Agras T100 Night Rescue on Solar Arrays: 100 kg Payload, 100 L Tank, and the Battery Math That Saved the Crew

TL;DR

• The DB2000 battery pushed 100 kg of IR-confirmed water across 12 km of solar trackers in -10 °C temps—still landing with 22 % reserve.
• Coaxial twin-rotor lift cut hover draw by 18 %, letting us run centimeter-level precision passes at 8 m s⁻¹ without a second battery swap.
• Spherical radar + IPX6K rating shrugged off condensation drip and 40 kt rotor-wash spray drift while the ground crew stayed dry.

From Spray Drift to Search Pattern—A Dusting Pilot's Flashback

Back in '19 I was called to a flaming wheat stubble field trapped under 3 000 V transmission lines. My old single-rotor rig could barely hold 25 kg, so every battery cycle was a coin toss between spray drift and a brown-out crash. We spent half the night walking the rows, swapping packs, wiping ash off nozzles, and still lost 14 % of the stand. That memory lit up the second I saw the flare on Solar Array 4 last Tuesday—same wind, same darkness, same clock ticking. Only this time the Agras T100 was on the truck, tank full, DB2000 already warm from the shop. I knew the numbers would be different; I just didn't know they'd be this different.

Why the T100 Was Built for Night, Water, and Watts

Coaxial Twin Rotor = More Lift per Amp

Two rotors sharing the torque cancel each other's swirl. Translation: 18 % less current draw in hover compared with a single-rotor of equal disk area. On a lithium pack that size, every amp you don't burn is 5 s of loiter you can spend looking for hot spots with the thermal cam.

DB2000 Chemistry Cold-soak Test

We logged internal impedance at 1.2 mΩ per cell at 20 °C, rising only to 1.6 mΩ at -10 °C. Older LiPo packs would have doubled that, sagging voltage and tripping low-bat warnings halfway through the grid. The T100 kept 92 % of nominal voltage until the 88 % SOC mark—exactly when we finished the last IR sweep.

IPX6K + Spherical Radar = No Time Wasted Wiping Lenses

Condensation beads rolled straight off the dome; radar tracked panel edges to ±2 cm even when the IR feed fogged. No hover, no wipe, no lost watts.

Technical Snapshot: Night Solar Rescue Load-out

Parameter	Agras T100 Spec	Scenario Value Recorded
Gross take-off weight	149 kg	149 kg (100 L H₂O + 8 kg IR pod)
Battery nominal energy	7 296 Wh	6 934 Wh usable (cold derate)
Average flight current	—	132 A at 49 V
Flight time (spray + search)	12–18 min	17 min 24 s
Reserve after mission	—	22 % SOC
RTK Fix rate	—	99.7 % (base 3 km away)
Wind gusts	—	18 m s⁻¹
Spray drift footprint (water fog)	—	<1 m lateral at 8 m s⁻¹ forward speed

Expert Insight
"People think nozzle calibration is only for crop chemicals. Wrong. When you're dropping 100 L of demineralised water onto 1 500 V DC bus bars, droplet size controls arc flash. I run XR110015 at 3 bar, VMD 210 µm. Anything finer rides the rotor wash and you get a conductive mist cloud—anything coarser cracks the tempered glass. That sweet spot cost me three burnt inverters to learn—save yourself the tuition."

Battery Efficiency Playbook for Night Ops

1. Pre-heat the pack to 25 °C while still on the charger. Every 10 °C rise gives back roughly 3 % capacity in the first two minutes of flight—cheap insurance when minutes equal acres.
2. Map your pattern so the heaviest tank segment coincides with the downwind leg. Less cyclic input equals 6–8 A lower average draw.
3. Keep forward speed between 7–9 m s⁻¹. Below 6 m s⁻¹ the coaxial system starts to milk the collective; above 10 m s⁻¹ parasitic drag climbs faster than the benefit of shorter flight time.
4. Use RTK "Fix-and-Float" toggle: lock to FIX on final approach to the hotspot, drop to FLOAT during transit. The CPU throttles the Kalman filter update rate, worth 200 mW—tiny, but over 18 min that's 0.15 % battery you just gifted yourself.

What to Avoid—Lessons from the Field

• Don't top off to 100 % SOC right before launch. Li-ion at 100 % plus cold soak equals higher voltage sag under load. Land at 95 %, wait 2 min, then launch—voltage stays flatter.
• Never trust a single IR spot reading. Panels can fool you: a 120 °C bus bar looks identical to a 45 °C reflective strip in 8–14 µm band. Cross-check with a 10 m hover and a handheld laser thermometer before dumping the full 100 L.
• Watch electromagnetic soup. A 50 ha solar farm is a 200 000 m² Faraday cage. Keep your base station >100 m outside the fence or you'll see RTK Fix rate drop to 60 % and the bird starts chewing amps while it hunts for corrections.

Common Mistakes That Drain the DB2000

1. Launching with partially charged spare packs "just in case." Cold-soak equalises them downward; you end up swapping sooner and burning more total watt-hours in hover.
2. Forgetting to switch radar to "Pole" mode. Default "Tree" adds a +2 m clearance buffer—nice, but every metre of climb costs 4 A in lift power over the array.
3. Running constant-rate spray instead of variable-rate. The T100 can modulate 0–12 L min⁻¹. A flat 10 L min⁻¹ across the whole pass wastes 14 L and 600 Wh when only 30 % of the farm is actually hot.

Frequently Asked Questions

Q1: Can the DB2000 handle repeated night flights in sub-zero temps without swelling?
Yes. The pack's internal heater blanket keeps cells above 5 °C during charge. Over 300 documented cold-soak cycles show <3 % capacity fade—well inside DJI's 80 % warranty threshold.

Q2: Will the spherical radar false-trigger on panel support pylons?
No. Set "Thin Obstacle" > 5 cm and the algorithm filters out I-beam echoes while keeping ±2 cm track of panel edges. We logged zero false-brake events across 1 800 pylons.

Q3: Is the T50 a better fit for smaller solar farms?
For arrays under 20 ha, the 50 kg payload of the T50 is plenty and saves one battery. Anything larger and the T100's 100 L tank halves sortie count—usually cheaper in man-hours. Contact our team and we'll run a SWATH-width simulation for your exact layout.

Ready to put real numbers against night-time watts, amps, and acres? Contact our team for a site-specific battery-efficiency map and nozzle-calibration kit.