CT155201 Hawk - Epilogue - Flight Safety Investigation Report

Report / June 10, 2011 / Project number: CT155201 - A Cat

Location: Cold Lake, Alberta
Date: 2011-06-10
Status: Investigation Complete

FSIR - download PDF version, 1.1MB

Epilogue

A crew of two qualified instructor pilots were conducting an instructor upgrade sortie, including a wingman syllabus mission, in a British Aerospace Systems Hawk aircraft when they heard a loud bang and noticed an increasing turbine gas temperature.  They then discontinued their training, analysed the aircraft systems, and turned the aircraft towards the Cold Lake airport.

The pilots set a medium engine power setting and commenced a shallow climb above 12,000 feet, above mean sea level.  After receiving their wingman’s report of smoke emanating from their aircraft and after noticing an increase in engine vibrations, the pilots shut down the engine.  Shortly thereafter, after determining that insufficient altitude remained to glide to the Cold Lake airport, they attempted to restart the engine.  During the restart, the wingman reported flames coming from the lead aircraft, after which the pilots then discontinued the restart and resumed their glide.

Unable to reach a runway, they carried out a controlled low level ejection.  The pilots parachuted in to a shallow swamp, receiving minor injuries, while the aircraft crashed and was destroyed.

The investigation concluded that the Hawk CT155 Adour Engine low pressure turbine (LPT) blade, which had a history of fatigue cracking at the trailing edge rear acute corner, failed prior to reaching its design life.

Four preventative measures were implemented to address LPT blade fatigue cracking, failure and liberation.  Additionally, the LPT blade design life was reduced to from 2,000 to 500 hours; it is expected that a new certification hours will return the design life to 2,000 hours by 1 March 2016.

Additional significant recommendations addressed pilot emergency handling procedures, forced landing glide profile determination, aircrew life support equipment, and amending Hawk pilot manuals and checklists.

CANADIAN FORCES FLIGHT SAFETY INVESTIGATION REPORT (FSIR)

FILE NUMBER: 1010-201 (DFS 2-2)

FSOMS IDENTIFICATION NUMBER: 147736

DATE OF REPORT: 12 June 2014

OCCURRENCE CATEGORY: A

AIRCRAFT TYPE: CT155 Hawk

AIRCRAFT REGISTRATION NUMBER: CT155201

DATE OF OCCURRENCE: 10 June 2011

TIME OF OCCURRENCE: 11:47 (local)

LOCATION: 2 NM South of 4 Wing Cold Lake, AB

OPERATOR: 419 Tactical Fighter Squadron

This report was produced under authority of the Minister of National Defence (MND) pursuant to section 4.2 of the Aeronautics Act, and in accordance with A-GA-135-001/AA-001, Flight Safety for the Canadian Forces.

With the exception of Part 1, the contents of this report shall only be used for the purpose of accident prevention.  This report was released to the public under the authority of the Director of Flight Safety, National Defence Headquarters, pursuant to powers delegated to him by the Minister of National Defence as the Airworthiness Investigative Authority for the Canadian Forces.

SYNOPSIS

A crew of two qualified instructor pilots were conducting an instructor upgrade sortie, including a wingman syllabus mission, in a British Aerospace Systems Hawk aircraft when they heard a loud bang and noticed an increasing turbine gas temperature.  They then discontinued their training, analysed the aircraft systems, and turned the aircraft towards the Cold Lake airport.

The pilots set a medium engine power setting and commenced a shallow climb above 12,000’, above mean sea level.  After receiving their wingman’s report of smoke emanating from their aircraft and after noticing an increase in engine vibrations, the pilots shut down the engine.  Shortly thereafter, after determining that insufficient altitude remained to glide to the Cold Lake airport, they attempted to restart the engine.  During the restart, the wingman reported flames coming from the lead aircraft, after which the pilots then discontinued the restart and resumed their glide.

Unable to reach a runway, they carried out a controlled low level ejection.  The pilots parachuted in to a shallow swamp, receiving minor injuries, while the aircraft crashed and was destroyed.

The investigation concluded that the Hawk CT155 Adour Engine low pressure turbine (LPT) blade, which had a history of fatigue cracking at the trailing edge rear acute corner, failed prior to reaching its design life.

Four preventative measures were implemented to address LPT blade fatigue cracking, failure and liberation.  Additionally, the LPT blade design life was reduced to from 2,000 to 500 hours; it is expected that a new certification hours will return the design life to 2,000 hours by 1 March 2016.

Additional significant recommendations addressed pilot emergency handling procedures, forced landing glide profile determination, aircrew life support equipment, and amending Hawk pilot manuals and checklists.

TABLE OF CONTENTS

1. FACTUAL INFORMATION
2. ANALYSIS
3. CONCLUSIONS
4. PREVENTIVE MEASURES
Annex A
Annex B
Annex C
Annex D
Annex E
Annex F

Top of page

1. FACTUAL INFORMATION

1.1 History of the Flight

1.1.1  The accident aircraft was the lead aircraft of a two plane formation operating south of Cold Lake Airport (CYOD).  Together, the lead aircraft, call-sign Zulu 21, and the wingman, call-sign Zulu 22, were designated as Zulu 21 Flight from 419 Tactical Fighter Training Squadron.  The mission was a one-versus-one basic fighter manoeuvre (BFM) training sortie.

1.1.2  The front seat pilot of Zulu 21 ( P1) a B category flying instructor undergoing an A2 category upgrade flight, was acting as the instructor in charge[1].  The rear seat pilot of Zulu 21 (P2) was an A2 category flying instructor who was monitoring and evaluating P1.

1.1.3  P1 met all prerequisites for the A2 Category upgrade check ride.  His A2 Check Flight Report and Category and Ability Report (FORM T-21) were duly completed and signed by the appropriate individuals [2].

1.1.4  P2 was the designated Zulu 21 aircraft captain and the Zulu 21 Flight formation commander.  The pilot of the second aircraft was a trainee on his first solo BFM mission of his Phase IV Fighter Lead-in Course.

1.1.5  The CYOD weather for the sortie was visual meteorological conditions (VMC) at all altitudes with surface winds of 200 degrees magnetic (°M) at 15 knots.

