Startled Shutdown: Fatal USAF E-11A Global Express PSM+ICR Accident

Startled Shutdown: Fatal USAF E-11A Global Express PSM+ICR Accident

On 27 January 2020 a USAF Bombardier E-11A (Global Express) 11-9358 crashed in the Deh Yak district in Afghanistan, killing both pilots.  The accident featured what a 1998 AIA/AECMA study termed a Propulsion System Malfunction + Inappropriate Crew Response (PSM+ICR).

A USAF Accident Investigation Board (AIB) issued its investigation report in November 2020.  A USAF AIB does not conduct a safety investigation (a separate Safety Investigation Board [SIB] is convened but its report is not released), instead as it explains:

In accordance with AFI 51-307, Aerospace and Ground Accident Investigations, this accident investigation board conducted a legal investigation to inquire into all the facts and circumstances surrounding this Air Force aerospace accident, prepare a publicly-releasable report, and obtain and preserve all available evidence for use in litigation, claims, disciplinary action, and adverse administrative action.

The Accident

The aircraft was operated by the 430th Expeditionary Electronic Communications Squadron (430EECS) from Kandahar Airport as part of Operation Freedom’s Sentinel.  It was equipped as a Battlefield Airborne Communications Node (BACN) communications-relay platform.

The accident flight was both an operational BACN mission and Mission Qualification Training for the co-pilot.

The aircraft commander, a Lt Colonel, was a current and qualified instructor and evaluator pilot in the E-11A.  He had 4736.9 flying hours of experience, which 1053.3 hours on type. He had also flown the KC-10, RQ-4, MC-12 and T-1.  

The co-pilot had completed Global Express ground and flight training at CAE on 10 November 2019, and received his basic qualification in E-11A on 17 January 2020. He had completed his first two MQT flights in theatre.  The AIB was told “he demonstrated knowledge of the aircraft systems, high situational awareness, asked great questions and was a great student.”  He had a total of 1343.5 flying hours, 27.6 hours on type.  He had previously flown the T-6 as an instructor and 127.4 hours in the B-1.

The aircraft took off at 11:05 local time and entered an orbit just west of Kabul at 42,000 feet altitude at c 11:36.  The flight was conducted under Visual Flight Rules (VFR).

At 12:50, the crew was cleared by ATC to climb to 43,000 feet. The engine throttles were advanced and the crew initiated the climb with the autopilot.

At 12:50:52, a fan blade of the left hand Rolls-Royce BR710 engine was released.  Post-accident photographs suggest the blade failure was contained in the nacelle.  This failure was accompanied by a bang, recorded by the CVR, which then stopped due to the g switch being activated by the out of balance vibration.  This Fan Blade Off (FBO) event resulted in an immediate shutdown of that engine by the engine control system and display of a L FADEC FAIL caution initially.   The left engine N1 dropped to 7.6% within five seconds of the FBO before spiking to an unrealistic 255.9%.

Within seconds the aircraft autothrottles also disengaged automatically. The autopilot remained engaged however. Bank angles remained essentially constant, consistent with a circular orbit, and the aircraft descended from 42,300 to 41,000 feet.

Ten seconds after the catastrophic engine failure, the crew retarded both throttles to 14º (the throttle lever range is from 0-40º) for one second, then slightly advancing the left throttle separately to 26º for one second, then retarding it to align with the right throttle (both now at 16º) for one second, before finally advancing the left throttle to 31º while retarding the right to idle (0º). 

The initial combined throttle movement was likely an attempt to reduce the airframe vibration and then the separate movements were to identify the vibrating engine.

The AIB comment that:

For the left and right accelerations, analysis indicates that the nose of the aircraft would have yawed to the left initially, but then rapidly right when the left throttle was moved up from 14 º to 26º and then back down to 16º. This analysis suggests that the sensation the crew felt…may have factored in their decision thinking the right engine had failed. 

