Low insulation resistance - Utility scale PV plant 5

[123MB]

Hi all, hope you're well.

I have been tasked with determining the cause for low insulation resistance in a utility scale PV system (2.5MW per inverter, 255x10kW strings per inverter, tens of inverters, and ~1000VDC operating voltage). The inverters are transformerless and the PV field operates ungrounded at all times.

The insulation reduces significantly with increasing humidity and lower temperature (as is typical) - the reduction is significant and can bring the IR down from ~500kOhm to less than 5kOhm per inverter. The installation has some damaged connectors between the PV modules - some are cracked and moisture ingress occurs into some of the connectors.

I am trying to determine the chances that damaged connectors are the cause of the extremely low IR readings. The connectors are nowhere near any bonded metalwork (they are atleast 10cm away), so I am finding it hard to believe that there could be current flow between the connector and the bonding system - except for one possibility - if the connectors and cabling were coated with moisture then it is theoretically possible for current flow to occur from the contact inside the connector, through the moisture layer to the outside of the connector, and then continuing through the moisture layout on the outside of the connector and cable all the way back to the nearest support post for the cable which would be ~20cm away which is grounded metalwork.

This is the only way I could conceive of the damaged connectors causing this issue. Can anyone provide any assessment of the potential for the above 'moisture path' to cause the low IR readings observed? The voltage used for the IR test is approximately 1000VDC as you would expect.

 

[FacEngrPE]

I would expect the problem is a combination of moisture and dirt.
Along the Atlantic seaboard it is not unusual to have enough current across pole top insulators to cause pole fires, unless measures are taken. Often this includes washing. A 1000V megohmmeter will be measuring this surface contamination effect.

On a large PV plant there might be enough surface available for leakage current to notice.

A similar effect occurs with electric motors, which is why cleaning, bakingdrying, and re-varnishing is often effective for restoring motors with low insulation resistance.

 

[123MB]

Interesting, thank you.

On the back of your post I looked up pole fires and found that a mitigation is the application of silicone to the insulators.

If we were to apply a spray RTV silicone to the cable, between the connector and where the cable is attached to the bonding system, that should prevent any current flow and this should be discernible on the insulation resistance measurement.


Thoughts?

 

[FacEngrPE]

Potting the connectors in RTV after a good cleaning to remove all of the dirt should solve you problem, but the connectors will not be easily disassembled thereafter.

3M splice kits or similar could be used.
A shrink sleeve or electrical tape over a freshly cleaned connector might also work.

But depending on the need to be able to disassemble the connector, it might be better to replace the damaged connectors. This might be necessary to segment the string to make components reasonably safe to work on.

Please note my experience with power distribution DC is limited to 250 VDC, grounded system.

 

[123MB]

At this stage my aim is to prove or disprove the contribution of damaged connectors to the low IR, not to implement a fix. If we stop any current flow between the connector and ground, ie by applying a silicone to the cable, that should have the desired effect…

 

[FacEngrPE]

Washing the cables leading to the connectors with demineralized (low conductivity) water might be enough for the test you describe. - I suggest washing at night as the electric hazards will be lessened.

If you must wash during the day, please review references for washing energized power lines. thread238-193065
957-2005 - IEEE Guide for Cleaning Insulators

 

[EdStainless]

Wash with high purity water and re-test.
Then silicone them and check again.
If you don't add the silicone then they will deteriorate much faster.
The connector need replacement as a simple matter of upkeep.
The real question is how much ageing and dirt on the cables is causing more problems.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, consulting work welcomed

 

[Gr8blu]

123MB: there are two means of subverting the insulation resistance path. One is the "clearance" approach - also called the arc path. This is a matter of straight-line clearances between live conductors and ground planes - or between live conductors of different polarities and/or voltages. The second method is called the "creepage" approach, which is when the link is between the live conductor and something at a different potential via the conducting debris on the surface of the insulation.