1.1.6  Zulu 21 Flight departed CYOD runway 31 right with a left turn to the Frog-Bronson training area [3].  Upon entering the Frog-Bronson training area and completing in-flight checks and warm up exercises, Zulu 21 Flight set-up for the first BFM exercise where Zulu 21 was to act defensively and Zulu 22 offensively.  Approximately 20 seconds after initiating a break turn, P1 advanced the throttle to maximum.  Seven seconds later at 10,960 feet above mean sea level (’ MSL) P1/P2 heard a loud bang.  At this time Zulu 22 saw a flame inside Zulu 21’s jet pipe along with a small puff of smoke.  While P1 continued to fly the aircraft, P2 noted that the turbine gas temperature (TGT) was increasing through 620 degrees Celsius (°C).

Figure 1:  Zulu 21 Flight Path and Times.  Data from the aircraft’s onboard air combat manoeuvring instrumentation (ACMI) pod was the primary source of flight path information.  Where this data stopped recording, Cold Lake radar data was used.  Google Earth was used to plot the data.  True north is up.

1.1.7  P1 retarded the throttle to 82% RPM and P2 noted that the TGT decreased back to normal parameters.  Neither P1 nor P2 reported seeing warning or caution lights on the caution and warning panel (CWP).  P1/P2 indicated they noticed mild airframe vibrations following the bang.  P1 made a radio call to discontinue the exercise and began a shallow climbing left turn towards CYOD.  At this time, Zulu 22 radioed that he had seen a flame coming from Zulu 21’s tailpipe.  P1 directed Zulu 22 to take spacing and follow them home because they had an engine malfunction.  An emergency was declared to the CYOD tower controller.

1.1.8  Zulu 21 was 44 NM south of the Cold Lake airport and 30 NM north of the Lloydminster airport (CYLL) when the engine malfunction occurred.  The engine continued operating and, although a mild vibration was felt through the airframe, P1/P2 decided to proceed back to CYOD rather than divert to CYLL.  P1 adjusted the throttle to approximately 86% RPM, which provided sufficient thrust to maintain a 1.7 degree climb angle (approximately 800’ per minute rate of climb) at 240 knots indicated airspeed (KIAS).  P1/P2 felt they would intercept the engine-out glide profile at some point as they approached CYOD.  They used neither distance measuring equipment (DME) nor air traffic control (ATC) assistance to determine their range, which was needed to calculate their glide profile to CYOD, but rather estimated their proximity to the forced landing glide profile by visual means.

1.1.9  Enroute to CYOD, P1/P2 reported that the magnitude of the vibration increased and became more severe.  P2 noted an oily smell in the cockpit.  24 seconds later, Zulu 22 radioed that smoke emanating from P1/P2’s jet exhaust pipe was intensifying [4].  As a result of the increasing vibrations and smoke reported by the wingman, P1/P2 believed the engine was going to fail catastrophically and decided that their only recourse was to secure the engine.  At 30 NM from CYOD, P1 secured the engine by completing the engine mechanical failure checklist procedure, Figure 2, and initiated a shallow zoom climb to trade excess airspeed for altitude until reaching approximately 190 KIAS.

Figure 2:  Mechanical Failure Procedure.  Taken from C-12-HWK-000/MC-000 Hawk Pilot Checklist, Revision Basic Change 6.

1.1.10  The aircraft altitude peaked at 13,490’ MSL at 26.8 NM from CYOD.  At 11:39:27 local time (LT), Zulu 22 radioed that the smoke was abating and at 11:39:45 (LT) Zulu 22 radioed that the smoke was gone.

1.1.11  Once the optimum engine-out glide parameters were established [5], P1 noted the visual aim point was short of the touchdown area for any runway at CYOD, indicating the aircraft was too low to safely glide to a runway.  P1/P2 discussed restarting the engine in order to reach a glide profile that would allow the aircraft to glide to a runway.  Following their discussion, restarting the engine became their desired course of action.  Two relight attempts were made; both were unsuccessful.

1.1.12  P1/P2 stated they prepared for the ejection without reference to the ejection checklist procedure, but instead carried out a head to toe check.  P1/P2 ejected below the minimum recommended ejection altitude of 2000’ AGL at 11:46:27 (LT).

1.1.13  At 11:46:35 (LT), Zulu 22 reported to CYOD ATC that P1/P2 had successfully ejected, that the aircraft had crashed near a residential building, and that there was a need for fire trucks on-scene.  At 11:47:16 (LT), Zulu 22 reported that both pilots were on the ground.

1.1.14  P1/P2 reported the ejection was a violent experience.  P1 reported a large triangular void in his harness above his shoulders following parachute opening and a sensation of slipping through his harness.  This required him to grasp both risers with crossed-arms in order to steady himself in the harness.

1.1.15  During descent, P1 noted he was drifting toward power lines and attempted to steer his parachute by dumping air.  He made one attempt to deploy his personal survival pack (PSP) but was unsuccessful, preferring instead to focus on avoiding the power lines.

1.1.16  P2 held onto the ejection handle until it was snatched away at parachute opening.  He required two attempts to locate and successfully release his PSP contents.  During the descent he manually deployed his life preserver in anticipation of a water landing.

1.1.17  Following the ejection, the aircraft continued to fly in a shallow left-turning descent until it impacted the ground at approximately 11:47 (LT), narrowly missing a house, Figure 3.

Figure 3:  Post-Ejection Aircraft Trajectory.  Trajectory based on Cold Lake radar data.  Centre of the ejection zone is indicated by a red triangle.  The ejection zone was calculated using a composite of ACMI and Cold Lake radar data in addition to known aircraft glide characteristics.  Google Earth was used to plot the data.  True north is up.

1.1.18  Zulu 22 proceeded to orbit the accident site and took on the rescue combat air patrol role, relaying time critical information to ATC until being recalled by 419 Sqn Operations.

1.1.19  The pilots landed very firmly in close proximity to one another in a shallow swamp approximately two NM south of CYOD.  They were then transported to the Cold Lake hospital and treated for minor injuries.

1.2 Injury to Personnel

 

InjuriesCrewPassengersOthersTotal
Fatal 0 0 0 0
Serious 0 0 0 0
Minor 2 0 0 2
Total 2 0 0 2

Table 1: Injuries to Personnel

1.3 Damage to Aircraft

Field Examination

1.3.1  The aircraft was destroyed as a result of the ground impact, explosion and post-crash fire.  There were five large pieces in the wreckage trail:  the main wings with the landing gear retracted, the fuselage, the engine including the exhaust jet pipe, the vertical fin and the horizontal stabilators.  Figure 4 shows the main crash site.

