Thermovision Camera for Transmission Line Inspection

Thermovision Camera for Transmission Line Inspection: Finding Hot Spots Before They Trip the Circuit

A compression joint that is 12 °C hotter than its neighbouring phase is not a cosmetic anomaly. It is a high-resistance crimp dissipating I²R heat into aluminium strands that lose tensile strength as they heat, and it will keep degrading until the joint anneals, the conductor strands break, or the line trips on a dropped phase. The joint gives no visible warning from the ground. A thermovision camera, correctly set up and correctly interpreted, gives that warning weeks or months early.

This article works through thermovision (infrared thermography) applied to overhead transmission lines: what the camera measures, how to configure it so the temperature reading is not fiction, how to correct for load and wind, how to classify what you find, and how drone-based inspection changes the workflow. The context is 132/230/400 kV practice of the kind run by Power Grid Bangladesh PLC (PGCB), but the physics and the standards apply anywhere.

Why Thermovision Matters on Bangladesh's Expanding Grid

PGCB's own network page lists, as of June 2026, roughly 3,282 circuit-km of 400 kV, 5,423 circuit-km of 230 kV and 9,687 circuit-km of 132 kV lines, plus 151 grid substations at 132/33 kV. That is more than 18,000 circuit-km of conductor, joints, clamps, and insulator strings, most of it running through wetlands, river crossings, and monsoon-exposed corridors where climbing every tower and lowering every joint for inspection is impractical.

Three factors make thermography a good fit here:

• Load growth. As lines are pushed closer to their thermal rating, marginal connections that were harmless at 40% loading become failure points at 90%.
• Humid, saline, monsoon environment. Aluminium-to-aluminium and aluminium-to-copper interfaces corrode; corroded interfaces raise contact resistance, which raises heat.
• Access. Char lands, haors and river crossings limit foot patrol. Aerial thermography reaches them without an outage.

ADB project documentation for the Dhaka and Western Zone Transmission Grid Expansion Project also refers to a Drone Inspection Center within PGCB's operation and maintenance department, so aerial inspection is already part of the institutional picture rather than a future proposal. (Confirm the current status and scope with PGCB before citing this in a proposal.)

How a Thermovision Camera Actually Measures Temperature

A thermal camera does not measure temperature directly. It measures infrared radiance in a long-wave band (typically 7.5–14 µm for uncooled microbolometer cameras) and converts it to an apparent temperature using calibration curves. Your inputs determine whether that number is right.

The radiometric measurement equation, in simplified form:

W_total = τ·ε·W_obj + τ·(1 − ε)·W_refl + (1 − τ)·W_atm

Where:

• W_obj = radiance of the target at its true temperature
• ε = emissivity of the target surface
• W_refl = radiance of the surroundings reflected from the target (for outdoor lines: the sky, the sun, nearby towers)
• τ = atmospheric transmittance over the path
• W_atm = radiance of the intervening atmosphere

ISO 18434-1:2008 (thermography — general procedures for condition monitoring of machines) covers the procedures for determining and compensating reflected apparent temperature, emissivity and attenuating media. It references ASTM E1862 (reflected temperature) and ASTM E1897 (transmittance) for the underlying test methods. ISO's page for 18434-1 shows a revision in progress, so check the current status before you quote it in a formal procedure.

The emissivity trap on bare conductors

Bare and weathered aluminium conductors are low-emissivity, highly reflective targets. Field values commonly fall well below the 0.95 that many cameras default to; treat 0.2–0.5 as a planning range and confirm on site with a reference (electrical tape patch on an accessible component, or a contact thermometer on a comparable dead-line sample). The consequence of getting this wrong is large.

Worked example: emissivity error on a hot conductor

A conductor surface is truly at 80 °C with ε = 0.30. The camera sees a clear sky with a reflected apparent temperature of −20 °C. Using the simplified total-band model (T in kelvin, atmospheric term neglected):

T_app⁴ = ε·T_obj⁴ + (1 − ε)·T_refl⁴
       = 0.30 × (353.15)⁴ + 0.70 × (253.15)⁴
T_app  ≈ 294.7 K ≈ 21.5 °C

If the operator leaves the camera at ε = 0.95, the camera back-calculates the object temperature from that 21.5 °C apparent value and reports about 23 °C, against a true 80 °C. The conductor looks cool. The hot spot is invisible in absolute terms.

