Application Case: Transformer Thermal Inspection on a Fixed Maintenance Route
Direct Answer & Executive Summary: An anonymized transformer-maintenance workflow for defining inspection distance, target coverage, lens selection, and radiometric reporting before a thermal-camera purchase.
Last updated: 2026-09-23
Engineering challenge: finding a small hot connection from an approved observation point
A transformer route may contain broad surfaces such as radiators and very small targets such as bushing terminals, tap connections, or cable lugs. An operator can easily take an attractive thermal image of the whole assembly while the connection that matters occupies only one or two detector pixels. The image then understates the apparent temperature of a small hot spot. This is a measurement geometry problem before it is a camera brand problem.
Consider a maintenance team planning routine energized inspections at a substation or industrial plant. Access and electrical approach boundaries are established by the site's qualified safety team. The buyer needs an instrument that resolves each specified target from those approved positions and retains enough context to compare the same asset on a later visit. This is an industry workflow, not a report of a fault found at a named customer site.
1. Optical Distance and resolution on the inspection route
For every view, the route sheet records the asset identifier, target feature, estimated target width, closest approved viewing point, likely obstructions, load condition, and whether the operator needs a screening image or a defensible temperature measurement. It also records whether the target has a painted, oxidized, or reflective surface. An infrared camera cannot reveal a connection hidden behind an opaque closed enclosure, so the inspection plan must state what is actually visible.
The simplest first-pass geometry uses horizontal field of view divided by horizontal detector pixels to estimate angular pixel size. For a 640-pixel-wide detector, a 24-degree lens is roughly 0.65 mrad per pixel; a 14-degree lens is roughly 0.38 mrad; a 6-degree lens is roughly 0.16 mrad. Multiply by distance to estimate the projected pixel footprint. These are illustrative calculations from nominal fields of view, not the measured IFOV specification of every camera-lens pair.
| Example inspection distance | Approximate pixel footprint with 24-degree lens | With 14-degree lens | With 6-degree lens | Implication for a 10 mm target |
|---|---|---|---|---|
| 5 m | 3.3 mm | 1.9 mm | 0.8 mm | A wide view may show several pixels across it; confirm focus and access |
| 15 m | 9.8 mm | 5.7 mm | 2.5 mm | A wide view is poor for target temperature; narrow optics improve coverage |
| 20 m | 13.1 mm | 7.6 mm | 3.3 mm | Even a 6-degree view provides only about three pixels across the target |
| 30 m | 19.6 mm | 11.5 mm | 4.9 mm | A 10 mm feature remains challenging; change view or target definition |
Three pixels across a target can be a screening starting point, but accurate temperature measurement may need more target fill depending on the instrument's measurement spot, focus, and calibration. A 10 mm bolt at 20 m is therefore not automatically measurable merely because a datasheet says 640 x 480. If the route cannot provide sufficient coverage, the team must revise the observation point, lens, target size, or inspection objective. Digital zoom cannot recover missing detector detail.
2. Control the conditions that make thermal comparisons meaningful
Electrical load is central. Compare similar components under comparable load, and record phase current or other available load evidence with the image. A cool-looking connection at low load is not proof that the same connection would remain cool under normal operation. Wind, rain, sun exposure, and background reflections can also alter apparent temperature. The route card should prompt the operator to record these conditions rather than conceal them in a generic temperature threshold.
Highly reflective metal surfaces are especially troublesome. Emissivity settings and reflected apparent temperature compensation should be checked against the actual target material. Porcelain, painted steel, oxidized metal, copper, and aluminum cannot safely share one assumed emissivity. Where a reliable absolute temperature is unavailable, a comparison with adjacent like-for-like connections at similar load may still be useful, but the report must preserve the method and its limits.
Thermal sensitivity, often expressed as NETD, helps describe the camera's ability to distinguish small temperature differences under specified laboratory conditions. It does not correct a target that is undersampled, out of focus, reflected, or viewed through an obstruction. The buyer should treat NETD, lens, detector size, focus behavior, and radiometric storage as a combined system.
3. Select equipment with the actual route in hand
The FLIR T865 is a 640 x 480 platform with several interchangeable field-of-view options; the manufacturer's published configurations include 24-, 14-, and 6-degree lenses. HIKMICRO's G61 is a Chinese-brand 640 x 480 alternative with lens and reporting options to confirm for the local market. Neither camera should be awarded the job from the detector count alone. Request the exact lens part number, its field of view, published IFOV, minimum focus distance, expected battery endurance, and supported radiometric export.
Ask the supplier to photograph a representative connection at the farthest approved viewing point. Retain the original radiometric file, visible reference image, focus setting, distance estimate, and a report exported by the software that will actually be delivered. If a second lens is needed for radiator overviews and distant terminals, ask how the route names the lens used for each image. A change of lens between visits must be visible in the record.
The procurement evaluation should also include calibration and service: how the camera is checked, what a lens change does to the measurement setup, whether route templates can be exported, and whether operators can retrieve old images by asset ID. A camera that makes one sharp demonstration image but cannot support an audit trail may be the wrong purchase for a maintenance team.
4. Close the loop from observation to maintenance action
A useful route report stores asset ID, image file, date, load, environmental notes, view location, lens, focus, emissivity and reflected-temperature assumptions, target region, and a reference image. It marks a suspect area for qualified electrical review, rather than asserting that one apparent hot pixel proves a defective component. After a repair or load change, the team repeats the approved view so that the follow-up can be compared with the original conditions.
For the camera-side selection checks, consult the switchgear thermal imaging guide. Its lens, IFOV, radiometry, and report questions apply to transformer routes as well, although the access geometry differs.
Questions the buyer should ask
- What is the smallest decision target at the farthest approved viewing position?
- Does the proposed lens provide sufficient detector pixels on that target in a real demonstration?
- Which surfaces will require emissivity or reflected-temperature handling?
- Can the operator retain and export radiometric files with asset and load context?
- Who defines safe access and the maintenance response when a suspect area is found?
Sources for technical review: FLIR T865 lens and imaging specifications and HIKMICRO G61 specifications.
*Disclaimer & Scope Note: This is an anonymized inspection-planning scenario with illustrative optical calculations. Site electrical safety rules and qualified personnel determine permitted observation positions; no specific fault or maintenance result is claimed.*
Long-Range Optical Resolution Verification
Transformer bushing and radiator inspections at extended working distances ( m - 10 m) require narrow angular IFOV. Verify your imager's spot size and × 3 MFOV threshold using our Thermal Camera FOV, IFOV & 3x3 MFOV Calculator.