Figure 4:  Main Crash Site.  Aerial view of the crash site from the initial impact point to the main wreckage area in the direction of the wreckage trail.

1.3.2  As shown in Figure 5, the engine was found intact with two thirds of it enclosed by a large section of the centre fuselage, which was attached by its three mounting points.  The left side section of the fuselage over the engine was consumed by post-crash fire along with most of the electrical and non-metallic plumbing located on the exterior of the engine.

Figure 5:  Main wreckage area.

1.3.3  A field examination of the engine revealed the low pressure compressor (LPC) intake was exposed.  It was in good condition with no apparent distortion or visible damage to the LPC blades.  The jet exhaust pipe was found separated from the engine.  The last foot of the jet pipe was crushed to less than half its original diameter.

1.3.4  The jet exhaust pipe was moved aside to gain access to the engine’s exhaust mixer and low pressure turbine (LPT) section.  One full LPT blade was missing at the 5:30 position from the LPT rotor disc.  The next eight LPT blades in rotation had sustained minor damage.  There were three LPT blades located approximately 180 degrees from the missing blade with their tip shrouds missing, Figures 6 and 7.  While examining the LPT section with a borescope, the lower portion of the missing LPT blade, called the fir tree section, was located forward of the LPT rotor disk and could not be extracted onsite.

Figure 6:  Missing LPT Blade.  The LPT disk is viewed from the back of the engine.

Figure 7:  Post-Disassembly LPT Disk.  The LPT disk is viewed from the back side.  Rotation is counter clockwise in this view.  Note the three blade tips missing approximately 180° away from the lost LPT blade.  The mass imbalance is equivalent to less than a single 64 gram blade because of the counterbalancing effect of the lost blade tips.  There is minor damage to a number of the other blades.  The centre orifice is the anti-ice/cooling tube outlet.  Note the soot inside the tube.

Engine Disassembly

1.3.5  Examination of the front of the engine did not reveal any obvious damage to the LPC blades, apart from a minor impact mark on the leading edge near the root of one blade.  Visual examination of the LPT confirmed that a single blade aerofoil was missing.  The root of the fractured LPT blade was located and recovered from the annulus between the low pressure (LP) nozzle guide vanes and the LPT disc.

1.3.6  During the disassembly several impact and puncture marks were examined; however, it appeared that the blade aerofoil release was contained within the engine casing.

1.3.7  Visual examination of the LPT blades revealed a large portion of the shroud from the first blade trailing the released blade was missing, and there was significant mechanical damage to the trailing edges of the forward eight blades.  The shrouds from three blades that were located diametrically opposite the released blade were missing.  Several other blades exhibited impact damage to their trailing edges, although the aerofoils of the remainder of the blades were relatively undamaged.  The damage to the eight forward blades was considered to have been caused by released material from either the fractured blade or from one of the released shrouds.

1.3.8  The disassembly continued until the engine was stripped to a modular level.  No further significant pre-crash damage was identified.  On previous engines that had suffered a similar LPT blade fracture, both the high pressure turbine (HPT) and LPT locating bearing securing bolts were found loose as a result of the out-of-balance LPT, however, the HPT and LPT bearing securing bolts on this engine, engine 7825, were found to be correctly secured.

LPT Blade Examination

1.3.9  The assessment of the LPT, including the fractured blade root remnant, was initially conducted at Quality Engineering and Test Establishment (QETE), followed by further examination in the Materials Laboratory at Rolls-Royce (RR), the engine manufacturer, in Bristol, UK.  The released aerofoil from the fractured blade was never recovered.

1.3.10  In accordance with the build record positions the fractured blade was numbered 62, with the remainder of the blade set numbered accordingly.

1.3.11  Assessment of the aerofoils on the eight blades leading into the primary fracture of blade #62 and of the blades diametrically opposite to blade #62 that had lost their shrouds revealed that the damage was secondary to the primary blade loss event.

Non-Destructive Testing

1.3.12  Ultra high sensitivity fluorescent penetrant inspection was performed by QETE on the roots of the remaining 93 blades.  This procedure identified 13 blades with crack indications in the rear acute corner of blade #10, 11, 14, 15, 19, 22, 38, 56, 59, 64, 70, 81, and 94.

Remnant of Blade #62

1.3.13  The appearance of the fractured remnant of blade #62 was commensurate with the four previous LPT blade fractures in NATO Flying Training in Canada (NFTC) engines, which had also fractured from the rear acute corner.  The part number of the previous fractured blades was AX66831, whereas the part number of the fractured blade from engine 7825 was AX72453.  Further analysis of remnant blade #62 is provided in Annex D.

Root Cause Analysis Blade #62

1.3.14  As a result of the root cause analysis (RCA) a number of causes were identified that would affect the stress in the area under the platform at the rear pressure side shank fillet of the blade.  A summary of these findings follows:

a. The Nimonic 90 retaining plate chocking can increase the nominal root shank acute corner stress by as much as 11%;

b. Blade modelling using observed evidence suggested that the retaining plates can load off-centre on the blade resulting in change in the nominal stress from between -11% to +8%;

c. Retaining plate load being carried by only one blade can increase the nominal stress as much as 21%;

d. Malformation of the blade retaining plate groove increased the nominal root shank acute corner stress by 12% irrespective of the corner radius;

e. Parametric modelling using optical scanning data identified variations in the peak stress in the root shank acute corner due to differences in corner radii, shank offset and a malformed blade retaining plate groove; the variations in the nominal stress was +35%, +7% and +13% respectively;

f. The effect of the blade crystal orientation was assessed and it was found that the range of nominal stress variation using the extremes of crystal orientation was +10 to -33%;

g. An increase in steady state stress due to dynamic stress was derived from strain gauge testing, this being an average 4% increase on peak stress at the location of crack initiation during cold weather running; and

h. Whole blade fatigue testing showed that the presence of a surface layer on the shank corner of the blade reduced the fatigue life of the LPT blade by approximately a factor of two when the blade was excited in the first edgewise mode.  The engine does not operate within the speed range at the first edgewise mode and so it contributes in part to the cracking seen, but was not considered to be the sole root cause.

