Aerial thermal image of a commercial rooftop array, scattered bright anomalies across the modules

Solar Thermal Fault Diagnosis: Why Shape Matters More Than Temperature

Almost every solar thermographic report you will ever be shown leads with hot spots. Bright marks on a dark array, a temperature figure beside each one, a severity label. It looks like a finding. Very often it is not.

A hot spot is a symptom. It is not a diagnosis. The same temperature rise, on the same array, on the same afternoon, can be produced by at least four entirely different faults. One of them is a safety matter. One costs you a measurable slice of annual yield. One is recoverable if it is caught early enough and permanent if it is not. And one is not a fault at all, but the module working exactly as designed.

Reported as a number alone, all four look the same. They attract the same severity band, the same red marker, and frequently the same recommendation, which is usually some version of investigate further. That is how an owner ends up paying to send someone onto a roof to look at a junction box that was never faulty.

What separates them is not how hot the anomaly is. It is what shape it is, where it sits, and whether it repeats. This piece is about reading that.

Blog Content TL;DR...

Why the temperature figure on a solar thermal report is only half the answer.

  • A hot spot is a symptom, not a diagnosis. The same temperature rise can come from four different faults with four different costs.
  • IEC 62446-3 severity is measured against healthy modules in the same conditions, not against ambient air, and it sets the band rather than the cause.
  • Symmetrical patterns mean electrical conditions at module or system level. A tidy shape is a circuit doing something.
  • Asymmetrical patterns mean localised cell or component damage. Cracks, shunts and debris are irregular and do not repeat.
  • Warm junction boxes in consistent positions across many modules are normal, and reporting them as faults sends people onto roofs for nothing.

The first question on a thermogram is not how hot it is. It is what shape it is, and whether it repeats.

CAA Certified and Insured

Subject

Solar PV thermography, fault morphology, and IEC 62446-3 anomaly classification on commercial arrays

Post Tags

Solar PV, IEC 62446-3, Thermography, Bypass Diode, PID, Microcracks, Fault Diagnosis, O&M

Skills Applied

Qualitative Thermal Analysis, Solar Loading Analysis, IEC 62446-3 Solar Thermographic Inspection

Author:

Steve Fisher, ITC Level III Certified Master Thermographer (Infrared Training Centre; Certification #205722059). Building, solar and industrial thermographic survey, analysis and reporting for Drone Media Imaging.

What the Number Does and Does Not Tell You

Severity is measured. Cause is diagnosed. They are not the same job

Aerial thermal image of a commercial rooftop array, scattered bright anomalies across the modules
Aerial thermal image of a commercial rooftop array, scattered bright anomalies across the modules

What the temperature figure actually tells you

Under IEC 62446-3, the temperature difference between an anomaly and its local reference sets the severity band. That reference is not the ambient air. It is the working temperature of comparable healthy modules in the same conditions, so a panel at 55°C on a hot afternoon may be entirely normal while the same reading in March would not be.

The bands run from advisory, at one to four degrees above the reference, through to critical above twenty. A cell-level anomaly beyond forty degrees is a mandatory safety trigger regardless of anything else. Findings below four degrees are recorded as observations rather than classified anomalies, but they are recorded, because an early-stage fault is worth knowing about before it becomes an expensive one.

Severity is not cause

Here is the part that gets skipped. The band tells you how big the thermal event is. It tells you nothing at all about what is producing it.

A twelve degree rise above reference could be a single cell shunting, a bypass diode conducting across a third of the module, an early potential induced degradation gradient, or a connector degrading in the string. Same number, same band, four different faults, four different costs and four different things to do about them.

Which is why Drone Media Imaging assigns two things to every classified anomaly: the IEC severity band, and separately a consequence classification of safety, yield or degradation trajectory. The first is measured. The second is judged, and it is the one that tells an owner whether to act this week or at the next service visit.

what does a solar panel hotspot mean, bypass diode thermal signature, IEC 62446-3 severity bands explainedwhat does a solar panel hotspot mean, bypass diode thermal signature, IEC 62446-3 severity bands explained

Symmetry, and What It Gives Away

The rule that does most of the work

There is one distinction that separates most solar thermal faults before any temperature is considered at all.

Symmetrical patterns point to electrical conditions at system or module level. Asymmetrical patterns point to localised defects in the cells or components.

It sounds almost too simple to be worth stating. In practice it is the single most useful discriminator in the discipline, because the two families have completely different causes, different urgencies and different people who need to fix them.

