DUASxx18 - Drone took off with incorrectly fitted rotor arms

Initial Report

In preparation for night-time multirotor drone (~25kg MTOW) maritime surveillance flights, the drone was set up as per pre-flight procedures by the Hub Operator (HO) (personnel in charge of setting up the drone, given the Remote Pilot (RP) is located in a remote operations centre away from the actual flight location). The RP noted the slot time to operate in as approaching, in addition to it being dark, with the HO using their mobile phone to provide lighting during the initial drone set up.

The drone took off, reaching a cruise altitude of 50m where it began to fly its pre-planned routing towards Waypoint 1. During the flight the RP noticed abnormal flight path deviations (no geofence breech), until the Control Unit (CU) alerted of a ‘Flight Zone Exceedance error.’ As an automated response, the drone auto-initiated land-mode as per its design, at which point the RP intervened by initiating hover. During this time the drone had climbed to 56m, requiring the RP to further intervene and descend to 50m whilst concurrently initiating a Return to Base (RTB), given the unstable nature of the flight. The drone landed safely; however, it was noticed the flight path was unusual/abnormal during the landing phase.

On landing, the RP communicated the abnormal behaviour to the OEM support engineer, including the Flight Operations Manager where a full investigation was conducted. The investigation concluded that Motor Arm 5 & 6 were fitted incorrectly with 5 being in slot 6 and vice-versa. There were perceived commercial pressures to make the allocated slot time, which lead to SOPs not being followed correctly and vital steps being missed.

Investigation outcomes:

• The remote pilot and hub operator were current and competent on the type of aircraft flown, both having operated recently and on a regular basis. There were no identified training issues on either record.
• The HO had limited visibility due to inadequate lighting and was trying to put the drone together whilst holding a mobile phone light in one hand.
• SOPs dictate that pre-verification photos should be sent from the HO to the RP to ensure the drone is set up correctly; the wrong arm position was missed by the RP.
• The drone flown had no defects and was subject to no firmware updates prior to flight and was considered airworthy.

To mitigate for further reoccurrence:

• HOs are now issued with head torches to aid drone setup at night-time.
• Checklists have been updated to include a double photo verification process by both the RP & the HO specifically for drone arm position, creating a more robust independent check.
• The organisation operating the drone conducted additional training to familiarise all RP & HOs with the new checklist procedures.
• The slot time for the operation was increased as the team debrief highlighted commercial time pressure for the operation.

CHIRP Comment

This is a detailed report of the occurrence as well as the follow up actions, that has been logged in the Operator’s Safety Management System. During set up, fixing the rotor arms in the wrong position under low-light conditions is clearly an HF occurrence. However, we would also suggest that, from a manufacturing perspective, it could be regarded as a design oversight that it is physically possible to do this. So, whilst it is an operator’s report, we suggest it might be worthwhile feeding back to the manufacturer that the next generation of the drone should have a physical modification that prevents fitting a rotor arm in the wrong place.

We like the list of actions taken after the occurrence, such as additional photographic verification of the arms being locked in the correct position and then being sent by the HO to the RP, as well as ensuring the pilots are equipped with head torches. There was clearly some time pressure coming from the allocated slot time, but a verbal cross check of actions performed by the HO whilst handing over to the RP, might well have done the trick.

One other aspect of the report caught our eye. It seems that the flight was over water and, on a Flight Zone Exceedance error being triggered, the drone initiated an auto-land. The RP was then required to switch to hover, before the pilot descended to the flight planned height limit. If the option exists on the UAS used, we suggest that the RP sets the default trigger to “Hover” or “RTB”, when flying over water, rather than the option “Auto Land”. We would also recommend that this setting should apply when flying over a built-up area.

Whilst the decision to equip the crew with head torches is good, we would probably go a little further and suggest that the take-off and landing site should really have some form of permanent static flood lighting.

Finally, it is worth noting that most Operators perform a control check just after the initial lift off at a height of around 10 metres. The purpose of doing this is to ensure that the movement of the sticks corresponds to the correct movement of the aircraft. We wonder if doing this would have resulted in the controllability issue being noticed slightly earlier in the flight.