1.4  Collateral Damage

1.4.1  The main crash site was spread over a distance of 230 metres (m), as depicted in Figure 4, and affected three privately owned properties, all straddling a gravel-surfaced north-south, east-west road intersection.  The aircraft first impacted the field to the southeast of the intersection.  Damage to this field consisted of ground scars and overturned soil and 50 m of destroyed barbed wire fencing.  A 20 m portion of the fence in the field to the southwest of the intersection was also damaged.  An electrical line crossing over the road intersection was grazed by debris resulting in the rupture of its strands.

1.4.2  Most of the collateral damage was concentrated on the property to the northwest of the road intersection; this was where the main wreckage came to rest and caught fire.  In this field, 50 m of electrified fencing and 20 m of a tree line were levelled by the aircraft as it came across the land.  Fire damage was evident, covering roughly 5,000 m2.  Finally, soil contamination from fuel and other small amounts of un-burned aircraft fluids were apparent around the main fuselage section and the wing.  Soil contamination covered an area of roughly 200 m2 to an unknown depth.

1.5 Personnel Information

 

 Front Seat Pilot (P1)Rear Seat Pilot (P2)
Total military flying time (hours ) 3400.5 2078.2
Flying hours on type 112.0 1086.2
Flying hours last 30 days 12.1 29.2
Flying hours last 90 days 27.4 84.2
Duty hours last 48 hours 20.0 14.0
Duty hours day of accident 3.0 3.0
Annual proficiency check Valid, 3 May 2011 Valid, 9 Feb 2011
Currency Valid Valid

Medical category

Valid Valid

Instructor category check

Valid, B Cat*
11 May 10
*Extension granted until 10 Jun 2011

Valid, A2 Cat
3 Aug 2010

Supervisory check 24 Feb 2011 7 Feb 2011

Practice forced landing assessed

3 May 2011, Level 5

9 Feb 2011, Level 5

Emergency simulator 16 May 2011 10 May 2011
Ejection seat / egress check 4 Jan 2011 9 Feb 2011

Table 2:  Personnel Information

Front Seat Pilot

1.5.1  P1 was a current and qualified Hawk B category qualified flying instructor (QFI) and was the Commanding Officer of 419 Sqn with over 3,400 hours of flying experience.  His post-wings flying experience was on the twin-engine CF188 Hornet and the single engine CT133 Silver Star.  He completed conversion training to the Hawk in May 2009.  He obtained his initial QFI category in Dec 2009 and then commenced instructional duties.  His last B category QFI supervisory check was satisfactorily flown 24 Feb 2011 and he was in the process of upgrading to an A2 category QFI.

1.5.2  His last annual proficiency check was satisfactorily flown 3 May 2011.  During that mission a bird strike required him to conduct a practice forced landing (PFL), which he completed flawlessly to a level 5 [6].  He completed his quarterly emergency simulator mission 16 May 2011 and his ejection seat/egress check on 4 Jan 2011.

1.5.3  P1 had eight hours of sleep the night before.  He was not performing military duties at night (i.e. night flying or shift work) and he had not travelled across time zones prior to the accident.  P1 was awake for less than six hours prior to the accident, which did not occur during a circadian rhythm low.

Rear Seat Pilot

1.5.4  P2 was a current and qualified A2 category QFI with 2,078 hours of flying experience.  Prior to arriving at 419 Sqn as an Instructor, his post-wings flying experience was the Hawk phase IV and then on to the single engine F-16 Fighting Falcon.  He began conversion training to the Hawk in Jan 2007 and, after obtaining his QFI category on 11 Apr 2007, he remained at 419 Sqn as an instructor.  His last A2 category QFI supervisory check was satisfactorily flown 07 Feb 2011.  He completed his annual proficiency check on 9 Feb 2011 when PFLs were practiced and rated level 5.  He completed his emergency simulator mission on 10 May 2011, and his ejection seat/egress check on 9 Feb 2011.

1.5.5  P2 had eight hours of sleep the night before.  He was not performing military duties at night (i.e. night flying or shift work) and he had not travelled across time zones prior to the accident.  P2 was awake for less than six hours prior to the accident, which did not occur during a circadian rhythm low.

Wingman – Pilot

1.5.6  The Wingman was a recent wings-graduate in the early stages of his Phase IV NFTC fighter lead-in course on the Hawk.  He had recently arrived at 419 Sqn and had accumulated 14 course hours; he had 101 hours total flying time on the Hawk.

1.6  Aircraft Information

1.6.1  The accident aircraft was a single-engine British Aerospace Systems Hawk, advanced jet fighter trainer with a dual tandem pilot seat configuration.  The aircraft are given the Canadian Forces (CF) designation CT 155 Hawk.  It is equipped with Martin-Baker Mk BA10LH ejection seats, which are installed in both the front and rear cockpits.  The rear cockpit has, for the most part, the same instrumentation and aircraft control configuration as the front seat does.

Adour Mark 871 Engine

1.6.2  The Adour is a twin spool, turbofan, medium by-pass engine.  It has a total of seven stages of compressor:  two LPCs and five HPCs.  The compressors are driven by separate single stage turbines via co-axial shafts.  Each shaft rotates in an anti-clockwise direction when viewed from the rear.

LPT Blade Configuration

1.6.3  The initial standard of LPT blade fitted to the Adour engine was to part number AX66831.  Due to cracking in the aluminised coating on the aerofoil the method of coating was changed from pack aluminising to vapour aluminising, which resulted in the change of part number to AX72453 (Modification AO1112).  The conclusion of the assessment regarding the previous LPT blade fractures was that the cracking originated at the rear acute corner due to a non-conforming sharp radius that was coincident with an area of high stress.  To resolve this issue, an improved method of hand dressing the rear acute corner to achieve a satisfactory radius was implemented.  To identify that a blade had been re-worked or manufactured with a correctly formed rear acute corner, a “#” symbol was marked on the blade after the part number.

1.6.4  There were four previous similar LPT blade fractures in NFTC engines:  engines 7828, 7802, 7807 (215) and 7818 (205).  The fractures in these blades originated in the rear root neck corner, under the blade platform.  The RCA into those fractures established that the stress at this location was higher than is generally considered acceptable under current design understanding.  As result of these fractures, a campaign was initiated to replace all the LPT blades with blades having correctly formed root neck radii.