However, the AIB observe that:

Because of the minimal time spent in each throttle position, the above throttle movements could not be used reliably to assess engine response. 

It appears the crew erroneously concluded it was in fact the right engine that had suffered the failure as 9 seconds later the right engine run switch was set to OFF, manually shutting down the right engine. 

Seven to eight seconds after this the recorded airframe vibrations dropped appreciably for reasons the investigators could not determine, undoubtedly reinforcing the crew’s confidence in their decision.

Though the cockpit displays are discussed, the AIB don’t draw any conclusions on their comprehensiveness and comprehendability, nor do they discuss whether any smoke or fume clues may have been present.  

At 12:51:19, 27 seconds after the FBO, the AIB says left throttle was advanced to full power (40º) then brought to idle. At 12:51:23, both throttles were advanced from idle to full power.  However, by this point both engines had been shut down.  Within seconds the Flight Data Reorder (FDR) stopped recording due to a loss of engine generated electrical power.

The investigators believe that the crew ran the DUAL ENGINE OUT checklist, albeit “with some deviations that are consistent with potential perceptions and conclusions about the status of the engines”.  In particular they maintained a higher airspeed that the 200 kts within the checklist. The Ram Air Turbine (RAT) appears to have successfully deployed, providing hydraulic and electrical power.  The Auxiliary Power Unit (APU) was started at some point below 28,300 feet.  The AIB do not comment on why the CVR & FDR recordings did not resume.

The crew meanwhile had made a Mayday call, announcing to Kabul ATC that they had lost both engines and, remarkably, that they intended to proceed to Kandahar.  At 12:52L the aircraft had been approximately 17 NM from Kabul International Airport, 28 NM from FOB Shank, 38 NM from Bagram Airfield but 230 NM from Kandahar, the crew’s chosen destination.

Kandahar was well outside the E-11A glide radius suggesting they expected to airstart at least one engine. 

The crew should have waited, in accordance with the checklist, until they had descended to 30,000 feet to attempt an airstart.  The AIB note that:

Evidence suggests and flight simulation profiles demonstrate that the time the aircraft was at or below 30,000 feet in either the windmill or the ATS-assisted [sic – APU-assisted] airstart envelope was approximately 14 minutes. As long as the engine run switches were on (directed in the FCOM DUAL ENGINE OUT checklist), there is a high probability that a right engine airstart should have occurred

There is no data to confirm whether any airstart attempts were made.  It is possible, if the crew continued to misdiagnose the failed engine, that attempts were wasted on the failed left engine.

At 13:03:30, the crew announced to ATC that they were now going to land at FOB Sharana.  Sharana was reported to have a 1000 foot ceiling, confirmed by other pilot’s testimony.

The aircraft was not able to reach Sharana and at c 13:09, the impacted the ground c 21 NM short of Sharana, on a heading of roughly 140º, consistent with a direct flight path towards the FOB.  The terrain was largely flat and snow covered however the aircraft impacted berms and ditches c 1-2 m high, skidding to a halt in approximately 340 m. The wings were ripped from the aircraft.  Much of the fuselage was destroyed by a post-crash fire.

The crew likely had less than a minute to maneuver after descending through the clouds.  Investigators found the leading edge slats out and the flaps extended, suggesting that the aircraft had been configured for landing.  

At approximately 13:09 the aircraft’s Emergency Locator Transmitter (ELT) activated and USAF A-10s in the area were diverted to conduct a search.  The accident site was located but the weather prevented a recovery that day according to the AIB.  The crew’s remains were recovered the next day along with the CVR. A second team returned to recover the FDR and deliberate further destruction of the wreckage occurred.

The AIB Investigation & Analysis: Signs of Bias?

Due to the nature of a USAF AIB, and its focus on litigation and disciplinary / administrative action, some important matters receive little attention. 

The major omission in this case is the actual cause of the fan blade off event! 

