Corey, I have been studying IR diagnostics quite a lot, but I haven't had a whole lot of opportunity to "get my hands dirty" with it, so I'm not sure how much help I'll be. I would like to know how many modules are in each string.
I'm going to share some random thoughts on this, but nothing here is at all conclusive, and if anyone can correct some flaws in my logic, I'd love to hear it! It doesn't look like a simple, obvious problem. Your curve trace has a predicted Isc, with a dampening current to the knee, it has a sharp knee, then it has a low voltage at Vmp, and it stays really low all the way to Voc. A high series resistance would explain the dampening current to the knee, but it wouldn't explain the low voltage at all. If there was high resistance in the cells, you should see signs of diodes activating unless all of the cell groups had relatively uniform resistance. Generally, the hot cells in the checkerboard pattern are underproducing compared to the cooler cells. Because they can't pass on as much current, they are actually wasting some of the energy that the other cells are producing. This wasted energy turns into heat. The often times, they waste more energy than they are producing themselves. Alternatively, there could be other issues (or damage) at the hot cells. The fact that the checkerboard pattern goes away when you shut off the inverter may be helpful. If you short out a module, it will almost always look like a checkerboard in an IR shot, and if a diode shorts out, that third of the module will generally look like a checkerboard. This is the confusing part... Because the checkerboard appearance goes away, that would most likely indicate that it's not an issue of a short circuit failure (such as short-circuited diodes, or short-circuited modules). On the other hand, your curve trace looks somewhat like what you'd expect if you had some short-circuited diodes. It looks like you have a calculated Voc of around ≈860, and a measured Voc of around ≈725. That's about 84%. My initial first guess based on this curve trace alone would have been that perhaps roughly 16% of the cell groups might be shorted out. So if you had 20 modules per string, that would be 30 cell groups (assuming 3 diodes per cell group). With these hypotheticals, you would have somewhere around 5 shorted cell groups. Now that I'm looking at this closer, the knee of the curve is very quite sharp. I wonder if this could be caused by a higher resistance short somewhere out in the array (or in a conduit somewhere). (Perhaps this could be from compromised insulation or water intrusion somewhere??) Assuming that this isn't an intermittent problem, this would be easy to test under Voc. Just take a good DC amp clamp into the array while the inverter is shut down, and test the PV wires several places to try to find current in the array. Under this hypothesis, as the curve tracer starts pulling current, and the measured voltage falls from Voc, it gets to a point where suddenly, the fault isn't passing as much current. This would cause the sharp knee on the curve trace. As the voltage gets lower, the current at the fault would also get lower and the measured curve would increasingly get closer to the predicted curve. This model would also explain why the checkerboard phenomenon goes away when you turn off the inverter... it would be because the current going across the fault would be much lower than Isc, so when it's resting, there isn't enough current flowing to produce the checkerboard effect. Using this hypothesis, the checkerboard phenomenon might not be indicating a problem in the modules at all. If there is a high resistance short somewhere in the array, then that would cause the actual current in the modules to be above the predicted Imp because the current is being lost before it gets to the inverter. Because nearly all modules get the checkerboard phenomenon at Isc, it wouldn't be surprising if many normal modules would also get the phenomenon if they're operating between Imp and Isc. Again, take these thoughts with a grain of salt. If anyone sees somewhere that I'm going wrong, I'd love to hear from you! Hopefully, something here helps! Thanks, Kienan Maxfield Solar [email protected]<mailto:[email protected]> (801) 477-0-SUN (477-0786) (801) 631-5584 (Cell) ________________________________ From: RE-wrenches <[email protected]> on behalf of Corey Shalanski <[email protected]> Sent: Thursday, May 17, 2018 4:19 PM To: [email protected] Subject: [RE-wrenches] PV Cell Temperature Variation On a recent utility-scale PV installation we noticed a strange phenomenon on a single string on a single inverter. An IV curve (attached) reveals that the measured performance characteristics for this string (solid line) are significantly less than the expected values (dotted line). For reference, the ambient temperature was 31°C and the irradiance was 1010 W/m². The phenomenon that more so caught our attention is apparent on a thermal image (also attached) of the modules in this string. For lack of a better term I would describe the distribution of cell temperatures as resembling a "checkerboard" or "scattershot" (random) pattern, ranging between roughly 55°C and 70°C. Interestingly this phenomenon was only apparent while the inverter was operating, i.e. with the inverter turned off the modules revert to a much more uniform temperature distribution nearer to 55°C, instead varying by only a couple degrees across the entire module/string. There was no apparent physical damage to the modules. Can anyone offer any suggestions about what might be causing this phenomenon? -- Corey Shalanski Joule Energy New Orleans, LA
_______________________________________________ List sponsored by Redwood Alliance List Address: [email protected] Change listserver email address & settings: http://lists.re-wrenches.org/options.cgi/re-wrenches-re-wrenches.org List-Archive: http://www.mail-archive.com/[email protected]/maillist.html List rules & etiquette: www.re-wrenches.org/etiquette.htm Check out or update participant bios: www.members.re-wrenches.org
