Industry

Powerline & network

Find the fault before it finds you. Corona, thermal, vegetation and asset condition from one flight program.

Tell us the site and the deliverable. We'll name the model.

This job

Inspecting an electrical network

Network inspection is the recurring task of finding defects on assets that are energised, spread across long corridors, and expensive to reach.

Four separate things get looked for. Corona and partial discharge, which indicate insulation breakdown before it becomes a fault. Thermal anomalies at joints, clamps and conductors, which indicate resistance. Vegetation encroaching on clearance envelopes. And physical asset condition, from pole rot to damaged hardware.

Each one needs a different sensor. Most networks want all four.

What makes network inspection difficult

Scale and access. A distribution network is thousands of kilometres of corridor, much of it across private land, difficult terrain or road reserve. Ground inspection is slow and partial.

The defects are invisible. Corona does not show in visible light. A hot joint looks like a cold one. Vegetation clearance is a three-dimensional measurement that a photo cannot give you.

And the consequence of missing one is disproportionate. A defect that costs a few hundred dollars to fix becomes a fire, an outage and a regulatory event if it is not found.

How the day actually runs

Split the inspection by what you are looking for, because the sensors and the flight profiles are different.

Corona work runs slower and closer, with the UV payload looking for discharge signatures across the insulator string and hardware. It works in daylight, which is the point of a solar-blind system.

Thermal runs along the conductor and pauses at joints, clamps and terminations. Load matters: a joint under light load may not show the anomaly it will show at peak.

Vegetation runs as a LiDAR corridor capture, then gets processed against your clearance envelope rather than assessed by eye.

Asset condition rides along with the visual and zoom sensors on the same aircraft.

Substations and switchyards are a walking job. Handheld SLAM captures the enclosed structure that the aircraft cannot get inside.

What to fly, what to walk

What you are looking for	Capture with	Why
Corona, partial discharge	ZQP UV34B or UV10	Detects discharge in daylight, nothing else does
Thermal faults at joints	Zenmuse H30T or Matrice 4T	Radiometric thermal with optical zoom
Current measurement	CZI MF4 or MF40	Direct airborne measurement on conductor
Vegetation clearance	Zenmuse L3 on Matrice 400	LiDAR corridor, measured envelope
Asset condition, general patrol	Matrice 4T	Zoom, wide and thermal in one aircraft
Recurring corridor patrol	Dock 3	Scheduled flights, no crew travel
Substations, switchyards, plant rooms	SHARE C10-32	Enclosed, cluttered, no sky view

The network is an aerial job. The substation is a walking job.

What makes a finding defensible

Different sensors, different verification.

Thermal findings need a stated emissivity assumption, a reflected temperature, and a note of load at the time of capture. A temperature without those three is a number, not a finding.

Corona findings are qualitative by nature. The value is in detection and location, and in comparing the same asset over time.

Vegetation clearance is measured, not judged. LiDAR gives you distance from conductor to canopy against your defined envelope, which is auditable in a way that a photograph is not.

Whatever the sensor, geotag the finding to the asset ID. A defect you cannot locate on the network is a defect nobody will fix.

What gets handed over

A defect list tied to asset IDs, with imagery, severity and location, in the format the asset management system ingests.

The recipient is usually an asset manager or maintenance planner, not a pilot. They want to schedule work, not review flights. A folder of images is not a deliverable.

What gets it sent back: findings that cannot be located, thermal readings with no stated conditions, and severity ratings with no defined basis.

The common miss

Flying the corridor and not the assets. Continuous imagery of a line looks like coverage, but the defects live at the hardware, and hardware needs a closer, slower look.

The other one is thermal captured at the wrong time. A joint under light load at nine in the morning may look perfectly healthy and be the thing that fails at four in the afternoon. Capture at representative load or state clearly that you did not.

And findings without asset IDs. A hot joint on an unidentified pole is not actionable.

Start with these

One or two kits that fit this job. Family pages cover the rest of the line.

Compound UV34B (Four-In-One UV Payload)

The UV34B Corona Camera Payload is the premium four-in-one UV inspection system offering the widest field of view (17...

From $71,428.00

View kit

SHARE C10-32

Complex sites, captured in fewer passes.

From $21,999.00

View kit

Choosing a model

Do I need a dedicated corona camera, or will thermal find the same faults?

They find different things. Thermal shows resistance, which appears at joints and clamps carrying current. Corona shows electrical discharge from insulation breakdown, which often produces no measurable heat at all. A failing insulator can be invisible to thermal and obvious to a UV camera. Networks that inspect for both carry both, because one is not a substitute for the other.

Is the Matrice 4T enough, or do I need an M400 with a dedicated payload?

The Matrice 4T covers general patrol, asset condition and thermal on a single compact aircraft, which suits most distribution work. Step up to a Matrice 400 when you need an interchangeable payload the M4T does not carry, such as a UV corona camera, a current meter, or a LiDAR unit for vegetation corridors. Many networks run both.

Accuracy

How reliable are thermal readings from a drone?

Reliable when the conditions are recorded. Radiometric thermal gives a defensible temperature only with a stated emissivity, reflected temperature and load at the time of capture. Without those, you have a relative comparison rather than an absolute measurement, which is still useful for spotting anomalies but is not a number to put in a report unqualified.

Can LiDAR really replace manual vegetation assessment?

For clearance measurement, yes, and it is more defensible. LiDAR gives measured distance from conductor to canopy across the whole corridor, assessed against your clearance envelope, rather than a judgement made from the ground or from a photo. Species identification and growth-rate assessment still need an arborist. The measurement is automated; the interpretation is not.

Workflow & data

Can findings feed straight into our asset management system?

That is the goal and it is mostly a matter of matching your schema. Findings need an asset ID, a location, a severity and imagery. Thermalmapper produces inspection reports with AI-flagged faults that you spot-check and export, which removes most of the manual review. We work through the export format during handover.

Compliance

What about flying near energised assets and over private land?

We give CASA-aware deployment guidance as part of handover, and we can talk through the operational approvals most networks work under. What we do not do is provide the approvals themselves, which sit with you or your ReOC holder. Landholder access and network-specific safety cases are similarly yours to manage, and most utilities already have established processes for both.

Next step

Bring a real site

Start with a call. If it looks like a fit, we bring the kit to your site and capture the section that usually causes trouble. You keep the data.