6.4 km of fence in 19 minutes.
A VK-166 Recon flew the 6.4 km perimeter of a 40 ha solar site at night in 19 min, with no pilot, and reported four faults by thermal and zoom. Conditions, method, results, and what we changed.
On the night of 1 October 2026, a VK-166 Recon flew the full perimeter of a 40 ha solar site in the English Midlands: 6.4 km of fence, three gates and two inverter yards. The sortie took 19 min from launch to landing. No pilot flew it, and it found four faults.
A walking patrol of the same fence takes the site's night team about 1 h 50 min.
Conditions
- Wind: 8 m/s from the west, gusting 11 m/s.
- Temperature: 7 °C, falling to 6 °C by landing.
- Light: overcast, no moon; launch at 22.40.
- Ground: wet grass after rain at dusk.
Method
The operator set one area, the site boundary, and one task: fly the fence line and report anything that differs from the last daylight survey. The float planned a closed route 10 m inside the fence at 30 m, and a hold of 40 s over each gate and inverter yard.
It flew at 8 m/s along the straight runs and slowed to 4 m/s at the corners. The 640×512 thermal camera led, as it does at night. When the thermal view showed something that did not match the baseline, the 12× zoom camera turned to it, recorded 10 s of 4K video and the float moved on.
Results
- Distance flown: 6.9 km, including the climb and the holds.
- Time on task: 19 min.
- Battery at landing: 41%.
- Faults reported: 4.
The four faults:
- A run of 4 m where the mesh had pulled away from two posts.
- The north gate, closed but not latched.
- A cabinet in the east inverter yard, 18 °C warmer than the five beside it.
- Brambles over the fence along 60 m of the south side, thick enough to hide a break.
The site's engineers checked all four the next morning. The warm cabinet had a failed cooling fan.
What failed, and what we changed
On the north side, the west wind came across the fence line in gusts of up to 11 m/s, just under the VK-166's limit of 12 m/s. The float held its route, but the zoom camera lost its lock on the fence twice and recorded 6 s of open field instead.
We changed two defaults. When the crosswind is above 9 m/s, the route now sits 15 m inside the fence on the windward side, not 10 m, and the float slows to 6 m/s on that side. The zoom camera holds its lock better at the wider angle. Both changes add 1 min to this perimeter.
A fence is long and the night is dark. A float that knows what the fence looked like yesterday only has to look for what changed.
More.
Swarm mapping over open water.
Three CF-1211 floats mapped 126 ha of open water in 27 min as one coordinated swarm, with no pilot and no landmarks to stitch by. Conditions, method, results, and what we changed.
110 km out and back.
A ZZ-R12 Long flew 110 km out and back along a disused rail corridor in 1 h 58 min, without a pilot, and landed with 26% of its energy in reserve. How it plans a route that long, and where it decides to turn.
Survey at first light.
Two CF-1211 sorties mapped 88 ha of arable land before sunrise, in winds gusting to 13 m/s at 9 °C. What we measured, what failed, and what we changed.