UA30 農業無人機：以流量、速度及作業幅寬核查施用率

A spraying drone can follow its route accurately while its treatment report remains wrong. The usual failure is a mismatch between three quantities: liquid leaving the aircraft, distance travelled over the ground, and the width that actually counts as treated. A dashboard that displays only tank capacity cannot resolve that mismatch.

The UA30 is a complete 30L agricultural drone listed by UNITED UAV. Here, it is a concrete example for developers building application reports and technical teams evaluating a complete aircraft package. The useful question is: what evidence would let another person reproduce the reported application rate?

Disclosure: UNITED UAV sells this aircraft. This article was drafted with AI assistance. Calculations and sample records below are synthetic engineering examples, not UA30 field measurements.

Treat the aircraft as a system

A complete spraying aircraft combines flight control, positioning, a liquid circuit, atomizers, power, mission planning and operator decisions. Their outputs interact. Changing groundspeed changes area covered per second; changing pump output changes volume delivered; changing altitude or operating conditions can change the usable deposition pattern.

For procurement, separate the supplied configuration from available options. Request an itemized aircraft, controller, battery, charger and spraying-system list. Establish whether positioning corrections, terrain sensing, spreading equipment and data-export access are included in that particular quotation. A photograph or a broad feature description does not settle those questions.

Also separate a liquid tank's volume from payload mass. Thirty liters of a formulation does not necessarily weigh thirty kilograms. Density, usable fill level, hardware and the approved mass envelope remain relevant. Software should keep volume and mass in different fields rather than silently converting between them.

This matters when comparing complete machines: a larger tank may reduce refills, but it cannot establish daily output without flight duration, charging, servicing and field geometry.

Start with the units

Let Q be total measured flow in liters per minute, v be groundspeed in meters per second, and w be the validated track spacing in meters. For a straight, steady pass, the nominal application rate is:

R [L/ha] = 10,000 × Q / (60 × v × w)

The conversion follows directly from area per second, v × w, and volume per second, Q / 60. Aerial calibration guidance also expresses this relationship using speed in kilometers per hour. The I2I aerial-equipment calibration SOP documents that form; its public-health operating recommendations should not be transferred wholesale to agricultural spraying.

For an illustrative 6 m/s pass with 6 m track spacing and 4.32 L/min total flow, the result is 20 L/ha. If speed falls to 3 m/s while flow remains unchanged, it becomes 40 L/ha. A route can therefore look geometrically correct while the nominal volume rate doubles.

These are deliberately chosen example inputs, not an asserted UA30 operating point or a pesticide recommendation. Select the actual target and operating envelope through the applicable product label, qualified agronomic advice and aircraft instructions.

Integrate intervals before calculating the mission average

Telemetry rarely arrives at a perfectly uniform rate. Integrate each interval separately: delivered volume equals flow multiplied by elapsed minutes; nominal swept area equals groundspeed multiplied by spacing and elapsed seconds. Then divide total delivered volume by total area.

Do not average individual rate values without weighting. Five seconds at 6 m/s and five seconds at 3 m/s cover different areas. With the example flow and spacing above, those intervals deliver 0.72 L across 270 square meters, giving approximately 26.67 L/ha. The simple arithmetic average of 20 and 40 is 30, which is wrong for that combined volume and area.

A small offline checker makes the arithmetic inspectable:

from math import isfinite

def nominal_rate(intervals):
    liters = area_m2 = 0.0
    for dt_s, speed_mps, spacing_m, flow_lpm in intervals:
        values = (dt_s, speed_mps, spacing_m, flow_lpm)
        if not all(isfinite(x) for x in values):
            raise ValueError("non-finite observation")
        if min(dt_s, speed_mps, spacing_m) <= 0 or flow_lpm < 0:
            raise ValueError("invalid moving-pass interval")
        liters += flow_lpm * dt_s / 60
        area_m2 += speed_mps * spacing_m * dt_s
    if area_m2 == 0:
        raise ValueError("no valid area")
    return liters * 10000 / area_m2

assert abs(nominal_rate([(5, 6, 6, 4.32)]) - 20) < 1e-9
assert abs(nominal_rate([
    (5, 6, 6, 4.32), (5, 3, 6, 4.32)
]) - 26.6666666667) < 1e-8

This function assumes interval-average measurements, straight moving passes and no overlap. It accepts neither raw asynchronous sensor messages nor stationary discharge. Production ingestion must classify those conditions before invoking it; deleting troublesome intervals would create a misleading report.

Make missing evidence visible

For every interval, retain aircraft identity, mission identity, configuration revision, timestamp basis, elapsed time, spray state, speed, total measured flow, spacing source and quality flags. Distinguish measured flow from a pump command. A requested output is not evidence that the same volume passed through the nozzles.

