RISO (Resistance Isolation warning - 'translated from the foreign') is what the better inverters report when there is a connection to earth in the middle of the string of photodiodes on a panel ) In a panel with puncture damage, this is quite believable.

Panels are inherently current limited - you will never get more electrons per second than photons that land on the thing - in practice quite a lot less. But faulty panels are a bad thing, and the partial earthing means that things that should float wrt earth will not, and may well be capable of if not electrocuting someone, at least throwing them to the extent they drop off the roof, which  is just as serious. And the inverter will not be able to make use of their output.

Lay off the meggar - it will not take sensible readings and may damage good panels which are essentially strings of low voltage diodes each one generating half a volt or so open circuit.
A more useful test is to look for a voltage between wires and frame, with some sensible load - in the old days a test lamp, maybe now a resistor. Anything voltage line to frame that does not fall away to be just a few volts once a  load has been connected in parallel with the meter is a sure sign of a mid-string fault.

The only sensible thing to do is to all-pole isolate or  unplug the obviously damaged panels, and if you have to remove them in daylight, then you should reduce the output current available by covering  the active surface, and treat the wire tails and any exposed damaged panel regions where the seagulls have landed (*), as if they might be  live relative to the panel frame.

Tricky one, I do not envy you.

Mike

* Seagull has landed sounds like a line from a Mike Myers movie - a low grade version of " The eagle has landed"

Thanks for the response. Helpful comment. The IR test was not possible due to the voltage between positive to earth and negative to earth (the structure) readings of 600-800V present on all of the 17 faulty strings. I'm trying to make sense of the risk factors at play. Adding additional parallel faults from the other 16 faulty strings to the mix only raises my concerns.

We are in the process of implementing access equipment to start disconnecting panels to reduce the overall voltage of the strings. 

Keen to understand the risks in this sort of scenario in more detail. It's a common fault after all.

Many thanks!  

Charles Yabsley said:

generally dissipates down to a level were we can test within 2-10 seconds. This is commonly known as a RISO issue, i'm aware

You will not easily get useable results from a typical Megger.  I think what you're seeing here is the capacitance of the array, amongst other things, and if normal and expected behaviour. This specific behaviour is not a fault.

As Mike says Riso is a symbol for insulation resistance, and an "RISO error" is generally the inverter reporting a low value. The inverters measure insulation resistance taking this and the operating voltage of the array into account. An Riso alarm does need to be addressed and should not be ignored. Bear in mind though that the inverter's measurement can be spoofed and also that an appropriate threshold varies with the physical size of the installation and environmental/installation conditions.

Diagnostic IR tests should be done by someone equipped (and skilled) with a suitable PV test device. Not least these testers are designed to operate with the array voltage on. Personally I find the HT Instruments PV-ISOTEST quite well equipped; it can carry out PI and DAR tests as well as simple IR and in another configuration can help locate faults within a string; I've not used a Seaward device for a while to be fair so can't comment on their more recent offerings.

On the other hand, it does sound like there are some major faults here that don't need an IR test.

mapj1 said:

Panels are inherently current limited -

This they are. On the other hand it is worth noting that under short circuit conditions, tens of amps arising from a P-N short do still pose a considerable fire risk, particularly if this is via faulty connections etc.

Charles Yabsley said:

Even with the inverter DC isolators in the open position (Off), on all 22 strings, is there a risk to life or a chance of fire with the system in its current condition.

Yes.

Given the condition you describe, there may also be current flowing in conductors even if the isolators at the inverter are off, due to parallel paths. These paths are likely high resistance and hence may well get hot. The presence, or absence, of these should be identified urgently.

Also, by the sound of it, there's a risk that someone has cheaped out / been done over by cowboys, so it's probably worth verifying that the isolators/switch disconnectors are appropriate for DC on load switching.

I agree breaking down the installation and isolating damaged sections is advisable. However, as well as covering up the array do take precautions (which might getting a DC clamp meter on conductors prior to working and using insulated equipment and gloves, tools and arc-rated PPE) when making any disconnections as you might be trying to break substantial DC current.

If it's really bad, with many interpole faults that can't be easily found and isolated, I might consider making a deliberate, temporary short circuits that can be broken safely (e.g. linking P-N downstream of a suitable switch disconnector) so that that there's negligible volts across the array and current through (hopefully relevatively high impedance) fault paths is minimal. This would be an emergency procedure to prevent fire while the installation is being broken down, not a long term answer, and could make things worse if there are series faults / resistances (e.g. dodgy connectors) so only to use after your own assessment - I cannot see the installation. That said if the faults are limited and the array (or section thereof) can be successfully open-circuited, that will usually be safer.

Charles Yabsley said:

Does the building structure pose a risk of electrocution?  (nearest building 35m). Trucks and vehicles are operating within the building 12 hours a day.

Sounds like a ground mounted array? Without seeing the installation, if the inverters are disconnected (so there's no AC ripple present) I'd suggest this is unlikely for general users of the building unless they're able to get between two parts of the array somehow. As you approach the array the risk assessment will need to be considered carefully.

Don't forget that Riso testing only identfies faults where there is a return path. Cables with faulty insulation dangling in free air may not be detected.

mapj1 said:

Tricky one, I do not envy you.

Nor do I. Good luck!

Do you mean Fauna (animals)- or Flora (plants) ?

Neither is good, but if things are nesting under the panels that may bite, peck or sting when disturbed, and/or may have chewed the insulation off the string wires then that is another startle hazard waiting for anyone unwary going near.

Losing footing on wet green stuff is a different hazard, but again, at any sort of height, even the apparently comedy risks become far more serious.

And you will need a large amount of tarpaulin, or arrange for the sun to be turned down or off .i.e. evening working  Once sun is no longer striking panels directly, the output will be single figures % of the 'full sun' value.- there is a secondary  reason that solar panel thefts occur at night...

Once separated the individual panels are not too bad but still avoid touching exposed copper or any damaged regions - the open circuit voltage will still be present down to very low light levels, even if there is no longer any great current behind it.

Mike.

Thank you for your thoughts, support and time on this.

The safety solutions and methods proposed fit in with our response. I like the idea of creating a short-circuit on the string, nice idea. 

Re IR testing: We have always been advised to test IR above VOC. We would generally consult the panel manufacturers for recommended test voltages, but this is has proven tricky. I'll take a look the meters suggested. The Seawood makes for a quicker test, but ultimately offers little extra test parameters than our meters and clamp testers. I'll look into the others you mentioned, although they look a little like the TIS meters... not a fan of their build quality.

My evolving thoughts on this:

The array posses a low risk or electrocution from ground level (unless someone is playing silly buggers with the inverters/DC isolators) 

The risk of electrocution at roof level (amongst the array) is high - Danger notice issued for this area (RAMS to reflect situation)

The risk of fire or short-circuit is high at roof level; the building does not hold any flammable materials or substances and has comprehensive fire procedures and detection on site (seagulls = canaries) - Danger notice issued surrounding this issue (RAMS to reflect situation)