1.6.5  Coincident with the replacement with conforming blades, modification AO1692 was introduced.  The prime reason for this modification was to change the material of the retaining plate from Nimonic 75 to Nimonic 90 for additional strength.  The modification was introduced to alleviate the problem of deformation to the retaining plates, which has led to blades moving forward in the disc slots and to the release of retaining plates.  At the last shop visit, engine 7825 (201) was fitted with LPT blades to part number AX72453 with the # identifier to confirm a correctly formed root radius and it had modification AO1692 embodied.  Figure 8 shows typical LPT blade and the failed LPT blade.

Figure 8:  Failed LPT Blade.  This figure presents an undamaged LPT blade (on the left) for comparison with the recovered fragment of the failed LPT blade.  The accident LPT blade airfoil section was not found and likely exited the jet pipe shortly after the blade failed.

Engine Controls

1.6.6  Low Pressure Fuel Cock.  The low pressure (LP) fuel cock is a two position (OFF/ON) mechanical control mounted to the left console of the front cockpit only, and connects the aircraft fuel system to the engine and the gas turbine starter (GTS).  In the OFF position, fuel to the engine and GTS is cut off.

1.6.7  Throttle.  The throttle is mounted on the left console of both cockpits, and controls the high pressure (HP) fuel cock.  In the fully aft position (cut-off), the HP fuel cock is closed.  An idle lever stop prevents the throttles from being moved to the fully aft position.  An idle stop lever, when raised, permits the throttle to move from idle to cut-off.  Fore/aft movements of the throttle are converted to rotary movements of the fuel control unit (FCU) throttle valve, which increase/decrease fuel flow to the engine spray nozzles and engine RPM, respectively.

1.6.8  Ignition Switch.  A two-position engine ignition switch (NORMAL/ISOLATE) is mounted on the left console of the front cockpit only, and controls the power supply to two (upper and lower) high energy ignition units.

1.6.9  Start Master Switch.  The start master switch is a three-position switch (OFF/ON/START) where the START position is spring-loaded to ON.  The switch controls the electrical power supply to the GTS.  When selected to START, the GTS start sequence is initiated.

1.6.10  Relight Button.  A relight button is located on the throttle in both cockpits and performs a number of functions.  With the start master switch ON, the ignition switch at NORMAL and the throttle at ½ inch above idle or less, pressing and releasing the relight button starts the GTS.  When airborne, pressing and releasing the relight button also energizes the high energy (HE) igniters for 30 seconds.

Adour Engine History

1.6.11  Prior occurrences identified an issue with fatigue cracking of the LPT blades between the fir tree root and the blade platform.  This failure mechanism was called a root rear neck acute corner machine tear or more simply, a root failure.  This type of failure was thought to be the result of fatigue cracking caused by an inadequate shank (root) to platform radius, which could eventually cause the blade to fail between the fir tree root and the blade platform.  This portion of the blade is hand-dressed to achieve the correct radius.  This type of failure first began to appear in 2006 and, for reasons not fully understood, has only been seen in Adour engines in use by the United States Navy (USN) and NFTC.  In 2011 two USN blades, reworked to the new corner standard, failed with the same mechanism and in 2013 a Royal Australian Air Force blade failure occurred with the AX72453 blade standard, suggesting that NFTC-unique flying operations may not have been a dominant causal factor.

1.6.12  Because of their location, LPT blade cracks cannot be detected by first or second line maintenance processes at NFTC.  In response to the root blade failures, RR instituted corrective action through the issuance of a non-modification service bulletin (NMSB) (NMSB 87106-7254-01) by specifying that the hand dressing of the blade corners (radius) be classified as a critical operation that could only be completed by specifically trained and approved personnel.  In addition, RR introduced an independent “overcheck” inspection of the root neck corner radii following dressing.  Eventually a revised NMSB was issued (NMSB 87106-7254-02) to address quality control issues.  The NFTC engines that had completed this modification were referred to as Group B engines.

1.6.13  A previously issued record of airworthiness risk management (RARM) (RARM-155-2007-002 Revision 9), put in place to incorporate the quality control initiatives of RR and the mix of modified/non-modified engines in use at NFTC, was closed in Apr 2010 after all engines in use at NFTC were changed to Group B engines.  The RARM was closed when the risk of LPT blade failure was believed to have been reduced to an acceptable level of safety.  Bombardier Military Aviation Training (BMAT) and the Directorate of Technical Airworthiness and Engineering Support (DTAES) agreed with RR, that the Group B engine LPT blades were suitable for a 2,000 engine flying hour (EFH) life.  In order to provide RR with time to improve their LPT blade inspection plan [7] BMAT limited Group B engines to a 1,000 life.  Engine number 7825 was installed in CT155201.  The LPT blade failed on CT155201 at 617 engine flying hours (EFH).

 

ItemHoursComments
201 airframe hours 4367.6*

*All times are from records available just before the flight. The accident flight was estimated to be 21 minutes long (0.4 hours).

Engine # 7825

time since new (TSN)

3711.3* Nil

Engine 7825

time since overhaul (TSO)

1823.6*

Engine 7825 underwent an overhaul at RR Engine Services at Kingsville Texas in Jun 2006 following an excessive vibration entry in the maintenance log while installed in 208.

Engine 7825

time since LPT replacement

616.9*

25 Sep 2009 Engine 7825 had 94 (P/N AX72453) LPT blades installed in Module 8 at 3100.4 engine hours. These LPT blades are NMSB87106-7254-02 compliant (Rework and Inspection of Blade Root Neck Radius).

Table 3:  Airframe Hours and Adour Engine # 7825 Hours and Inspections

Adour Engine 7825 History

1.6.14  A review of BMAT-supplied (Maintenix) airframe and engine maintenance records was conducted to identify maintenance issues of significance, commencing from the time engine 7825 became a Group B engine.  Of note, BMAT’s engine maintenance ability is limited to replacement of the LPC and stator assemblies on the front of the engine.  Any required maintenance of the engine core results in the engine’s return to RR.