Why symmetry means electrical

A module or a string is a series circuit. When something interrupts or bypasses part of that circuit, the affected section heats evenly, because the same current is being dissipated across the whole of it. That produces clean geometry: a whole module uniformly warm, a whole string uniformly warm, a neat rectangle covering exactly a third or a half of a module.

Cell-level damage has no such discipline. A microcrack follows the fracture. A shunt sits where the manufacturing defect is. Debris sits where it landed. The result is irregular, and it does not repeat across neighbouring modules, because nothing systematic caused it.

So the first question on any thermogram is not how hot. It is whether the shape is tidy, and whether the same shape appears somewhere else on the array. A repeating pattern across many modules on the same inverter input is telling you about the inverter, not about the modules, and no amount of module-level investigation will find it.

Reading Five Common Patterns

Four readings, and what each one means

A neat rectangle covering a third or a half of a module. That is a bypass diode conducting. It may be doing its job, protecting shaded cells, or it may have failed short and be conducting permanently. The distinction matters, because one is the system working and the other is a module producing a fraction of its rated output and heating a junction box while it does so.

A gradient across a module with the cells nearest the frame warmest and the centre coolest. That is the signature of potential induced degradation, driven by system voltage between the cells and the grounded frame. Caught early it is substantially recoverable through reverse-bias treatment. Left until the encapsulant has chemically altered, it is not.

A single cell hot, in a linear, diagonal or star-shaped mark. That is a microcrack disrupting the metallisation on the cell. A rounded, amorphous single-cell mark in the same position means something different, a shunt rather than a fracture, and it behaves differently over time.

A whole string uniformly elevated. Nothing is wrong with those modules. Something upstream has disconnected them, and the fault is in the wiring, a connector or the inverter, not in the panels being photographed.

The one that is not a fault at all

Junction boxes run warm. That is normal. When the warm junction boxes appear in consistent positions across many modules, that is a module design characteristic and not a finding.

An isolated junction box hot on one module while its neighbours are cool is a genuine connector or internal diode issue. The difference between those two readings is whether anyone gets sent onto the roof, and it is decided entirely by whether the pattern repeats.

The Classification in Brief

IEC 62446-3 severity bands, and the second axis that sits alongside them
Thermal image showing a frame-edge temperature gradient across a module, the signature of PID

The reference. Severity is measured against the working temperature of comparable healthy modules in the same conditions, not against ambient air. Get the reference wrong and every band on the report is wrong with it.

The bands. One to four degrees above reference is advisory, a monitoring observation below the classification threshold. Four to ten is low to medium. Ten to twenty is medium to high. Above twenty is high or critical and warrants immediate investigation. A cell-level anomaly above forty degrees is a mandatory safety trigger.

The threshold. Findings at four degrees above reference and beyond are classified anomalies. Below that they are recorded as observations with a reassessment recommendation, rather than dropped.

Two axes, not one. The IEC band measures the thermal event. The consequence classification, safety, yield or degradation trajectory, states what it means for the asset. A medium band finding can carry a safety consequence where an absolute temperature trigger is met, and the two are assigned independently for every classified anomaly.

Governing Standards

  • IEC 62446-3:2017 specifies the thermographic method for photovoltaic systems, including the outdoor infrared inspection procedure and the anomaly classification framework referred to throughout this article.
  • The Drone Media Imaging Consequence Classification runs alongside the IEC severity band, expressing what a finding means for the asset as safety, yield or degradation trajectory. Both are assigned to every classified anomaly.
  • This is qualitative detection. Thermography identifies where a module or string is behaving abnormally. It does not measure electrical performance, and confirming a suspected electrical fault requires string-level testing under IEC 62446-1.
  • Findings are reported as observations warranting further investigation by a suitably qualified contractor. Drone Media Imaging carries out no remedial or corrective work of any kind.
  • Survey, analysis and reporting are conducted by Steve Fisher, ITC Level III Certified Master Thermographer (Infrared Training Centre; Certification #205722059), practising in accordance with the frameworks of ISO 18436-7:2014, ASNT SNT-TC-1A and ANSI/ASNT CP-105, and BINDT CMGEN Appendix B as UK training governance.
Make an Enquiry

Get a Second Reading on Your Last Report

If you have an inspection report full of hot spots and no clear account of what is causing them, that is worth a conversation. Tell us the array size, roughly when it was last inspected, and what the report recommended.

We also analyse and report on imagery captured by others. Where a Level 1 operator has flown the array, the thermograms can be reviewed and classified to Level 3 standard without anyone returning to site.

Make an Enquiry

related posts