Emissivity setting error on a bare aluminium conductor

The practical takeaways:

1. On low-emissivity metal, do not trust absolute temperature. Use comparative ΔT between identical components (phase A joint vs phase B joint) at the same viewing angle, distance and background.
2. Joint sleeves, clamp bodies with darker oxidised or painted surfaces, and connector hardware often have higher emissivity than the bare strand. These are the better measurement targets.
3. Avoid viewing angles that mirror the sun or bright sky into the lens. Shift your position or use the low-sun window.

Selecting the Right Camera: Specifications That Matter for Lines

Detection and measurement are different jobs. Any camera can show a bright spot; only a properly specified radiometric camera can put a defensible number on it at 30 m.

Table 1: Thermovision camera specification guide for transmission-line work

Parameter	Practical target	Why it matters on a line
Detector resolution	640 × 512 minimum for line work; higher for standoff	Pixels on target decide whether the joint temperature is measurable or just visible
Thermal sensitivity (NETD)	≤ 50 mK	Distinguishes small phase-to-phase differences on a healthy circuit
Spectral band	7.5–14 µm LWIR	Standard for electrical inspection; less affected by sun glint than mid-wave
Lens	Telephoto (50 mm or longer for standoff work)	Determines measurable spot size at distance
Radiometric output	Full per-pixel radiometric image files	Needed for post-flight ΔT analysis, not just a colour picture
Adjustable inputs	ε, reflected temp, distance, humidity, atmospheric temperature	Required for corrected measurements
Accuracy class	Typically ±2 °C or ±2% of reading	Quote it in reports; ΔT criteria must exceed measurement uncertainty
Paired visual camera	Yes	Lets you identify the component and its tower/span for the work order

Spot size and pixels on target

The number that catches out most new aerial thermographers is spatial resolution. A target must fill enough pixels for the camera to measure it, not just see it. A commonly used rule is a minimum of about 3 × 3 pixels for a reliable temperature reading; your camera's datasheet gives the exact measurement field of view, and that figure governs.

IFOV (mrad)         = pixel pitch / focal length
Min. measurable spot = N × IFOV × distance      (N = pixels across, e.g. 3)

Worked example: 640 × 512 detector, 12 µm pixel pitch

Table 2: Measurable spot size versus lens and standoff distance (3 × 3 pixel rule)

Lens	IFOV	Horizontal FOV	Min. spot @ 20 m	@ 30 m	@ 40 m
25 mm	0.48 mrad	17.5°	28.8 mm	43.2 mm	57.6 mm
50 mm	0.24 mrad	8.8°	14.4 mm	21.6 mm	28.8 mm
100 mm	0.12 mrad	4.4°	7.2 mm	10.8 mm	14.4 mm

A 25 mm lens at 30 m can measure a target of roughly 43 mm or larger. A sleeve or clamp body about 50 mm wide would only just qualify, and a 30 mm strand-break hot spot would not. Moving to a 50 mm lens halves the minimum spot at the same distance. For live 400 kV work where standoff is set by clearance rules rather than by what you would like, the telephoto lens is the correct answer, not flying closer.

Minimum measurable spot size versus distance for 25, 50 and 100 mm lenses

What to Look For: Fault Signatures on Overhead Lines

Where thermovision finds faults on an overhead line span

Table 3: Component-level thermal signatures on overhead lines

Component	Thermal signature	Probable cause	Follow-up
Compression joint / splice	Localised hot spot at or beside the sleeve, hotter than the same joint on the other phases	High contact resistance: poor crimp, strand corrosion, incomplete oxide removal	Priority replacement; check adjacent spans for the same crimping batch
Suspension or dead-end clamp	Hot clamp body or hot conductor exit	Loose bolts, corroded interface, bimetallic contact	Retorque or replace; inspect for strand damage
Jumper / bolted connection (dead-end tower)	Hot palm or bolt group	Loose or corroded bolted lug	Retorque to specification, clean, inhibit compound
Repair sleeve / armour rod	Warm or hot segment against bare conductor	Damaged strands under the repair	Inspect; consider full span replacement
Insulator string	Individual discs warmer or cooler than neighbours (ceramic/glass); warm section in polymer	Leakage current, contamination, internal defect, punctured disc	Follow with visual, UV corona or electrical test
Substation-side connections	Hot terminal pad, hot disconnector blade contact	Loose or corroded terminal, worn contact	Outage-window repair
Conductor mid-span	Uniformly warm segment	Localised strand damage or overloaded section	Compare against adjacent spans; check load

A note on cool anomalies: ASTM E1934 treats warm exceptions in electrical equipment as usually caused by increased resistance, loose or deteriorated connections, overloads, unbalanced loads or faulty components, and cool exceptions as usually caused by failed components. A cold disc in an insulator string is a diagnostic, not a reassurance.