1.6.15  The review searched for bird strikes, over-temperature events, over-speed events, icing events and any unscheduled maintenance that might have been of interest to the investigation.  This action found only two events:

a.  On 5 Oct 2009, engine 7825 was installed in 212.  Event ID 2184768, dated 08 Oct 2009, identified an excessive vibration encountered during the engine post installation ground run.  The engine mounts were re-torqued and the hydraulic pumps, generator, air starter, and exhaust clamps were checked for security.  Vibration test set D [8] pre-use verification was carried out serviceable at 75 microns and it was determined after discussion with the engine technician specialist that the engine vibration was within limits; and

b.  On 4 Mar 2010, engine 7825 was moved and installed in CT155201.  Event ID 2268573, dated 5 Mar 2010, shows the engine encountered vibrations peaking over 50 microns at the 100% power setting.  The No. 1 hydraulic pump was replaced, the jet exhaust pipe brackets were inspected for security, the engine mounts were inspected and re-torqued, and the oil pressure filter was cleaned and examined.  The engine HP and LP rotating assemblies were checked for freedom of movement and no unusual noise was noted.  The magnetic chip detectors were sampled and checked serviceable.  On 11 Mar 2010, the engine was ground run serviceable within vibration limits.

1.6.16  A maintenance inspection was performed on the day prior to the accident flight of 10 Jun 2011, when the three engine chip detectors located in the gear box were analysed in accordance with (IAW) the scheduled 25 hour inspection.  No anomalies were noted.

Avionics

1.6.17  Central Warning System.  The central warning system (CWS) provides visual and audio warnings to the pilot.  Whenever a warning is given, the master warning light in each cockpit flashes, the annunciator panel displays either a red or amber light, depending on the failure, and an audio tone is generated, which is followed by a specific voice message given twice.

1.6.18  Engine Monitor Panel.  Engine indications are displayed on the engine monitor panel (EMP) on the main instrument panel in each cockpit.  If the generator fails, these displays are powered by the battery through the essential services bus.  Engine RPM is displayed both digitally (which goes blank below 12% RPM) and on a monochrome horizontal analogue strip gauge.  TGT is displayed in a similar manner to RPM.  If the TGT rises above 606°C the display will begin to flash as a warning.

1.6.19  Head-Up Display.  The aircraft is equipped with a head-up display (HUD) in the front cockpit and a HUD repeater in the rear cockpit.  The HUD displays flight information including pitch, roll, heading, airspeed/mach, altitude, vertical speed, angle of attack (AOA) and a velocity vector (Figure 9).  The AOA is depicted on an analogue scale consisting of a series of dots and a pointer.  A double dot indicates 5.5 units of AOA, which is approximately the best glide AOA[9].  The display can be repeated on a monochrome HUD monitor in the rear cockpit.  If the generator fails or falls off-line, electrical power to the HUD is lost and it ceases to function.

Figure 9:  Hawk HUD Symbology.  The image shows HUD symbology during a Manual of Flying Training-prescribed PFL pattern and is provided for illustration purposes only.  This figure only presents an approximate image of the accident aircraft’s glide profile.  Note the indicated airspeed and AOA are close to 185 KIAS and 5.5 units; however, the flight path angle is -3.9° instead of -4.7°, which is the flight path angle when the aircraft is established on the ideal glide profile.  The two do not correspond because the image is a still photo of a transient event where there were small up and down flight path angle changes.  Note that the horizon line of the pitch ladder does not align with the actual horizon.  This is because the flight path information is based on a flat earth and curvature of the earth becomes apparent at altitude.  Although this will result in an overestimation of how far the aircraft can glide, those errors are relatively small below 15,000 ft MSL (<2 NM) and decrease rapidly as altitude decreases.

1.6.20  Standby Flight Instruments.  Each cockpit is fitted with a set of standby instruments to show attitude, altitude, speed and heading.  Each cockpit is also equipped with a standby AOA indicator.  These instruments are powered by the essential services bus, which can also be powered by the battery, and are grouped around the regular multipurpose display directly in front of the pilot.

The Ejection System

1.6.21  The Canopy Fragilization System.  The canopy system uses miniature detonation cord (MDC) to fracture the canopy during ejection, for ground rescue assistance or for ground emergency egress.  MDC is a lead-sheathed detonating cord and when it detonates it fractures the canopy into pieces and sprays very small particles of molten lead and sharp Plexiglas throughout the cockpit area.

1.6.22  Ejection Seats.  Each cockpit has a Martin-Baker ballistic catapult/rocket system Mk BA10LH ejection seat that can provide escape capability at all altitudes and speeds within the aircraft’s flight envelope and down to zero speed and zero altitude in a level attitude.  Minimum and maximum crew seat boarding weights are 147 to 238.5 pounds, respectively.  The seats can be fired individually or in an automatic sequence by command ejection where the rear seat is followed by the front seat.

1.6.23  Leg Restraint System.  A leg restraint system is located on the front of the ejection seat.  The function of the system is to secure the occupant's legs to the seat during the ejection.  According to Hawk AOIs, during strap-in, the occupant is to “Adjust the leg restraint lines to give just enough length for the application of full rudder in both directions.  Pull any excess leg-restraint line down through the snubbing units.”  Martin-Baker literature indicates the leg restraint lines are designed to retract the legs from a forward position.

1.6.24  Main Parachute System.  The ejection system incorporates the GQ1000 main parachute.  The GQ1000 is a 17-foot diameter parachute with an average total descent rate under a full canopy of approximately 28’ per second (FPS).  The high descent rate and total velocity of the GQ1000 parachute increases the risk of injury to the aircrew.

1.6.25  A GQ1000 parachute Records of Airworthiness Risk Management (RARM) was initiated on 22 July 11, where it was noted that, “the GQ1000 parachute is installed on a number of different aircraft including other Hawk fleets, and was previously installed on the CF188 prior to the upgrade to the NACES seat.  The GQ1000 is known to have descent rates above the standard accepted and, therefore, carries a higher risk of landing injuries post ejection.  In addition, the GQ1000 is a “forward drive” parachute by design.  This means in addition to the descent rate there will be a forward movement at landing adding to the total landing velocity and increasing the risk of injury.”

1.6.26  A RARM with two revisions has been completed and approved by representatives of the Operational Airworthiness Authority, the Technical Airworthiness Authority and 2 CDN Air Div.  The RARM identified a risk of medium due to higher than accepted descent rates leading to the probability of injury to aircrew when using the GQ1000.  The RARM activity log dated 24 Sep 2013 stated that a GQ5000L replacement parachute was unlikely to be acquired due to budget pressures; however, mitigation to achieve an acceptable level of safety included the installation of an automatically-deployed PSP that reaches the ground before the pilot does, and, thus, lowers the pilot’s landing weight and his subsequent risk of injury.

