I have a scatter plot from a fairly large dataset of a solar array output.
Generated from measured V & I and allowing output power and impedance to be calculated and scatter plotted against each other
This data covers a vast array of different insolation and panel temperature conditions.
As you can see there is a clearly identifable correlatory curve between panel output power and output impedance. We haven't got a measure of insolator or temperature for this dataset, but those trends are intrinsically embedded in the data of course.
Now my question, if i repeat this plot in say 10 years time, what might i expect to see? Is it possible to crunch this data into some sort of panel State Of Health value? Anyone done anything similar?
For me it's interesting because it should be independant from panel cleanlyness as i think a dirty panel that therefore generates less energy over time (total integrated energy) is not the same as one that has a higher impedance, ie has aged? So if you just look at total integrated energy to determine SoH then you need to include changes in panel cleanliness and shading and insolation etc?
Your plot calculated from of the panel V, I or incremental internal resistance delat V/delta I?
It seems to be a simple arti=thmit fincetion of P, V, I , R and no relation to the panel life.
We have seen solar panel life depends on
Type of panel construction, quality of mfg eg cheap Chinese or Panasonic, Samsung, etc,
The solarisation degradation is between 3%/yr t 0.5%/yr for most panels depends on the above and the integrated solar intensity over the years.
Consult the panel mfg specs and the integrated solar intensity for your location.
https://en.wikipedia.org/wiki/Solar_irradianceJon
To measure the series resistance Rs, I think a panel would have to be tested in isolation, in the sun.
A dummy load with two steps could be used, The dummy load would need to overload the panel to a position well clear of the nose, tracking to the left along the constant current line. Then take V & I measurements at two I points , get delta_V, Delta_I hence Rs.
Refer to the standard diode model.
Hmmm. If you want to diagnose any 1-cell defect, a single working point won't cut it.
Recommending an proper I/V curve at a fixed power setting, esp. on the high side of the voltage.
Now my question, if i repeat this plot in say 10 years time, what might i expect to see? Is it possible to crunch this data into some sort of panel State Of Health value? Anyone done anything similar?
Similar? No but I had a project that went through a billion data point. We were cross referencing system generation with data from other systems in Belgium on the 15 minute scale and we were able to identify partial shading on panels.
You should be able to determine if a panel is faulty by looking at it with a thermal camera. Faulty cell (that 150x150 cell) will heat up more, because the rest of the string will push current through it.