Align measurements to a common clock. If position arrives several hundred milliseconds after flow, a speed change can be paired with the wrong discharge interval. Document the alignment method, expected delay and maximum accepted gap. Preserve the original messages so a revised alignment can be evaluated later.

When data disappear, report unknown coverage and volume separately from accepted intervals. Do not fill a disconnected flow sensor with zero. Zero means an observed absence of flow; missing means the system cannot establish it.

Before promising integration, ask which exports the delivered UA30 configuration actually provides. Obtain sample files, units, timestamps and field definitions. This article proposes an analysis format; it does not claim that the aircraft exposes a particular public SDK, protocol or telemetry endpoint.

Separate flown area from uniquely treated area

Distance multiplied by spacing is a useful diagnostic denominator, but it counts a repeated pass again. Turns, overlap, revisits and boundary crossings can make nominal swept area larger than the unique area treated inside the field.

For a spatial report, represent each accepted pass as a footprint in an appropriate projected coordinate system, clip it to the approved field boundary, and calculate the union. Retain overlapping coverage separately rather than discarding it: repeated application can matter even when the unique-area total stays unchanged.

A geometric buffer is still only a model. The effective width needs deposition assessment for the actual setup and conditions. Positioning accuracy does not prove uniform deposition across that width. Report nominal swept-area rate, unique-area volume rate and any observed deposition results with distinct labels.

This distinction also prevents an attractive but unsupported conclusion: a low average rate across a large polygon does not prove that every part received an appropriate dose. Underapplication and repeated application can coexist.

Reconcile liquid and the complete work cycle

Compare integrated discharge with an independently measured liquid balance. Account for starting volume, refills, recovered liquid, residual contents and any documented draining or leakage. A disagreement can point to calibration error, an omitted refill, a clock problem or a different measurement boundary.

Investigate the residual rather than forcing both totals to match. Preserve uncertainty from collection vessels, scales, density measurements and flow calibration. Acceptance tolerances should follow measurement capability and operational requirements, not a convenient number chosen after the demonstration.

At an illustrative 20 L/ha, a nominal 30L tank gives 1.5 hectares by division alone. Actual dispatch planning must account for usable liquid, required reserve, non-treatment discharge and the energy needed to finish and land. That division says nothing about whether one loaded flight can cover the area.

For throughput, measure the whole cycle: preparation, takeoff, productive passes, turns, return, refill, battery handling, charging and cooling. Distinguish battery energy from ground charging capacity. Neither charger wattage nor a full tank establishes sustained hectares per hour.

Use discrepancies to choose the next measurement

When reported volume disagrees with the independent liquid balance, start with the liquid circuit and its measurement boundaries. Check priming, residual liquid, refill entries and whether the recorded quantity is measured discharge or a requested setting. Repeating the route faster will not isolate a flow-meter problem.

When volume agrees but the rate changes between otherwise similar passes, inspect speed alignment, track spacing and spray-state transitions. A lagging speed message can produce a large instantaneous spike while the total volume still looks reasonable. Keep those spikes attached to their timestamps; smoothing them into a reassuring average removes diagnostic evidence.

When both arithmetic checks agree but deposition is uneven, investigate the physical application pattern. Nozzle condition, liquid properties, aircraft motion, height and wind can matter. A consistent digital report does not establish biological effectiveness or drift control.

Sensitivity calculations help prioritize measurements. If true flow is 5% above the recorded value while actual effective width is 10% below the assumed width, the true nominal rate is about 16.7% higher than reported: 1.05 divided by 0.90. Neither discrepancy looks dramatic in isolation, yet their effects compound.

For a complete UA30 evaluation, this creates a practical division of work. The software team checks timing and arithmetic; the equipment team establishes configuration and calibration; the application specialist assesses deposition and permissible operating conditions. Define who supplies each record before the demonstration, so an unexplained result leads to a specific follow-up measurement rather than another sales comparison.

Ask for a reproducible acceptance record

A useful UA30 demonstration should leave more than a finished-area screenshot. Request the exact configuration, approved operating conditions, calibration record, raw export, refill log and a report showing rejected or unknown intervals. Run the same dataset through an independent calculation and compare the results.

Begin with controlled calibration and an appropriate nonhazardous test liquid under the aircraft supplier's procedure. Keep physical aircraft protection and operator control independent of this reporting script. The checker evaluates recorded evidence; it neither commands flight nor determines chemical application settings.

The UA30 product listing identifies the aircraft discussed here. For a technical evaluation, define the required data exports and acceptance evidence alongside the hardware configuration. A complete spraying system becomes easier to compare when its reported work can be reconstructed from measurements, boundaries and explicit assumptions.