1.6.27  Personal Survival Pack.  Provisions for survival after ejection are stored in the PSP.  The kit is in a fibreglass container, secured to the parachute harness by a strap that passes through two rings on the pack, one on each side.  Only one connector is required to be released for the PSP to deploy.  The left and right side connectors are located near the left and right lower back region and are not visible to the pilot.  They must be found by feel.

1.6.28  Survivor Locator Beacon 2000.  The survivor locator beacon (SLB) is located inside the PSP and was designed to automatically activate and deploy a 43 inch cable antenna following seat/man separation during the ejection sequence.

1.6.29  Upon ejecting, P2 was whipped by the longer antenna flailing in the airstream.  Though no injury occurred, P2’s outer visor was cracked and layers of his G-suit fabric were cut by the flailing antenna.  P1 did not suffer any injuries from the flailing antenna.

1.6.30  Subsequent to this accident, a modification to the deployment of the 43 inch cable antenna was completed wherein the antenna routing was changed to deploy only upon PSP deployment, currently a manually initiated action.  This configuration change ensures that both the antenna deployment happens later in the ejection sequence when the various components are more stable and the antenna deploys further from the Pilot, thereby mitigating the possibility of the antenna contacting the pilot and resulting in injury.

1.7 Meteorological Information

1.7.1  The weather recorded at CYOD at the time of the accident was as follows:

CYOD 101148 25011KT 12SM FEW 090 FEW 220 23/07 A2990

Forecast Winds valid 18Z, 10 Jun 2011and Issued 12Z for use 11-21Z (Heights MSL):

 

Altitude (’)Direction (T) / Speed (knots)Temperature (°C) 
6,000 210 / 21 8
9,000 190 / 11 2
12,000 200 / 07 - 4
18,000 250 / 09 - 15

Table 4:  Forecast Winds and Temperatures

1.7.2  The local weather at the time of the accident was characterized by generally clear skies, good visibility and westerly surface winds.

1.7.3  The weather was not a factor in this accident.

1.8 Aids to Navigation

1.8.1  All CYOD navigation aids were serviceable at the time of the accident.  The aerodrome is equipped with a TACAN system positioned near the centre of the aerodrome; however, the accident aircraft’s TACAN receiver was selected to an air-to-air mode for the tactical portion of the student training mission, vice providing distance information to CYOD airport.  There were no relevant notices to airmen (NOTAM) that affected operations at the time of the accident.

1.9 Communications

1.9.1  At the time of the accident, all CYOD ATC radios were serviceable and all frequencies were available.  The accident aircraft was fitted with one UHF/ VHF capable radio and another UHF-only radio.  The accident aircraft communicated with both the wingman and CYOD ATC using standard UHF pre-set frequencies.

1.10 Aerodrome Information

1.10.1  CYOD has two principal parallel runways (13/31 left and right) which were both active at the time of the accident.  Runway 13 left/31 right is 12,600’ and runway 13 right / 31 left is 10,000’ in length, both oriented 128ºM/308ºM.  A third runway, 04/22 is oriented 038ºM/218ºM and is 8,270’ in length.  Complete aerodrome information can be found in Annex C.

1.10.2  CYLL has one hard surfaced runway oriented 255ºM/075ºM and is 5,579’ in length.  Complete aerodrome information can be found in Annex C.

1.11 Flight Recorders

1.11.1  From data recovered from the on board engine data acquisition unit (DAU), the investigation determined that the LPT blade loss occurred at 11:35:33 (LT).  At this time, Zulu 21 was in a left turn through 187°M at 11,960ft MSL at 270 KIAS; CYOD bore 359°M at 43.4 NM and CYLL bore 177°M at 30 NM.

1.11.2  When the engine was shut down at 11:39:19 (LT), CYOD bore 341°M at 30 NM and the aircraft was passing 12,755’ MSL at 260 KIAS.  At 11:39:51 (LT), when the aircraft was at its peak altitude of 13,490’, Zulu 21 was 26.8 NM to CYOD.

Video Camera and Recorder

1.11.3  The Hawk aircraft is not equipped with a crashworthy cockpit voice recorder (CVR) or flight data recorder (FDR) but it does have a video camera and recorder system that records audio and the view as seen through the HUD by the camera in the front cockpit.  The video camera and video cassette recorder (VCR) are not crashworthy.  The VCR records the forward view outside the aircraft, the HUD navigation and weapon aiming symbology display and both cockpit intercom and radio transmission and reception.

1.11.4  The video camera is controlled by a video status and control unit on the right console in the front cockpit.  The VCR is located on the right console of the rear cockpit and is powered by the generator.

1.11.5  The VCR unit was recovered from the crash site.  The VCR door was found closed with the rotary latch in the CLOSED position.  The VCR unit was only mildly damaged and was not exposed to excessive heat.  The VCR door was opened normally using the rotary latch.  The VCR cassette tape was never found.

1.11.6  P1 directed P2 to remove the video cassette in an attempt to save it from damage prior to the aircraft crashing.  P2 removed the cassette from the VCR unit and placed it in his lower right G-suit pocket before ejection.

1.11.7  Zulu 22’s video cassette was recovered.  It was found to have captured the entire flight from take-off to shut down.  The communication audio and HUD data from the Zulu 22 video cassette was used in this investigation. 

1.11.8  A CF188 Hornet aircraft, which has a similar HUD recording system, was waiting to take-off when the accident took place.  The CF188 pilot turned on his HUD recorder and captured the radio communications on the ATC frequency during the later stages of the accident flight.  Comparing the ATC frequency communications from the CF188’s video cassette with Zulu 22’s video cassette allowed the investigation to correlate Zulu 22’s video cassette running time to GPS time for comparison with all other recorded data sources.

Data Acquisition Unit

1.11.9  Engine component data is collected by the aircraft DAU for analysis.  The DAU is downloaded every two weeks or upon reaching 20 flying hours, whichever comes first.  BMAT sends RR a monthly report compiled of details about the cyclic exchange rates of select engine components.

1.11.10  The DAU records environmental data, aircraft structural data and engine data at eight samples per second [10].  This device is not crashworthy.  Recording of engine parameters, altitude airspeed and OAT data commences when engine RPM increases above 20% and stops recording when RPM decreases below 15%.