Moisture, temperature, and altitude can all affect an insulation resistance measurement. The presence of cracked (or otherwise broken) insulation materials may result in lower measurements as there will now be a reduced "clearance" path and a reduced "creepage" path. Add in the presence of other surface contaminants such as dirt, leaves, moisture, oil mist, carbonized insulation, etc. and the "creepage" path is further reduced as the conductivity of the path itself increases.

You mention that you have damaged insulation in the neighborhood of the connectors. First step to long term reliability is to clean off the potential tracking (creepage) path with high purity water - although in the really short term, a simple wipe with a lint-free cloth might be sufficient to break up the path to get a better reading. Putting this cleaning into part of the ongoing maintenance cycle is imperative to prevent a recurrence. Second step is to either replace - or at least reseal - the errant connectors. Then inspect the remaining connectors and address their condition on an as-required basis ... fixing or replacing as you go. Finally, make the SYSTEM inspection part of the ongoing maintenance cycle: chances are good that this type of thing will occur again, given enough time and enough environmental exposure. To improve reliability over the much longer term, think about finding another connector and/or cable that stands up better to the actual environment where the equipment is installed.

Converting energy to motion for more than half a century

 

[jfpe]

We see a lot of problems caused by failing module connectors, but only when they are laying on racking or in a puddle.

Consider isolating strings within a combiner box and using a Seaward tester. This will give you Insulation resistance of the individual string (along with Voc and Isc). The Seaward let's you test the entire string including modules and home runs. Most module manufacturers don't want you to use a Megger on their modules. From there, you can see if it a problem with a small number of individual strings, many individual strings, or if every single string has a problem. Once you know, you can zero in more definitively. We typically see connector failures. We have also seen widespread individual module insulation resistance failures due to production problems. With either of these situations, you can split strings further and test individual modules. Also, we've seen sites where every single string shows low insulation resistance due to wire management, module quality, etc. Hopefully, you don't have this situation.


-JFPE

 

[123MB]

Thanks Jfpe,

In your last paragraph you say ‘due to wire management’… can you elaborate on this cause? What about the wire management?

 

[waross]

If you are reading 500kOhms to ground from each conductor, and assume 1000kOhms conductor to conductor, that's 1 Amp at 1000 Volts.
That will be 2 Amps if the measurement is conductor to conductor.
At 1000 VDC that will be between 1 kW and 2 kW losses.
From that we may conclude that the low resistance is distributed.
1 kW of losses concentrated in one faulty connector will shortly be evident.
My benchmark when assessing equipment with very low insulation resistance, combined with management insistence that it be energized is to calculate the losses and consider the losses as a fraction of the normal losses of the device.
Then make a judgment call, with suitable warnings to management.

--------------------
Ohm's law
Not just a good idea;
It's the LAW!

 

[jfpe]

For bad wire management problems on PV sites, I've seen home run wires cut by being dragged over a sharp edge in the racking during install. I've seen where the final torque on the modules and racking is done after the home run wires are in place and the home run wires get crushed. It may not create a hard ground fault, but you may see low insulation resistance. Over zealous tightening of stainless steel cable ties will create the same problem.


-JFPE

 

[123MB]

Thankyou waross and jfpe, much appreciated.

Waross, what you've said makes good sense, thankyou.

Jfpe, thanks for that - The site does have some issues that you've described present, so they definitely could be contributing.

 

[mgtrp]

If you are reading 500kOhms to ground from each conductor, and assume 1000kOhms conductor to conductor, that's 1 Amp at 1000 Volts.

I think there might be a decimal point in the wrong place with that calculation. 1000V / 1000kOhm = 1mA, not 1A. Which is a lot less heating in any problematic connector, and unlikely to give the type of unambiguous evidence of damage that 1-2 kW might.

 

[itsmoked]

1000V / 1000kOhm = 1mA, not 1A

Keith Cress
kcress -

http://www.flaminsystems.com

 

[itsmoked]

Have you tried a thermal imager? If any leakage causes even a half degree rise it should become obvious with an imager. While it may not 'see' all leakages cases you have it could quickly point out any that are larger which could help you focus on where to look at other spots more effectively.

Keith Cress
kcress -

http://www.flaminsystems.com