Aerial Accuracy Logo
Send a message

IEC TS 62446-3 Thermal Inspection Methodology

Guide to outdoor infrared thermography concepts for photovoltaic systems, including anomaly classification, severity assessment, and professional deliverable formats.

Quick Reference: Severity Thresholds

S1 - Monitor
ΔTn < 10°C
~15% module output loss. Schedule for next maintenance window.
S2 - Plan Repair
10°C ≤ ΔTn < 20°C
~30% module output loss. Schedule repair within 30-90 days.
S3 - Dispatch Now
ΔTn ≥ 20°C or safety risk
~60% module output loss. Immediate action required.

Anomaly Classes

String Outages

Entire string appears uniformly cool compared to neighbors. Zero or near-zero output from affected string.

Causes: Blown fuse, tripped breaker, disconnected DC cable, inverter fault
Thermal signature: Uniform cool string (all modules same temp)
Loss: 100% of string output

Bypass Diode Issues

1/3-module hot bands or step-changes across substrings. Diode conducting due to cell shading or failure.

Causes: Shorted diode, partial shading, cell cracking
Thermal signature: 1/3 of module hot (bypass activated)
Loss: 33% (one diode) or 66% (two diodes) per module

Nonuniform Thermal Signatures

Report frame-level nonuniform thermal signatures with capture context; pixel sampling is planning information, not individual-cell detection performance.

Candidate causes for ground testing: Cell micro-crack, PID, snail trails, delamination
Observed pattern: Localized nonuniform heating in the source frame
Impact: Qualified testing determines condition and operational impact

Wiring/Polarity Heating

Hot home-runs, connectors, combiner inputs, or reversed strings creating resistive heating in DC circuits.

Causes: Loose connector, corroded terminal, reversed polarity, undersized wire
Thermal signature: Linear heat along cable runs, point heat at connectors
Loss: Variable; safety hazard priority

Temperature Normalization (ΔTn)

Raw temperature differences (ΔT) vary with irradiance levels. A 10°C difference at 500 W/m² represents a more severe defect than 10°C at 1000 W/m². Normalization removes this bias for consistent classification across different measurement conditions.

Normalization Formula

ΔTn = (Tdefect − Tref) × (1000 W/m² ÷ Gmeas)
Tdefect

Radiometric temperature of the anomaly (°C)

Tref

Reference temperature from healthy modules (°C)

Gmeas

Measured irradiance at time of capture (W/m²)

Reference Temperature Selection

Tref should be the median temperature of 6-10 adjacent healthy modules in the same row and orientation. This accounts for:

  • Local ambient temperature variations
  • Wind cooling effects across the array
  • Module age and soiling consistency
  • Tracker position and sun angle

Worked Example

Illustrative inputs, not delivered field measurements:

  • Tdefect = 68°C (elevated area in the source frame)
  • Tref = 52°C (median of adjacent modules)
  • Gmeas = 860 W/m² (from site sensor data or a reference irradiance source)

Calculation:

ΔTn = (68 − 52) × (1000 ÷ 860) = 16 × 1.163 = 18.6°C

Classification in this example: S2 (Plan Repair) — 10°C ≤ 18.6°C < 20°C

Project reports use source imagery, capture context, and qualified ground evidence to prioritize investigation targets; they do not issue individual-cell or per-module absolute-temperature diagnoses.

Measurement Requirements & False-Positive Controls

Reliable defect detection requires controlled measurement conditions. Log these parameters for every finding to ensure defensible classifications.

Irradiance (G)

Preferred:≥700 W/m²
Minimum:≥600 W/m²

Irradiance verified via site sensor data or reference irradiance sources, logged with the flight record. Low irradiance reduces thermal contrast, masking defects.

Wind Speed

Recommended:≤5 m/s
Maximum:≤10 m/s

High wind convectively cools modules, reducing ΔT and masking hotspots. Record wind speed at time of capture.

View Angle

Required:≤30° from normal

Oblique angles reduce emissivity accuracy and cause reflections. Note module tilt/azimuth and sun angle for edge rows.

Image Quality

Focus: Sharp edges, no blur
No smear or ghosting artifacts
Adequate resolution (≥5×5 px/cell)

Reject frames with motion blur, focus issues, or lens condensation.

Repeatability Check

Re-observe suspected anomalies after 1-3 minutes to rule out transient effects (passing clouds, temporary shading, bird droppings). A repeatable thermal signature confirms the defect is real; transients disappear.