Coincidental to considering what data might have been evident on the CVR if it had not failed at the same time, the AIB casually mention a BR710 failure on a civil Global Express XRS in 2006.  In that event the aircraft commander…

…reported the first moments following blade separation as disorienting initially, with airframe vibrations of such magnitude as to lead the crew to wonder if they had experienced a mid-air collision.. He described a loud bang, and sustained vibration through the rest of the flight, sufficient to break stemware in the galley. The pilot in command also stated that he could not determine which engine had failed based on aircraft vibration and sensation alone without looking at the instruments.

The FDR data available suggest the vibration was c25% higher in the E-11A case.

The civil crew the 2006 Global Express XRS event took one to two minutes to verify engine indications and determine the failed engine prior to initiating shutdown.  In the USAF case there were just 24 seconds between the left engine FBO event and the crew’s shutdown of the right engine.  

The investigators considered it likely that the crew’s actions were hastened by a startled sense of urgency due to aircraft vibrations and other auditory/sensory cues.  The AIB note that:

There is research in the aviation community that reviews the effects of startle. The FAA Advisory Circular [AC120-111 Upset Prevention and Recovery Training], [issued] 4 January 2017, defines startle reflex as the uncontrollable, autonomic muscle reflex, similar to fight or flight, that is elicited by exposure to a sudden, intense event that violates a pilot’s expectations. It also identifies the startle response is the reaction to an “unexpected event that violates a pilot’s expectations and can affect the mental processes used to respond to the event”.

The AIB contend that:

The startle reflex lasts up to 1.5 seconds, while the startle response lasts 30 seconds or more.

This is not mentioned in AC120-111 but is consistent with 2015 EASA Startle Effect Management research report produced by NLR.

Furthermore, the AIB note that dual engine out scenarios were not practiced as part of the E-11A training syllabus.  Investigators found these were considered by many pilots to be improbable events and so were not frequently discussed either.  This of course does ignore that a single engine failure, as actually occurred here, are routinely practiced for (and probably disproportionately too!).  One issue seemingly not examined by the AIB is what type of engine failures were simulated in training.  A simple power loss gives very different clues than a rarer but more dynamic FBO event, which is more likely to induce the startle they discuss.

The AIB used the DoD Human Factors Analysis and Classification System (HFACS) guide to evaluate the relevant human factors, based on the work of Wiegmann and Shappell.   

HFACS has four main categories or levels: Organizational Influences, Supervision, Preconditions, and Acts.  

Each level is further divided into multiple sub-categories.

HFACS should encourage identifying factors beyond the ‘person level’ to consider wider, systemic organisational issues.  However, as discussed above, as the AIB is however primarily focuses on identifying blame and liability, unlike a safety investigation.

The AIB identified three actions (marked AExxx) and one precondition (marked PCxxx): 

• AE107 Rushed or Delayed a Necessary Action: i.e. a rushed engine shutdown resulting in the serviceable engine being shutdown
• AE206 Wrong Choice of Action During an Operation: i.e. deciding to head for Kandahar
• PC206 Overconfidence: i.e. in the likelihood of restarting at least one engine
• AE102 Checklist Not Followed Correctly: i.e. apparent deviations from the Dual Engine checklist 

No supervisory or organisational factors are identified by the AIB.  This is despite the training for startle generally and the training for an FBO event both being obvious potential latent human factors. 

This conveniently mean no human factors involving a human who was outside the cockpit of the E-11A were identified by the AIB.  This neatly illustrates how tools like HFACS can be abused when applied with a greater interest in apportioning blame than establishing understanding.  In other words, no matter how useful HFACS can be, it doesn’t preclude the classifications being biased by the classifier’s own beliefs and desires. 

In conclusion….

The AIB President found by a preponderance of the evidence that the cause of the mishap was the Mishap Crew’s error in analyzing which engine had catastrophically failed (left engine). This error resulted in the Mishap Crew’s decision to shutdown the operable right engine creating a dual engine out emergency. 

