Dramatic AW139 Accident at Houma: Skillful Recovery from a Latent Production Defect

Posted by on Sep 23, 2024 in Accidents & Incidents, Crises / Emergency Response / SAR, Design & Certification, Helicopters, Human Factors / Performance, Offshore, Oil & Gas / IOGP / Energy, Safety Management | 0 comments

Dramatic AW139 Accident at Houma, Louisiana: Skillful Recovery from a Latent Production Defect (ERA N811TA, 24 September 2022)

On 24 September 2022, Leonardo AW139 offshore helicopter N811TA, of ERA Helicopters, was involved in an accident near Houma, Louisiana. The two pilots and four passengers were uninjured.

AW139 N811TA After Coming to Rest (Credit: Houma Fire Department)

The US National Transportation Safety Board (NTSB) released their safety investigation report on 11 September 2024.

The Accident Flight

The helicopter was returning from the then Murphy Oil Thunder Hawk offshore installation in the Mississippi Canyon (MC) Block 736 of the Gulf of Mexico.  The aircraft commander, who was Pilot Flying,  had 2160 hours of experience, 480 on type.  The copilot, the Pilot Monitoring, had 3760 hours, 218 on type.  The helicopter had ben flown for 7,491 hours in the 11 years since delivery.

About 7 minutes from their destination of Houma-Terrebonne Airport (HUM), the occupants smelled “burning plastic” throughout the helicopter. Initially there was no smoke evident or any abnormal cockpit indications.  The helicopter was otherwise behaving normally and the flight crew decided to turn off the air conditioning system in case it was the source of the smell.

However, a few minutes later there was a loud “whoof sound” accompanied by smoke emanating from the rear of the overhead circuit breaker panel. Within seconds the cockpit was engulfed with a “thick orange/brown smoke” resulting in “zero visibility”.

Simultaneously there was a low main rotor speed (NR) audio warning, a rapid overspeed of both engines, an upward movement of the collective control and a left movement of the cyclic control.

The copilot was unable to clear the smoke by opening the small ventilation window in the left-hand cockpit window. He then jettisoned the left-hand cockpit window and the smoke cleared.

At this point the collective was down and the cyclic was to the forward right. The controls required significant force to stay in position.  The power index (PI) was about 145% on both engines.  The NR slowly rose to be above 83% and eventually recovered to 100%.

Meanwhile, the helicopter climbed 3,500 to 4,000 ft as the crew were unable to establish a descent using normal flying control inputs. They tried selecting one of the engines to idle, but the NR decreased from 100% to the high 70s. Consequently, they returned that engine to a flight condition.

The pilots needed to use their ”full body weight” to keep the collective down, but the helicopter neither descended or slowed in that condition. The pilots found they could descend if they forcibly pushed the cyclic further forward, but the helicopter descent was at 170 to 186 knots indicated airspeed (KIAS).

They declared an emergency and requested fire and emergency medical services at Houma to be notified, and briefed the passengers.

On arrival over Houma the crew made a high speed descent from 6,000 ft to 1,000 ft and flew an orbit of the airport to verify flight controllability. The tower controller confirmed that the landing gear was extended.

However, the crew were unable to control engine power in manual mode via the beep switches on the collective. They tried to reduce airspeed by reducing the No. 2 engine to idle using the engine mode switch on the lower console with the No. 1 engine remaining at maximum continuous power. The helicopter only decelerated to about 140 KIAS and the pilots decided that autorotation would be the only way to achieve a safe landing speed.

On the first landing attempt, the crew aligned the helicopter with the Houma runway 36 centerline at 140 KIAS with the No. 1 and No. 2 engines in flight and idle modes, respectively. The No.1 engine was them also set to idle, but the NR rapidly decreased from 100% to about 75% so the crew reselected the No. 1 to flight mode and a go-around was conducted.

On the second attempt, the crew began the descent from 400 ft agl with No. 2 engine at idle and No. 1 being alternated between flight and idle modes when the NR went to 70% and 85% respectively.  As a result the helicopter descended to about 50 ft agl and an airspeed where an autorotation was practical with both engines at idle.

The helicopter made a run on landing and slid off onto the grass to the right.

The main landing gear collapsed but the helicopter came to a stop upright.

AW139 N811TA After Coming to Rest (Credit: Houma Fire Department)

The aircraft commander applied the rotor brake and the co-pilot turned off the fuel and electrical switches. The passengers egressed the helicopter. The engines had to be shut down by moving the overhead engine control levers to the full off position.

A CRM Lesson: “Keep Talking”

Not surprisingly due to their airmanship and crew resource management the pilots received an award from HAI for saving the aircraft and all on board.  When asked “what do you want people to remember?” they said:

Keep flying. Keep talking and keep engaged. Push those scared thoughts out of your head and keep working the problem.

