Bell 407 N595RL accident near Galliano, Louisiana

Probable cause

The pilot’s failure to ensure the helicopter was clear of obstacles before takeoff from the helideck, which resulted in the helicopter’s right landing skid pivoting about a helideck perimeter light during takeoff and a dynamic rollover. Additionally, the pilot’s improper takeoff technique likely contributed to the development of dynamic rollover.

NTSB analysis

The on-demand passenger flight was departing from an offshore production platform when, upon liftoff, the helicopter entered an abrupt right roll and crashed into the helideck and then descended into the water. Recorded parametric data indicated that shortly after liftoff, about 2 ft above the helideck, the helicopter was in a 32° right roll with a right roll rate of about 68 degrees per second when the device stopped recording.

A review of the helicopter’s in-cockpit video camera revealed that the pilot did not land the helicopter in the center of the helideck during the landing that preceded the accident takeoff. Additionally, the pilot did not reposition the helicopter before the accident takeoff. Based on video evidence, the position of the helicopter on the helideck resulted in the aft portion of the right skid to be adjacent to a helideck perimeter light. Examination of the helideck revealed impact gouges in the helideck surface that matched the bolt head pattern of the helicopter’s right skid tube. These gouges likely were created when the helicopter was in a steep right bank angle. The location of the impact gouges in the helideck surface further supports that the aft portion of the right skid tube was in contact with the helideck perimeter light at takeoff. The perimeter light housing, whose attachment hardware to the helideck was not frangible, was found significantly deformed. Based on the physical and video evidence, the helideck perimeter light became the pivot point for a dynamic rollover to occur during takeoff.

The helideck perimeter lights were 2 inches higher than the construction standard of 6 inches. However, because the helicopter’s right skid was already in contact with the perimeter light before the takeoff, the out-of-compliance height of the perimeter light, by itself, did not contribute to dynamic rollover.

Examination of the helicopter wreckage found no evidence of preimpact failure of the airframe, the main and tail rotor systems, or the engine. The main rotor blades and hub exhibited signatures of powered impact damage consistent with engine power delivery to the rotor system when the blades impacted the helideck. Additionally, the recovered engine control unit data confirmed that the engine was functioning normally up until the main rotor blades impacted the helideck. Examination of the flight control system found no evidence of preimpact fractures, disconnections, or restrictions. The lateral hydraulic servo actuator, which controls the helicopter in the roll axis, exhibited normal functionality during bench testing.

A review of the pilot’s previous takeoffs revealed that he typically did not follow company policy to bring the helicopter into a 3-5 ft hover check before continuing with the takeoff. The pilot’s improper takeoff technique (without a brief 3-5 ft hover check) would have decreased his ability to identify and react to any anomalies during the takeoff, including the onset of a dynamic rollover.

All three requirements for a dynamic rollover (thrust exceeding helicopter weight, a pivot point other than the helicopter’s center of gravity, and a rolling moment) were present during the accident takeoff. Based on the video evidence, the pilot was likely unaware the helicopter’s right skid was in contact with the helideck perimeter light before takeoff.

Toxicological testing of pilot’s samples detected low levels of ethanol in blood and vitreous fluid, high ethanol levels in liver tissue, but no ethanol in urine. This ethanol pattern is not consistent with consumption and is likely from postmortem production, as the levels vary significantly amongst specimens and there was no ethanol detected in the urine. Therefore, the detected ethanol did not contribute to the accident.

Text from the NTSB. Open the NTSB docket for CEN23FA071 for the full record.