Inspection Conditions: Getting Valid Data Outdoors

Thermovision line-inspection workflow

Load

Resistive heating scales with the square of current. A line at 10% load will show almost nothing even from a serious defect. NFPA 70B-2023 (Section 7.4) requires thermography to measure ΔT between similar components under similar loading and against ambient air, and to document the result. Field practice commonly quotes a minimum of about 40% of rated load for meaningful electrical thermography, but treat that as a guideline from practitioners and set the actual threshold in your own procedure. Record the line current, from SCADA or the line CT, at the moment of each image.

Wind

Wind removes heat from the hot spot and makes it look milder than it is. The Infrared Training Institute's wind compensation table gives the following multipliers to apply to the measured temperature rise:

Table 4: Wind speed correction factors (Infrared Training Institute)

Wind speed (m/s)	1 or less	2	3	4	5	6	7	8	9 or more
Correction factor	1.00	1.36	1.64	1.86	2.06	2.23	2.40	2.50	Not recommended

Above about 8 m/s, do not inspect. Bay of Bengal pre-monsoon and monsoon gusts on exposed coastal corridors often exceed that, so wind logging is a scheduling input, not an afterthought.

Sun, sky and rain

• Solar loading heats dark hardware and creates false hot spots. Solar reflection off shiny sleeves can produce false readings in the camera. The low-sun window (dawn, dusk) or overcast conditions give the cleanest ΔT.
• Rain and wet surfaces cool components unevenly and add reflective films. Do not inspect during or immediately after rain.
• Humidity and path length reduce atmospheric transmittance. Enter measured humidity and distance in the camera. At Bangladesh humidity levels and 30–40 m standoffs, this matters more than it would in an arid climate.

Load and Wind Extrapolation: Turning a Field Reading into a Risk Number

A hot spot measured at partial load and in a breeze understates what happens at peak. Two corrections are commonly applied:

ΔT_corrected = ΔT_measured × F_wind × (I_target / I_measured)ⁿ

• F_wind from Table 4
• n = 2 from I²R theory; empirical values around 1.6 are used where convective cooling dominates. Use both as a bracket.

Worked example: uncorrected versus corrected classification

A thermographer measures a joint on phase C that is 6 °C hotter than the identical joint on phase A. Measured line current is 240 A. Wind speed is 4 m/s. The circuit's planned peak is 600 A.

Step 1 – Wind correction:   6 °C × 1.86 = 11.2 °C
Step 2 – Load correction:
   n = 2.0:  11.2 × (600/240)²   = 11.2 × 6.25 ≈ 69.8 °C
   n = 1.6:  11.2 × (600/240)^1.6 = 11.2 × 4.33 ≈ 48.3 °C

The raw 6 °C falls in the "probable deficiency" band of the NETA similar-component scale. Corrected to peak conditions, the joint is projected to run tens of degrees hotter than its sibling, which puts it firmly in "repair immediately" territory.

Load extrapolation of a measured hot spot

Read this carefully before you apply it. Load- and wind-correction formulas are approximations, and the practitioner community is divided on them. The Snell Group's technical note on the subject points out that the surface temperature you image is often not the fault temperature at all, and that correction algorithms oversimplify the physics. The defensible use is: report the measured ΔT, report the corrected projection as a bracket labelled as an estimate, and let the work-order priority follow the conservative reading. Do not use a corrected number as a substitute for a measurement.

Joint heating in context

To see why the numbers are so sensitive, consider the heat generated in the joint itself:

P = I² × R_joint

At 600 A, a healthy joint of 40 µΩ dissipates 600² × 40×10⁻⁶ = 14.4 W. A degraded joint at 200 µΩ dissipates 72 W at the same current, five times as much, concentrated in a sleeve a few centimetres long. That is the energy that anneals aluminium strands and drives the failure to run-away.