The AIB President also found by a preponderance of the evidence that the Mishap Crew’s failure to airstart the right engine and their decision to recover the Mishap Aircraft to Kandahar substantially contributed to the mishap.

Our Final Words

The USAF AIB process suffers from the phenomenon of WYLFIWYF (What-You-Look-For-Is-What-You-Find), as discussed in our 2022 article: What Lies Beneath: The Scope of Safety Investigations

HFACS is no a panacea, especially is misapplied (which is easiest to do when used as an investigation tool rather than a classification taxonomy).  HFACS standardisation is therefore important, as was demonstrated when the European Helicopter Safety Analysis Team (EHSAT) analysed 473 helicopter accidents occurring between 2000 and 2010 using HFACS as one of two classification taxonomies.  While the accidents were initially classified by 10 national/regional teams, they were still subject to a standardisation review so as not to pollute the aggregated results. Aerossurance’s founder, Andy Evans, was industry co-chair of EHSAT.

This is not the only fixed wing special mission aircraft accident over Afghanistan that raises questions on training of USAF crews on specialist fleets.  We discussed another in this 2019 article:  MC-12W Loss of Control Orbiting Over Afghanistan: Lessons in Training and Urgent Operational Requirements

Safety Resources

You may also find these Aerossurance articles of interest:

• Professor James Reason’s 12 Principles of Error Management
• Back to the Future: Error Management
• Psychology of Blame
• Just Culture or Just Culpability?
• Wrong Engine Shutdown Crash: But You Won’t Guess Which!
• British Midland Boeing 737-400 G-OBME Fatal Accident, Kegworth 8 January 1989
• Metro III: Propulsion System Malfunction + Inappropriate Crew Response
• B1900C PSM+ICR Accident in Pakistan 2010
• Deadly Delay: Catastrophic USAF CV-22B Osprey Gear Box Failure
• USMC CH-53E Readiness Crisis and Mid Air Collision Catastrophe
• RCAF Production Pressures Compromised Culture
• Loss of RAF Nimrod MR2 XV230 and the Haddon-Cave Review
• C-130 Fireball Due to Modification Error
• AC-130J Prototype Written-Off After Flight Test LOC-I Overstress
• C-130J Control Restriction Accident, Jalalabad 
• Korean T-50 Accident at Singapore Airshow
• ‘Procedural Drift’: Lynx CFIT in Afghanistan
• Investigation into F-22A Take Off Accident Highlights a Cultural Issue
• USAF MQ-9A Reaper Lever Confusion: Human Factors
• USAF Engine Shop in “Disarray” with a “Method of the Madness”: F-16CM Engine Fire
• Inadequate Maintenance, An Engine Failure and Mishandling: Crash of a USAF WC-130H
• Inadequate Maintenance at a USAF Depot Featured in Fatal USMC KC-130T Accident
• USAF Parachutist Fatally Extracted Through Ventilation Door
• USAF T-38C Downed by Bird Strike
• USAF T-6A Texan II Lost in Inverted Stall
• When Red Bull Gives You More Than Wings…
• Contaminated Oxygen on ‘Air Force One’ 
• Friendly Fire: Civilian Shot in the Head During USAF F-16 Training
• ‘Crazy’ KC-10 Boom Loss: Informal Maintenance Shift Handovers and Skipped Tasks
• MC-12W Loss of Control Orbiting Over Afghanistan: Lessons in Training and Urgent Operational Requirements
• SAR AW101 Roll-Over: Entry Into Service Involved “Persistently Elevated and Confusing Operational Risk”
• A Second from Disaster: RNoAF C-130J Near CFIT

---

Aerossurance has extensive air safety, flight operations, airworthiness, human factors, aviation regulation and safety analysis experience, including with many fixed wing special mission aircraft, civil & military.  For practical aviation advice you can trust, contact us at: enquiries@aerossurance.com

Follow us on LinkedIn, Blue Sky @aerossurance.bsky.social and Twitter @Aerossurance for our latest updates.