Capture → Evidence → Report Workflow

1Capture Set

Radiometric Thermal

R-JPEG or TIFF with full temperature matrix. EXIF/RTK GPS tags intact.

RGB Visual

Matching visible-light image for defect identification and context.

Flight Log

Per-photo GPS, yaw/pitch/roll, timestamp (UTC).

Irradiance Sources

Irradiance verified via site sensor data or reference irradiance sources, logged with the flight record. Attach source data with ambient temperature and wind.

2Evidence Packet (Per Finding)

  • • Primary frame (radiometric) + RGB context image
  • • ΔT map with defect box and reference box annotated
  • • Computed ΔTn with Gmeas value
  • • Repeat frame if used for transient validation

3Dispatch-Ready Report Bundle

PDF Report

Human-readable with thumbnails, callouts, and findings summary.

CSV Data

Row-level records for CMMS import and analysis.

KML/KMZ

Pin-drop locations for truck tablets and field navigation.

GeoTIFF (Optional)

Georeferenced thermal orthomosaic for GIS integration.

CSV Schema (CMMS Drop-In)

Standardized CSV format for direct import into maintenance management systems. All timestamps in UTC, coordinates in WGS84 decimal degrees.

site_name, area_id, investigation_target_id, timestamp_utc, apparent_temperature_context, capture_conditions, evidence_paths, notes

Location Fields

  • module_pos: e.g., "S47-M2847" or X/Y within string
  • tracker_id: Tracker serial or logical ID
  • row_id: Row number within array block

Classification Fields

  • anomaly_class: string_outage, bypass_diode, hotspot, wiring
  • subtype: cell_hotspot, bypass_1/3, connector, home_run, string_down
  • severity_bin: S1, S2, S3

Measurement Fields

  • deltaT_normalized_C_1000: ΔTn at standard 1000 W/m²
  • irradiance_Wm2: Actual irradiance at capture
  • view_angle_deg: Camera angle from module normal

Evidence Fields

  • image_thermal_path: Path to radiometric original
  • evidence_zip: Folder with originals & charts
  • notes: Free-text repair hints

PDF Report Structure

Fast to read on-site. Designed for field crews and asset managers.

1. Summary Dashboard

  • • Total findings count by anomaly class
  • • Breakdown by severity (S1/S2/S3)
  • • Estimated MW at risk
  • • Inspection date, conditions, coverage area

2. Map Page

  • • KML snapshot with cluster pins
  • • Color-coded by severity
  • • Array block labels for navigation

3. Investigation Targets (One Per Page)

Title: Validated area identifier
Location context: documented with positioning limits
Capture conditions: documented from available evidence
Thermal + RGB evidence imagery; screening notes; ground-verification priority.

KML/KMZ Styling (Crew-Friendly)

Styled for immediate field use on truck tablets and mobile devices.

S1
Small Yellow Pin
Monitor
S2
Orange Pin
Plan Repair
S3
Large Red Pin
Dispatch Now

Pop-up Fields

  • • Anomaly class and subtype
  • • ΔTn value and severity
  • • String/module IDs
  • • Thumbnail image
  • • "Open evidence folder" link

ID & Naming Conventions

Survey ID

SiteCode-YYMMDD-Run##

Example: HORNET-251223-R01

Finding ID

SiteCode-A####

Increment per finding. Mirror across PDF/CSV/KML.

Image Files

FindingID-THERM.jpg — Processed thermal
FindingID-RGB.jpg — Visual context
Keep radiometric originals (R-JPEG/TIFF) in evidence folder.

Field Math Quick Reference

Raw Temperature Difference

ΔT = T_defect − T_ref

Both values from radiometric image.

Normalized Temperature Difference

ΔTn = ΔT × (1000 / G)

G = measured irradiance in W/m².

Reference Temperature Selection

T_ref = median(T_1, T_2, ..., T_n) where n = 6-10

Select adjacent healthy modules in same row. Median reduces outlier influence.

Our Equipment & Documented Field Method

For full camera specifications (NETD, GSD at altitude, calibration schedule) and a comparison of our Autel 640T specifications and field procedures with the inspection topics discussed above, see our technical reference:

Documented Solar Thermal Methods

Commissioned deliverables can include radiometric data, GPS coordinates, severity classification, and CMMS-ready exports; the project scope defines the capture method and reporting package.