Severity Classification: NETA ΔT Criteria

ANSI/NETA thermographic survey suggested actions

Table 5: Thermographic survey — suggested actions based on temperature rise (ANSI/NETA Table 100.18)

ΔT vs similar component, similar load	ΔT vs ambient air	Recommended action
1–3 °C	1–10 °C	Possible deficiency; warrants investigation
4–15 °C	11–20 °C	Probable deficiency; repair as time permits
—	21–40 °C	Monitor until corrective measures can be accomplished
> 15 °C	> 40 °C	Major discrepancy; repair immediately

Three points on using this table for transmission lines:

1. It was written for electrical equipment generally, and NETA's own note says heating varies with the square of the load current and that the values are guidelines in the absence of consensus standards. Utilities usually adopt it or a tightened variant into their own maintenance standards. Check whether PGCB, your client or your employer has a documented criterion and use that as the governing one.
2. Prefer similar-component comparison over ambient comparison. It cancels sun, ambient and much of the load effect.
3. Table 100.18 is quoted here from ANSI/NETA ATS documents. The current MTS edition (2023) is the maintenance standard; confirm the table numbering and wording in your copy before quoting it in a client report.

Drone Thermography: What Changes in the Workflow

Unmanned aircraft have moved thermography from occasional helicopter surveys to routine patrol. The trade-offs:

Table 6: Platform comparison for line thermography

Platform	Strengths	Limits
Ground handheld	Cheapest; good for substations and accessible towers	Poor view of conductor-level joints; parallax and reflection problems
Multirotor drone	Close angles on each phase; hover over hardware; georeferenced images	Battery endurance; wind sensitivity; standoff, airspace and permit constraints
Fixed-wing / VTOL	Long corridors per flight	Less agile at towers; requires larger payload margin
Helicopter	Long range; established practice	Highest cost; vibration; limited hover time
Fixed online sensors (substation)	Continuous monitoring	Limited to fixed points

Workflow points specific to aerial work:

• Detect versus measure. Industry practitioners stress that a thermal payload chosen for hot-spot detection may not have the optics or radiometric quality for temperature measurement. Specify which of the two you need in the tender.
• Pair thermal and visual images at every tower, geotagged, so a work order can identify tower, phase and component.
• Keep standoff. The drone holds position at a safe distance and uses optical reach to resolve the target rather than approaching the conductor. Some service providers publish fixed horizontal standoffs for lines above 33 kV; for your own procedure, derive clearances from the utility's live-line rules and the aircraft manufacturer's electromagnetic limits, not from a marketing page.
• Consistency. Fix altitude, speed, angle and camera settings across flights so that trend comparison between surveys is meaningful.
• Software assist. AI-based detection can screen large image sets quickly, but a qualified thermographer should confirm every flagged anomaly before a work order is raised.

Bangladesh permits: a real constraint on thermography timing

Drone operation in Bangladesh sits under the Civil Aviation Authority of Bangladesh (CAAB). Reporting in mid-2026 describes a draft Civil Aviation Rules 2026 released for consultation that would replace the 1984 rules and include a framework for commercial and non-commercial drone operation, and earlier draft Drone Regulations 2024 proposed licensing, registration and zone-based clearance (yellow and red zones needing CAAB permission and, in red zones, security clearance). Permit guides for 2026 also describe operations being limited to daylight hours under visual meteorological conditions.

That last point matters for thermography. The cleanest thermal window is often pre-dawn or dusk, when solar loading is low, and a daylight-only permit condition pushes you toward overcast periods and the early-morning slot just after sunrise. The rules are in draft and change often, so confirm current requirements directly with CAAB and your client's security and operations departments before planning flights.

Personnel and Reporting Standards

Thermography is an interpretation discipline, and a coloured image is not a finding.

Table 7: Standards and references for a thermography programme

Reference	Role	Status note
ISO 18434-1:2008	General procedures for thermography in condition monitoring, including emissivity, reflected temperature and attenuating media	Edition in force is 2008; a revision appears to be in progress, so check
ISO 18436-7:2014	Qualification and assessment of thermography personnel (three categories)	Reviewed and confirmed 2025 per ISO
ASTM E1934-99a (Reapproved 2024)	Guide for examining electrical and mechanical equipment with infrared thermography; defines end-user and thermographer responsibilities and report content	Active, reapproved 2024
NFPA 70B-2023 (Section 7.4)	Thermography requirements within electrical maintenance programmes; ΔT measurement and documentation	Now a standard rather than a recommended practice; written primarily for facility electrical equipment
ANSI/NETA MTS-2023	Maintenance testing specifications including thermographic survey criteria	Confirm current table wording
IEC 61472:2013 (+ COR1:2015)	Method for calculating minimum approach distances for live working, 72.5–800 kV	Governs live-line clearances, not drone clearances directly

ISO 18436-7 defines three certification categories. Category III personnel develop programmes, define procedures and acceptance criteria and supervise Categories I and II. A useful division of labour for a utility programme: a Category I or II thermographer acquires and screens the images, a Category III thermographer signs the criteria and the reports.

