all over the shop. Has the CoCT done the due diligence study?

Any legal defense the CoCT does raise, will have to be done in full knowledge, that their stipulation is contrary to the requirement of the national utility's (ESKOM's) considered opinion. (Who's the minnow now?)

The CoCT should and would lose that legal case, and then there would be a free-for-all with regards to all their prescriptive " computer says no " answers to  solar applications.

I am a power system protection specialist with 30+ years experience, but hopefully I have made this case from a reasonable layman's perspective. 

Rant over.

 

[___]

10 minutes ago, phil.g00 said:

This is an unreasonable expectation and demand by the CoCT. 

I have also noted in the past that the very steep requirements for anti-islanding might be overkill, and I've heard it not only here but elsewhere in the world too. So I'm not going to debate the reasonableness of the expectation, but I will debate whether the boogeyman is Cape Town. The only thing that is really required is compliance with NRS097-2-1 (that is a national thing), and I cannot remember if RoCoF is required, but on reading what I have at my disposal again (a bit old) it seems it isn't.

What is required is passive voltage and frequency monitoring (it merely has to be within the required range) combined with at least one active measure (such as attempting to shift the frequency, which with passive frequency monitoring should cause rapid disconnection).

[phil.g00]

1 hour ago, plonkster said:

What is required is passive voltage and frequency monitoring

Firstly, lets consider ESKOM's interpretation:

The safety issue is clear, no chance to fry the man doing repairs on the line.

No argument.

In that situation, no voltage = no frequency  by definition. 

And for good measure, there are are some defined specs around what levels of voltage and within what times, a specified quality of disconnection should take place.

Again, absolutely no problem with that, because now we are talking about the very real safety concerns surrounding the ability to liven plant.

1 hour ago, plonkster said:

combined with at least one active measure (such as attempting to shift the frequency

I don't know how ESKOM interpret's as the other active measure, I believe they see it as the second fail-safe contactor. It is a second isolation break that is actively switched.

(Lawyers can debate this point).

For sure "attempting to shift the frequency" isn't it and I 'll explain why later.

Now, lets consider the COCT's interpretation of the same law.

All of the above, great - again no argument.

But with the responsibility of voltage and frequency control at domestic level.

I'd like to illustrate how absurd that actually is:

 I have a 10kVA load, but because of my 3kVA of panels I only require a  7kVA contribution from the grid

In terms of frequency and voltages, at a domestic level a 3kVA load and 3kVA generation have an equal and opposite (negligible), affect on the power system.

Ah! say the bureaucrats,  and latch on to the keyword "opposite" effect. 

But I go one further, switching on a 3kVA load and switching off a 3kVA load also has an equal and opposite(negligible), affect on the power system.

If there isn't a case for this relay when you switch on your kettle, there isn't a case for when you switch it off.

Once, the safety case for livening the system is dealt with, whether you generate 3KVA or whether you switch off your kettle is indistinguishable to the greater power system.

Couple this argument with, their own non-compliance, the practical inability for ROCOF systems to be commissioned and maintained, the practical impossibility of a domestic solar installation to influence the greater power system's voltage and frequency and the undesirable effect this would have on power system stability if it was adopted at a widespread national level.

There is no safety argument to support it, and it is an absurd requirement.

 

[___]

I don't know man. It seems rather clear to me that if I have a small 10kva inverter tied with a large 100kva transformer, a grid outage will turn that large transformer into a pretty large low-impedance load and the voltage drop will be undeniable. But that is only as long as I actually remain connected to the transformer.

What if somehow the LV line to my house is severed? Now it's flapping in the breeze and there is no current flowing at all (because it is severed)... and then someone from the council comes, turns off the power at the transformer (which goes unnoticed -- severed line), and when he touches the line he is electrocuted because I'm accidentally feeding it. It seems to me you simply cannot take for granted the conditions on the line that you tie with... and hence a small amount of monitoring is necessary. You can't just look at voltage... you might be seeing the voltage you yourself is making on a disconnected line, and then we're not even considering that there might be other embedded generators on that phase as well (maybe the house 4 doors down has one too). The regulations require that it should disconnect even if there are other generators too.

