AAIB: offshore operators need more sim training
The final report on the fatal crash of a CHC Scotia-operated Eurocopter SA 365N Dauphin 2 in the Irish Sea on Dec. 27, 2006, underscores the importance of simulator training and the need for landing guidance systems on offshore platforms. The UK’s Air Accidents Investigation Branch (AAIB) noted that the crew would have benefited from extensive training for unusual scenarios in a simulator as well as improved landing guidance systems for oil and gas platforms.
The accident took place in darkness at about 6:33 p.m. local time near the North Morecambe platform. While attempting to land, the helicopter flew past the platform and struck the surface of the sea at a speed close to 125 knots. All seven aboard–the two crewmembers and five passengers–were killed.
There was no evidence of any technical malfunction that might have contributed to the accident. The investigation therefore sought to understand “why two experienced pilots were unable to stop a serviceable helicopter from flying
into the sea.”
The weather conditions were listed among the numerous factors. It was a particularly dark night, with an overcast completely obscuring any celestial illumination. Weather conditions were “poor” but still “above the required minimums and not unusual for such operations.” The report suggests it was not raining, and notes visibility was probably close to 3.7 nm.
The copilot, the pilot flying, was flying an unstabilized approach to the platform. He made it clear he had no proper perception of depth. The first part of the approach was marked by steady increases in the collective, tail rotor input, cyclic pitch and cyclic roll input. The radio-altimeter readout initially decreased then increased.
Between the start of the descent and the “fifty-five” call (a standard call referring to a target speed), the descent became progressively steeper. There was a steady reduction in collective demand and a steady, positive change in pitch attitude–rather than the adoption of fixed values. Investigators believe that the pilot intended to reduce speed throughout the descent until the helicopter reached 30 knots.
However, the Board noted that fixed attitude or power settings simplify glidepath control, particularly when approach-angle cues are limited.
There was a change in the pilot’s technique after the “fifty-five” call, as the collective demand increased steadily, the descent rate was arrested and the helicopter began to climb. Investigators concluded that the rapid succession–two in 35 seconds–of control strategies indicates the copilot’s difficulty in assessing approach angle.
The approach was flown essentially by reference to visual cues. In dark, overcast conditions, it is likely that some cues were degraded or absent. For example, without a distinct horizon, the pilot’s ability to assess pitch attitude and approach angle was probably compromised. Simultaneously, the lack of textural cues in the ground plane (in this case the sea surface) compromised the pilot’s judgment of pitch attitude and approach angle and estimation of range to the deck.
The AAIB asserts that, with improved helideck lighting, the crew might have detected deviations from a safe approach path at an earlier stage. Such modified lighting is recommended by the UK CAA and mandated by ICAO beginning this month.
On the day of the accident, the aspect ratio of the helideck would have been directly proportional to the approach angle. Therefore, in principle, the aspect ratio would provide a useable cue for ascertaining approach angle, investigators insisted. However, subsequent flight tests showed that a bright yellow perimeter light tended to obscure the circle of lights marking the helideck.
It is therefore possible that the crew found it difficult to judge the approach angle during the first phase of the approach. One way of meeting this challenge is to standardize control technique for an approach and minimize the number of variables vulnerable to change, the investigators said. This can be done, the report suggests, by starting the descent at a specified height and range and maintaining a stable pitch attitude. The drift in pitch during the first phase of the Dauphin’s approach (10 degrees) effectively prevented the crew from pursuing this technique. It would have tended to obscure any change in approach angle.
The second possibility is to supplement limited visual cues with instrument guidance. According to investigators, however, the radio altimeter was not conveniently situated for inclusion in the copilot’s scan during the majority of the approach.
The underlying causes, however, most likely stem from the limited visual cues available and the paucity of instrument cross-checks, according to investigators. In fact, the Board listed the commander’s inadequate monitoring of the approach as a contributing factor. Except for the “fifty-five” call and one height call at 400 feet, the commander did not provide any information that might have helped the copilot. He did not even draw attention to any of the extreme bank angles and pitch attitudes before the copilot requested assistance. Better monitoring would also have prepared the pilot-not-flying to intervene, investigators emphasized.
Before establishing visual with the North Morecambe platform, the commander was using the weather radar (in air-to-ground mapping mode) and the GPS to provide the copilot with range information. Once the crew had established visual contact with the platform, the commander provided no further range information. Available range information from the GPS, along with height calls from the commander, could have provided much of the required information for the copilot to fly a stable approach, according to the AAIB.
Noting the lack of instrument landing systems on offshore platforms, the report recommended that the EASA, which is responsible for such operational issues, complete research on GPS offshore approaches “without delay.” Such systems would “require substantial modifications to allow their use on oil and gas platforms” and would have to “cater to the tactical freedom necessary in helicopter operations.”
At some point, the copilot sought to hand over the controls to the commander but did not use standard phraseology. Instead, he said, “Help us out.” It took some time for the commander to understand the request, and the actual handover took four seconds. The commander’s first actions after taking the controls were correct, but he became concerned for the well-being of his copilot, who appeared to be upset at being unable to control the helicopter. This brief distraction from his instrument scan probably explains why he did not notice the increasing angle of bank to the right and the continuing descent into the sea. This also possibly explains why he did not hear the automatic voice alert device warning at 100 feet, the report continues.
Following the accident, CHC has identified the parameters to be monitored during an approach and has provided more specific guidance on the actions to be taken following disorientation or incapacitation. It has developed go-around procedures that include guidance for use of the autopilot coupler. A night traffic pattern has been developed and published, the investigators noted.
Investigators were also concerned that an appropriate synthetic training device (STD) for the SA 365N was available but not used. Therefore, the operator’s crews were denied “the extensive benefits that an STD can provide in both training and checking.” The report quotes European regulations that require abnormal/emergency situations be taught in a flight simulator.
Since the accident, the operator has continued to develop the implementation of a policy to train all pilots in STDs.
By the spring of last year, all SA 365 crews had received such training. All Sikorsky
S-92 and Eurocopter AS 332L crews have received simulator training, as have some of the AS 332L2 crews, with the remainder to be completed “as soon as possible.” All Sikorsky S-76 crews will have received training by the spring this year.