And:

I agree. If you’re by yourself, keep talking to yourself. Keep working the problem. I think that would have been a fatal flaw if we didn’t keep talking.

The Safety Investigation: The Failure Mode

The collective controls are interconnected by the C1 torque tube mounted laterally under the
left hand pilot’s seat.

AW139 Collective Flying Controls (Credit: Leonardo via NTSB)

Metal control rod, C2, then runs vertically up the left side of the fuselage, connecting through a linkage to the composite C3 torque tube that runs laterally to the mixing unit.

Investigators found thermal damage and a longitudinal split of the C3 torque tube.

ERA AW139 N811TA: Damage to C3 collective torque tube and the wiring support strip used for wire routing (Credit: NTSB)

Electrical wire P190A6-G (p/n 3G9B11A1911) connects the No. 1 generator to the No. 5
distribution bar on the No. 1 essential bus in the cockpit overhead panel.  It was found
chafed, consistent with contact with a line of rivets on the C3 collective torque tube.

ERA AW139 N811TA: Chafed wires and rivets on C3 collective torque tube (Credit: NTSB)

ERA AW139 N811TA: Chafed wires and thermal damage to C3 collective torque tube (Credit: NTSB)

Investigators found melted resin from the C3 tube and observed soot on the adjacent panels.

ERA AW139 N811TA: Melted resin found on the top of the overhead panel (Credit: NTSB)

The wire assemblies routed to the overhead circuit breaker panels are intended to pass under wiring support strips.  However, in this case the the support strip had its plastic electrical mounts incorrectly installed on the upper side, so the wiring assemblies passed closer to the C3 torque tube, allowing P190A6-G to chafe.

Correct Wire Routing with Cable Supports BELOW the Strap Assembly (Credit: Leonardo)

Incorrect Wire Routing with Cable Supports ABOVE the Left & Right Strap Assemblies as Found on AW139 N811TA (Credit: NTSB)

The Safety Investigation: The Production History

This AW139 was assembled at Leonardo’s Philadelphia, Pennsylvania facility, in 2011 using assembly task job cards first issued in 2007.  Card 3G0630A04112C4R, contained instructions on the installation and routing of electrical power cables from the nose to the cabin roof and the power distribution panels.

In the views looking upwards (from within the cabin), the power cables were depicted as solid lines with the strip assemblies as solid lines behind them.  This indicated that the routing of the electrical cables was under the strip assemblies, but the assembly job cards provided no specific prompt to route that way.  Erroneously however the two strip assemblies were shown with an identical geometry when in fact they were different.

In 2014, after N811TA was built but before the accident, the job cards were amended to correct the geometries and improved with further drawings & the statement:

Ensure that the moving parts of the flight control rods in the area has at least 0.5 inch clearance from the surrounding fixed parts.

We assume this is a ‘standard practice’ that from 2014 was added to the job cards as a specific reminder.

Although the NTSB don’t explicitly comment, once the wiring support strip assemblies are installed they provide a far more powerful indication of the wiring route than any drawing and so misrouting becomes inevitable unless caught by inspection.

Investigators found that the manufacturing drawings for the right-hand wiring support strip assembly showed two views.

These clearly show that the electrical supports were mounted on the side opposite that of the nut plates.

However, the drawings for the left-hand assembly contained a less compelling single view.

Investigators found that crucially on N811TA the electrical supports and nut plates were both
mounted upper side of both the left and right strip assemblies.  The wiring had thus been installed above and so more vulnerable to contact with the C3 torque tube.

Safety Actions

The operator’s parent company, Bristow Group, a major AW139 operator globally promptly initiated a fleet check that identified 9 additional helicopters with the issues affecting this wire bundle (8 with incorrectly routed wiring, and one with correct routing but with incorrect support strap hardware).

On 11 October 2022 Leonardo Helicopters issued Emergency Alert Service Bulletin (EASB) 139-731 to require inspections of the forward cabin roof ceiling harnesses and their installation to identify potential wire chafing. The ASB required a borescope inspection to identify if the cable assemblies were incorrectly routed above their respective wiring support strips. If misrouting was found, further inspection of the wiring for chafing and collective C3 torque tube for damage was called for.  Any incorrectly built wiring support strips were to be replaced.

Finally, the EASB also required a visual inspection of the wiring near the No. 1 diode module as there had been chafing found in that region on N811TA and on one other helicopter fleet checked by the operator.

After the release of EASB 139-731 a total of 23 issues were found (8 aircraft with incorrectly routed wire bundles; and 15 with evidence of chafing at the aft cooling diode).  To put this in context, by September 2019, Leonardo had delivered over 1000 AW139s so the drawings were routinely being correctly interpreted.