ASTM E1934 also cautions that an infrared examination gives data about equipment at the time of examination only and does not assure proper operation. The report should state the load, wind, ambient, emissivity and reflected-temperature settings so that the result can be reproduced.

Minimum report content for each finding:

• Tower or span ID, phase, component and GPS position
• Thermal and visual image pair with measurement points marked
• ΔT vs reference (and which reference), absolute temperature with the emissivity used
• Line current, ambient temperature, wind speed, humidity, distance, time of day and solar conditions
• Camera model, lens, calibration date and thermographer certification
• Classification, recommended action and reinspection date

Risk and Safety Instructions

Thermography of energised lines is low risk to the camera operator when standoff is respected and high risk when it is not.

• Never trade clearance for image quality. Use a longer lens, not a closer approach. Set drone standoff from the utility's live-line procedures, the manufacturer's stated limits and, for personnel-based live work near lines between 72.5 and 800 kV, the minimum approach distance principles in IEC 61472:2013. Induced fields near 400 kV can disturb compasses, GNSS and datalinks on small aircraft.
• Ground-crew position. Keep operators clear of the fall zone under the line, of vehicle-access hazards on embankments and of stepping onto unknown ground in flooded paddies. Do not stand under a conductor when working a handheld camera at a tower base.
• Substation work. Follow the permit-to-work system, keep to marked walkways, respect the site's arc-flash boundaries and do not remove covers or open panels on energised equipment to "get a better look." A thermographer's job is to image, not to intervene.
• Weather. Stop flights and ground surveys at the first lightning warning. Do not fly in rain. Wet insulators and wet hardware raise leakage and change the thermal pattern.
• Aircraft safety. Pre-flight checks, redundant return-to-home settings, spotter for the pilot, and a defined landing zone away from the corridor.
• After-finding discipline. Do not tell field crews to "tighten the bolts" on an energised joint. Repair happens under an outage or approved live-line method, followed by a reinspection to confirm the fix.
• Regulatory compliance. Confirm CAAB registration, operator licences, permits and insurance before every campaign.

Quick-Reference Decision Matrix

Situation	What to do	Watch out for
Line loaded below about 40% of rating	Defer the survey or record as a low-confidence baseline	Clean images at low load do not prove a healthy line
Bare conductor reads cool while a joint reads hot	Compare joints only, not conductor absolute temperature	Low ε makes absolute values unreliable
ΔT > 15 °C vs sibling phase	Treat as major discrepancy; plan urgent repair	Confirm with a second view angle first
ΔT 4–15 °C vs sibling phase	Probable deficiency; schedule repair and trend it	Apply wind and load projection as a bracket
Wind above 8 m/s	Do not survey	Corrections stop being credible
Pixels on target below about 3 × 3	Retake with a longer lens or closer angle within safe standoff	Reported temperatures will be too low
Individual insulator disc anomalous	Follow with visual and UV/electrical checks	Thermal alone does not confirm a puncture
Repair completed	Reinspect under similar load	Verification image is part of the record

Conclusion

Thermovision on transmission lines works when four things are true at once: the line is loaded enough to make a defect visible, the camera has enough pixels on the target, the emissivity and reflected-temperature settings reflect the actual metal and sky, and a qualified person classifies what is seen against a documented criterion. Skip any one of those and the survey produces images that look authoritative and mean little.

For PGCB-scale networks, where line-km is growing faster than patrol crews, drone thermography changes the economics. It brings phase-by-phase views of joints and clamps within reach of a routine programme. The gains come from disciplined method: standard flight profiles, radiometric payloads, load and weather logged with every image, conservative classification, and reinspection after repair. The regulatory environment for drones in Bangladesh is still moving, so build permit lead time and daylight-flight constraints into the inspection calendar from the beginning.

---

Have you run thermal patrols on 132 kV or 230 kV lines in Bangladesh, or compared drone and helicopter survey results? Share your experience in the comments — WAZIPOINT is building out a technical reference series on transmission and substation maintenance for South Asian field conditions.