Drone Thermal Inspections for Solar Arrays: Conditions, Standards, Limits, and Follow-Up
Thermal imaging from a drone can find hot cells, hot strings, and connection problems across a large array faster than walking every row. It is a screening and documentation tool, not a diagnosis: the image locates a temperature anomaly, and electrical testing on the ground confirms what it is. This guide covers the conditions that make thermography valid, the standard that defines the method, and the follow-up workflow that turns a thermal image into a repair decision.
What thermal inspection can detect
Under the right conditions, thermography reveals cells, substrings, and connections operating hotter or colder than their neighbors. Common findings include hot cells from mismatch or damage, hot bypass diodes, warm or cold strings pointing to string-level faults, and connection points with elevated resistance. The camera sees temperature differences; the pattern of those differences suggests the fault class. What it cannot do is measure electrical parameters: it cannot tell you the voltage of a string, the resistance of a connection, or whether a module is failing from PID, corrosion, or internal damage. Those require electrical testing on the ground.
Irradiance and weather conditions
Thermal contrast only exists while the array is producing meaningful power, so inspections are scheduled at high and stable irradiance — typically near solar noon on clear days — rather than in the morning, under clouds, or in the late afternoon. Wind cools modules unevenly and can mask small temperature differences, so low-wind conditions are preferred. Reflections from nearby surfaces, off-angle viewing, and uneven module loading (dirt, snow, vegetation) all distort the image; the flight plan and the interpretation must account for them. The equipment records irradiance, wind, ambient temperature, and viewing geometry alongside the thermal data so the inspection can be repeated under comparable conditions.
RGB and radiometric data
Visual (RGB) images are captured at the same time as thermal frames so every anomaly can be located on the module and compared with visible damage. Radiometric thermal data stores a temperature value at every pixel and can be re-analyzed after the flight; non-radiometric video cannot. The inspection report should pair radiometric stills with RGB stills, record the emissivity setting used, and note the temperature scale so a reviewer can verify the analysis. Without radiometric data, a thermal image is evidence that something was warm — not a measurement of how warm.
The standard: IEC TS 62446-3
IEC TS 62446-3 defines the technical specification for outdoor thermographic inspection of PV systems: equipment requirements, measurement conditions, image capture, and reporting. Inspections claiming to follow the standard should state the equipment, conditions, and report format used, and the thermographer should be trained in both thermography and PV system behavior. A drone vendor’s marketing page is not a source for the standard; the standard itself is published by IEC and sold through national standards bodies.
Flight planning and operator requirements
Flight planning covers the flight path and altitude needed for the camera’s resolution to resolve individual cells, the viewing angles that avoid reflections, and the airspace and property rules that apply. In the United States, commercial drone operations fall under FAA Part 107: the pilot needs a remote pilot certificate, the flight needs to comply with airspace and altitude rules, and some sites require additional approvals. Outside the United States, the local civil aviation authority’s rules apply. Operator qualifications matter because thermography skill and drone skill are separate; a qualified drone pilot is not automatically a qualified thermographer.
Data interpretation and false positives
Thermal patterns are read against the array’s wiring diagram: a hot substring, a whole string, or a single cell each point to a different fault family. False positives are common: reflection of the sky or ground, uneven soiling, snow remnants, vegetation shadow, and cable or connector heat from normal operation can all look like faults. The interpretation should therefore classify each anomaly by confidence and list the ground checks that would confirm or clear it, rather than presenting every warm spot as a defect.
Defect-confirmation workflow
The follow-up workflow: (1) log each anomaly with module, string, and thermal and RGB images; (2) perform ground confirmation — DC voltage, current, and insulation-resistance checks, plus connector and junction-box inspection — on the flagged circuits; (3) record the electrical measurements and the confirmed root cause; (4) carry out the repair or replacement and re-test; (5) archive the before-and-after images and measurements with the repair record. An inspection is complete when the anomalies are either confirmed and remediated or cleared by testing; a report that ends at the image is a screening result, not a repair decision.
Residential versus utility-scale suitability
Thermal inspection pays for itself fastest on large arrays where walking every row is expensive, but it is also useful on residential and small commercial systems when a specific failure is suspected or after a damaging event such as hail. On small systems, a handheld thermal camera and an electrician’s multimeter often answer the question at lower cost than a drone flight. On utility-scale sites, drone thermography is a scheduled maintenance tool with defined intervals and reporting standards. The method should be chosen to fit the array size, access, and the question being asked.