I'm not convinced that the residential city situation and an Eskom-on-a-farm-with-my-own-Transformer situation can be that easily compared, but this is not my area of expertise :-)

[phil.g00]

5 hours ago, plonkster said:

I'm not convinced that the residential city situation and an Eskom-on-a-farm-with-my-own-Transformer situation can be that easily compared

I concede that that is a valid argument.

And in fairness, ESKOM only considers a dedicated transformer scenario, so we are not comparing apples with apples.

Do the regulations draw a distinction between the two scenarios?

 

 

[___]

11 hours ago, phil.g00 said:

And in fairness, ESKOM only considers a dedicated transformer scenario, so we are not comparing apples with apples.

That does make it a LOT simpler, and I think you are quite correct, in such a situation checking for a low voltage is quite enough.

I suspect the complexities comes precisely from having multiple embedded generators (some that you might not know of) on the line, in other words, it comes from the distributed nature of the beast.

11 hours ago, phil.g00 said:

Do the regulations draw a distinction between the two scenarios?

I don't know, but maybe it is in the name:: small scale  embedded generator :-)

[phil.g00]

@plonkster, The devil is in the detail.

And also what is trying to be achieved by what equipment and for whom.

1. There are rules that apply to linesman at HV level ( that's the other side of the transformer). He must earth his plant, ( & he would even if there wasn't a law anyway).

2. On the LV side ( 400V & 230V), I would the practice is to test for dead, but earthing before working on it would probably be considered overkill.

I think that, linesman & electricians cannot be excused of basic safety precautions (legal requirements or not).

So for them, a "lock-out" system is important. Its "dead" and safe to work on and must stay that way.

Ok, that covers workers, but what about the public?

I'd argue that existing legislative precautions suffice, the public has no right to assume they wont get electrocuted, because a high % of they time they will and do, solar generation or not.

Sorry, public.

So that boils down to 3 basic safety features in my mind.

1. Killing the backfeed,  2 checking its dead, and 3 preventing it from being re-energized.

Of which you as a domestic IPP can aide in 1 & 3.

The ESKOM approach also specifies the requirement of an approved inverter, and they rely on that to perform step 1.

The worker would perform due diligence to check the plant is dead, step 2.

And their specified under-voltage lockout system performs step 3 and keeps it dead.

OK, if you go along with that logic, ( and if you don't, please say so), and its legally compliant,  it seems to me that the CoCT is interpreting the law in a way that independently duplicates the functionality of step 1 and step 3 . Whereas ESKOM seems interpret the two stage defense as 1 thing achieving step 1 and another thing achieving step 3.

But maybe, there are inconsistencies.

The Multi-Grid is CoCT compliant, what criteria  did it satisfy to achieve compliance as a single device?

[Gabriel_2018]

I can see this topic is interesting all over the world!!

[___]

1 hour ago, phil.g00 said:

The Multi-Grid is CoCT compliant, what criteria  did it satisfy to achieve compliance as a single device?

They sent it to a testing house (Primara) who made sure that it ticks the boxes required by NRS097-2-1. So I expect they would likely put it on equipment where they can simulate frequency and voltage levels quite easily and merely check that it disconnects within the required 2 seconds... and if it does, they sign it off. Additionally I would expect that they check the layout as well (do you have two disconnecting switches for example). I also heard there are some requirements for EMC.

[phil.g00]

My point is that the Multigrid differs from the rest of the range, not it what it measures but in the fact there are two disconnects.

A Quattro, for example combined with ESKOM's switch would achieve the same function, but it would not be approved in the CoCT.

What controls the second disconnect is the difference between ESKOM and the CoCT's interpretation of the same law.

The CoCT is seemingly insisting on a more elaborate system than the law stipulates.

[anotherbrownbear]

What type of relay is CoCT asking for? As far as I was concerned it is the same relay that is used for Eskom compliance. The only difference is the position and type of physical disconnection that is to be used. 

[___]

Cape Town technically only requires NRS097-2-1 compliance. At the moment I hear rumours that they actually reject applications that aren't on the list (which is wrong in my opinion, the list is a guide, something to make life easier for consumers).

