Uneven string current means strings that should match, on the same inverter or tracker, carry noticeably different currents. You can see it in string-level monitoring or in inverter string-failure detection, which compares each input with the mean or with the best string. Snow is one cause: when it only partly clears during melt or slide-off, clear cells and covered cells run at different currents. That loss lines up with snowfall and recovery periods and shows up on site cameras. Tracker twist is another cause: parallel strings on a twisted row no longer get the same irradiance. That loss depends on time of day and is worst at low sun angles, during backtracking and in wind. To confirm twist, compare the commanded tracker angle with angles measured along the row, and check row-end shadow lines in drone imagery. First, make sure the strings really are comparable. Normalize for different string sizes, and group inputs that share shading, alignment and module type. Then check the number of modules in the string, shading and module orientation. Do the current comparison under stable irradiance, or use a reference string or an IV curve test.
<!-- review notes
- sentence 3: support check: partial - Partial clearing mismatch and snowfall/recovery timing are stated, and camera observations appear only in the detection method, but not explicitly as the loss showing up on cameras; still largely supported.
-->
c2 drafted · lead
Low normalized row current
With tracker twist, the affected rows show lower irradiance-normalized current than neighbouring rows, even when the modules, inverters and sensors look healthy.
review note: support check: partial - Tracker twist lower normalized current than adjacent rows at low sun/backtracking is stated, but POA normalization and irradiance appearing normal are only loosely supported (irradiance sensors appear healthy).