The day after the EASB, EASA issued an Emergency Airworthiness Directive (AD), 2022-0209-E, to make the EASB mandatory in all nations that automatically recognised the State of Design ADs.  It took the US FAA until 2 November 2022 to issue a similar EAD, 2022-22-03 (ironically giving US operators less time to comply with the AD then it took to issue the AD).

On 24 October 2022, the strip assembly’s manufacturing job cards were updated with an 2-D orthogonal view (which the NTSB incorrectly call a 3-D graphic).

Also added was the following:

Note: Pay attention to correct assembly: the anchor nuts must be mounted on the side opposite to the electrical supports.

The assembly job cards now had an extra note and photograph added.

NTSB Probable Cause

The inflight loss of collective control of the helicopter due to thermal damage of a collective control torque tube that abraded with misrouted electrical wiring, which resulted in an electrical short and inflight fire.

Contributing to the accident were the incorrectly manufactured wiring support strip assembly that misrouted electrical wires near the collective control torque tube, the ambiguity of the support strip assembly drawing that allowed for its incorrect manufacturing, the inadequate quality control processes to identify the incorrectly manufactured support strip assembly, and the helicopter manufacturer’s inadequate assembly instructions that allowed the misrouting of the electrical wiring, due to the incorrectly manufactured strip assembly, on the helicopter production line.

Our Observations: Applying HCD 

As Murphy’s Law tells us: What can go wrong will go wrong: i.e. if a component can be assembled two ways (one correct and one incorrect) then misassembly is inevitable.  Aerossurance has long advocated a greater focus on human centred design (for example in this RAeS presentation in 2017) and has been pleased to support a major RAeS project: Development of a Strategy to Enhance Human-Centred Design for Maintenance.

The principles of HCD for maintenance also apply to production.

In this case an HCD solution would be to ensure the wiring support strip assembly electrical supports and nut plates can only be fitted to opposite sides of the strip (e.g. be fitting baulking features) so the components can only be fitted on the appropriate side.

Rather than reliance on technical documentation, if the support strip can only be assembled so the wire is supported below, then misrouting the wiring above the strip and missing that during inspection becomes extremely unlikely.

Misrouting of wiring, pipes and hoses during production or post-maintenance remains an all too frequent event.  Less than a year ago Aerossurance became involved in the follow-up to an oil leak on a modern offshore helicopter type.

However, its worth remembering that N811TA’s accident is currently a once in 4.2 million flying hour event, the AW139 being a popular and widely used type.

Our Observations: “The Scary Version” – Supply Chain Experience Levels

There have been supply chain disruptions across the helicopter industry for a decade:

  • From autumn 2014 oil price crash caused new helicopter orders to drop off and production of helicopters and spares to reduce.
  • From spring 2020 the global pandemic further slowed demand (though not as much as for the airline industry), resulted in layoffs & furloughs in some manufacturers and accelerated early retirements.
  • From spring 2022 the Russian attack on Ukraine not only caused challenges sourcing titanium but also cause a sudden spike in the oil price, an increase in offshore helicopter operations and therefore demand on the helicopter manufacturers.  That demand exposed the known and entirely predictable reductions in production capacity from reduced production and COVID-19.

“The scary version” is that:

  1. somewhere, on potentially any aircraft type, there has been slightly ambiguous documentation and components that could be misassembled,
  2. those became more vulnerable to error because of lower production work force experience (both recent and in total – and some types have had far greater production downturns than the popular AW139), something that Boeing, for example, has been shown by NTSB to have struggled with, and
  3. any production error might well now be built into in-service aircraft dormant until they dramatically reveal themselves as N811TA’s wiring did at some point in the future.

N811TA’s latent defect took almost 7,500 flying hours and 11 years to emerge as not every production defect manifests itself immediately or is readily detectable during base maintenance.  Though of course the consequence of misoriented plastic clips are rarely as significant as in this case.

We have also previously written about how these industry shocks have affected flight operations and another vital supply chain, that of suitable qualified and experienced personnel, the very resource that saved N811TA.

Are those two ‘supply chain issues‘ that your organisation’s SMS has formally considered?

You might only have a limited influence over one or other, but if they affect your risk the you should assess them and consider if collaboration with others might help.

UPDATE 26 September 2024: Opinion: Stay Tuned For The Era Of Aerospace Reorganization by Teo Ozsan and Jay Carmel

Safety Resources

The European Safety Promotion Network Rotorcraft (ESPN-R) has a helicopter safety discussion group on LinkedIn.  You may also find these Aerossurance articles of interest:

Aerossurance has extensive air safety, flight operations, airworthiness, aviation design & certification, design & production standards, human factors, aviation regulation and safety analysis expertise.  For practical aviation advice you can trust, contact us at: enquiries@aerossurance.com

Follow us on LinkedIn and on Twitter @Aerossurance for our latest updates.