7 signs your solar array
is underperforming.

The clearest signs of underperformance are a persistent gap from a defensible expectation, repeated equipment alarms, missing data, abnormal string or inverter comparisons, and a bill that no longer follows the facility’s load pattern. One low-production day is not enough to diagnose a fault: verify the data path, weather, grid status, and equipment state first.

Start with a defensible baseline

A production chart alone is not a diagnosis. Compare the affected period against irradiance or a validated model, then annotate planned shutdowns, utility outages, curtailment, clipping, construction activity, and meter changes. The NREL photovoltaic O&M guide emphasizes consistent operating data and failure definitions because ambiguous records make performance issues harder to manage.

For multi-inverter sites, peer comparison can be more useful than a portfolio average. Compare units with similar orientation, DC loading, and weather exposure. For a rooftop portfolio, compare production with the facility’s interval load so a billing change is not mistaken for an array fault.

The seven warning signs

1. Persistent production gap. Output remains below an appropriate baseline after accounting for weather, outages, and planned operating changes.
2. One inverter diverges. A peer unit shows materially different operating hours, faults, or AC output under comparable conditions.
3. Repeated restarts or derates. The same alarms return after resets or the equipment repeatedly reduces output.
4. Missing telemetry. The monitoring portal stops reporting even though equipment may still be operating.
5. String or combiner imbalance. One circuit departs from comparable circuits under similar conditions.
6. New physical warning signs. Hot spots, discoloration, damaged wiring, loose hardware, or enclosure problems require qualified assessment.
7. Bill-to-load mismatch. Imported and exported energy no longer follow the facility’s operating pattern.

None of these signs proves a root cause. Each is a reason to open a documented investigation. Preserve the date, time, equipment ID, alarm text, observed weather, safety condition, and raw production evidence before acknowledging alarms or changing firmware.

Rule out false alarms first

Confirm the meter and monitoring clock, communications uptime, sensor plausibility, and whether each value represents AC, DC, interval, or cumulative energy. A failed gateway, scaling error, or missing weather signal can create an apparent performance problem without an equipment failure.

Separate utility-driven states from plant states. California Rule 21 governs covered interconnection, operating, and metering requirements for generating facilities at CPUC-jurisdictional utilities. Protection or control changes should follow the approved configuration and the applicable utility process.

When to call an O&M provider

Escalate promptly when a safety device operates, an inverter cannot remain online, a visible electrical or structural condition exists, site-wide communications loss prevents operational control, or a production gap persists after basic data checks. Follow the site emergency plan for smoke, fire, exposed conductors, flood intrusion, or storm damage.

Prepare the evidence package before the call: ideally 12 months of interval production, recent alarms, inverter event logs, the weather source, a current single-line diagram, maintenance history, safe-location photos, and utility notices.

A five-step investigation sequence

1. Confirm the data. Check timestamps, meter scaling, communications, and the comparison period.
2. Classify the loss. Determine whether the evidence points toward DC, AC, facility-meter, grid, weather, or missing-data conditions.
3. Compare peers. Use similarly configured inverters, strings, or roofs rather than a broad portfolio average.
4. Inspect safely. Move from remote evidence to a focused field plan only when the information gap and safety prerequisites justify it.
5. Verify closure. Confirm the expected signal after corrective work and retain the result with any remaining limitation.

A closed work order should state what was observed, how it was tested, what was repaired, whether the result was verified, and what remains unresolved. If the operating window or weather prevents verification, record that limitation rather than declaring a completed restoration.

Records worth keeping

Maintain a current single-line and site plan, equipment inventory, serial-number map, commissioning baseline, utility notices, manufacturer manuals, maintenance logs, monitoring exports, alarm history, safety documents, and dated photos. These records shorten investigations and improve handoffs between owners, contractors, engineers, insurers, utilities, and warranty providers.