I only have a very old copy of NRS097-2-1, but it requires:

And then the rest is about how that disconnection device should work, including that it should have two series switches (4.2.2.2.5, which is where the Multiplus fails), and that it is allowed for the disconnection device to be independent (aka the Ziehl).

And then lower down you get to the Anti-islanding part, which has this to say:

Note that point 4.2.2.4.4 (which talks about RoCoF, it's cut off) is for synchronous generators, not for inverters.

Also note that 4.2.2.2.4 requires redundant monitoring... a single fault in your monitoring relay should not lead to a failure to disconnect. This pretty much rules out most frequency and voltage monitoring relays (such as this), and that is how you end up with the Ziehl (which has such a backup setup).

Edited December 15, 2018 by plonkster

[anotherbrownbear]

Ok, so that is for the inverter itself. CoCT and Eskom require an additional disconnecting device when connecting to LV network. I'm not sure what the minimum size of the plant is for CoCT to require this additional device but on Eskom it is for every plant. 

This additional protection basically does what any compliant inverter already does inside. But because of all the inverters that are being installed that do not have the relevant certification,  they have brought out a blanket rule. Added to the these plants are also sites that have decent equipment installed but not done in the right way. Like leaving out main disconnects and/or AC protection. 

[phil.g00]

On 2018/12/15 at 12:03 PM, plonkster said:

They sent it to a testing house (Primara) who made sure that it ticks the boxes required by NRS097-2-1

I think this is the problem, and explains the CoCT approach.

The law require "X" performance and "Y" quality of isolation from your connection. ESKOM understands this can be achieved through a combination of system components.

The testing house tests particular component in isolation of the context of a system. In other words if one component doesn't tick all boxes  it doesn't get certified.

I'll bet the Ziehl relay was submitted in something like Victron's  pre-configured double contactor box, and therefore it is seen to do it all.

Then the CoCT clerks insist on certification at component level as being decisive, over an above the judgement of system level certification by an authorized competent person.

[___]

2 hours ago, phil.g00 said:

Then the CoCT clerks insist on certification at component level as being decisive, over an above the judgement of system level certification by an authorized competent person.

And like you I do not like that one bit. That is nonsense. What I hope -- and surely sanity should prevail -- is that they will accept the word of a professional engineer over their own list.

Lists usually also have legal ramifications. Long ago (I'm talking around 2004) there were two ways that medical aids determined their levels of cover. The one was the BHF rate list (board of healthcare funders). The other was the SAMA rate list (South African Medical Association). Both were abolished because it was deemed to be a kind of price-fixing, and it got replaced by the NHRPL, the National Healthcare Recommended Price List (around 2009)... which ALSO got abolished because it was deemed to be a kind of price-fixing. Now... this list of recommended inverters... isn't that also a kind of fixing that benefits some setups and discriminates against other (arguably better) setups? I mean if the list is an aid that helps people select a compliant inverter, that is one thing... but if the list becomes the law unto itself, that seems to me something that could potentially be challenged on legal grounds.

Edited December 17, 2018 by plonkster

[sfdt]

On 2018/12/11 at 5:21 AM, phil.g00 said:

Does anyone know of anyone, when has successfully navigated the administrative sea to become legally grid-tied directly to ESKOM?

(With specific reference to small scale solar generation at LV level for own use (no-export), when the customer has his own dedicated MV/LV transformer connection).

 

 

A year later - has anyone successfully (or not) applied to become legally grid-tied directly to Eskom?

[Rautenk]

On 2018/12/13 at 6:33 PM, anotherbrownbear said:

The DGSL document is recently updated, so very current.

I hope @anotherbrownbear is still active or perhaps ANYONE can help here....

Where do one procure a Dead Grid Safety Lock? Considering it will become standard in SANS10142-1-2 (Wiring code for EGs) that will be accepted in our lifetime.
Now the problem is that DGSL was envisioned by Dr. Henri in Eskom, the value is definitely there to use this... BUT has anyone actually manufactured one and listed it on the Eskom side?