Operation, Monitoring, and Maintenance
What you must understand
Operations is about comparing measured output to a defensible expectation. Know performance ratio, availability, degradation (plan on roughly 0.5 % per year for modern modules), and how weather-normalized models are used. Monitoring platforms report at string or module level; learn to distinguish a communications outage (no data) from a production outage (data showing zero).
Maintenance includes visual inspection, soiling and cleaning schedules, vegetation control, snow clearing, torque re-checks, thermal imaging, and firmware and warranty tracking. Every intervention belongs in a maintenance log with date, finding, action, and follow-up. Safety governs everything: de-energize, verify, and use PPE before touching conductors.
Traps that appear in questions
- Declaring equipment failure when the gateway simply lost network.
- Calling clipping a fault.
- Skipping inspections after a storm or hail event.
- Ignoring heavy soiling in dusty or agricultural areas, then blaming the inverter.
Worked example
A system was modeled at 5,000 kWh for the period and delivered 4,200 kWh - about 84 % of expectation. Before opening a service ticket, check irradiance for the period, soiling, new shading (trees, new construction), snow, and monitoring gaps. Only after those are excluded do you move to string-level electrical testing. Rule of thumb: investigate the cheap explanations first - weather, dirt, shade, and comms.