1.11.11  The accident DAU was recovered in the debris field.  The data was downloaded and provided engine data for the duration of the flight until the recorder ceased operating after the engine was shut down.  There was no time stamp on the engine data; however, running time was determined based on eight data samples per second recording rate.  The point where the aircraft started the take-off roll, as indicated by an increase of airspeed, was then correlated with the ACMI data.  Since the ACMI recorded accurate inertial data as well as GPS time, the investigation was able to correlate the engine data with all other recorded data sources.

1.11.12  The DAU data was used to characterize the engine behaviour for the flight up until the point where the DAU stopped recording engine data (Annex A, Figures 1 - 6).  There were no anomalies noted prior to the loss of the accident LPT blade.

1.11.13  At 11:35:19 (LT), RPM was about 90%.  At 11:35:26 (LT), RPM rapidly advanced to full power seven seconds prior to the LPT blade loss event at 11:35:33 (LT).  The loss of the LPT blade was characterized by a momentary increase of RPM to 100% and low pressure rotor speed (NL)[11] to 100% and 108%, respectively, and a TGT increase from 579°C to 611°C and increasing over a period of about 1.25 seconds.  Over the next three seconds, RPM and NL decreased to 82% and 69%, respectively, and the TGT dropped to about 380°C.

1.11.14  Over the next 27 seconds, RPM increased from 82% to 84% then from 84% to 86% in two distinct steps while NL and TGT reached 76% and 400°C respectively by 11:36:04 (LT).  From this point in time until the engine was shut down at 11:39:19 (LT), mild noise of +/- 0.1% RPM was then observed in the RPM and TGT data.  The noisy RPM was likely due to vibrations migrating into the electromechanical sensors as a result of engine/airframe interactions in the presence of the LPT disk mass imbalance.

1.11.15  From 11:36:04 until 11:38:44 (LT), RPM increased linearly from 86% to 89%.  Very small perturbations were noted in the data but these did not affect the overall trend.  Then from 11:38:44 (LT) until the engine was shut down (thirty-five seconds later), RPM had increased to 93%.  After RPM increased to 93% the engine was shut down, with no pause at idle.

1.11.16  After the engine was shut down, RPM, and TGT followed exponential decline curves.  At 11:39:46 (LT), 27 seconds after shutting the engine down as RPM was decreasing through 21%, interference was noted in the RPM signal.  According to RR, this is a known issue observed in flight test data that is caused by the igniters firing and it indicates that a relight was being attempted.  No data was recorded after 11:40:06 (LT) when RPM decreased through 15%.

Air Combat Manoeuvring Instrumentation System

1.11.17  An airborne instrumentation subsystem (AIS) pod was installed on the aircraft’s left wingtip.  The AIS pod is the airborne terminal of the local Air Combat Manoeuvring Instrumentation (ACMI) system.  Provided that 24/28 volt DC power is available, the pod autonomously and continuously determines its own time-space-position-Information (TSPI), including position, velocity, altitude and attitude, by means of GPS and inertial data.

1.11.18  Speed, altitude and AOA data is also provided by a pitot static system located on the front of the AIS pod.  All data is recorded to a data cartridge located in a slot in the aft end of the AIS pod and records at a rate of 10 data samples per second.  ACMI data from Zulu 22 spanned the entire flight.  Zulu 22’s ACMI data included ZULU 21’s data up until 11:39:19 (LT) when Zulu 21’s engine was shut down.  Both sets of data were used in combination with all the other recorded data sets to help recreate and study the flight profile (Annex B, Figures 7, 8).

Air Traffic Control Radar

1.11.19  The CYOD ATC surveillance radar provides continuously-recorded primary and secondary surveillance radar information to ATC personnel.  The ATC surveillance radar data covering the time of the accident was provided to the investigation team for playback using radar playback software.  The radar received Zulu 21’s transponder information (mode 3A code squawk 1221) just after take-off until just prior to ground impact.  The radar ceased receiving Zulu 21’s mode C information at a distance of 29.7 NM from Cold Lake.  The last recorded altitude was 13,100’ MSL and climbing.

1.11.20  At 11:43:10 (LT), Zulu 21’s ACMI pod ceased to record, however, ATC radar data showed that Zulu 21 continued straight-ahead towards the threshold of runway 31 left before gently turned left just prior to impact.

1.12 Wreckage and Impact Information

1.12.1  There were two distinct wreckage locations:   P1/P2 and the ejection seats and the main aircraft crash site.

1.12.2  P1/P2 landed about 30 m south of Highway 28.  The ejection seats landed 100 m to the north of Highway 28.  The rear ejection seat landed on private property in a swamp and was submerged under water.  The front ejection seat landed in an agricultural field.

1.12.3  The main accident site, about one nautical mile to the northwest of where P1/P2 landed, covered an area of level ground straddling both sides of a four-way gravel road intersection.  The aircraft’s initial impact and direction of breakup was along a NW axis running through the road intersection.  The aircraft struck the ground with a shallow rate of descent in a near level pitch attitude, left wing low.  Ground scars indicate an aircraft heading of 296°M.  Overall, the length of the debris field from the initial point of impact to the farthest aircraft part was 228 m.

1.13  Medical

1.13.1  P1/P2 were transferred to the Cold Lake Medical Center and cared for by the 4 Wing Flight Surgeon.  A physical examination indicated both P1/P2 received minor injuries from the MDC splatter and the leg restraint system.  One pilot received a minor injury from the parachute harness.  Both P1/P2 were discharged from hospital post-evaluation.

1.13.2  Toxicology samples were drawn at the Cold Lake Medical Center and later couriered to the Division of Forensic Toxicology, Armed Forces Medical Examiner System, Rockville, Maryland.  P1 urine toxicology was positive for Diphenhydramine.

1.13.3  P1 was grounded on two occasions, 24 Feb 2011 and 29 Mar 2011, for an exacerbation of a chronic medical condition.  There was no documentation that P1 had been seen by a Flight Surgeon to be ungrounded after either of these medically imposed restrictions.

1.13.4  P1 required 100% oxygen to “feel better” during the flight.  P1 was seen by the Flight Surgeon four days after the accident for symptoms of the same chronic medical condition for which he previously had sought treatment.

1.14  Fire, Explosives Devices, and Munitions

Fire

1.14.1  There was no evidence of a self-supporting pre-impact fire.  A post-crash fire consumed surface vegetation and parts of the aircraft aluminum structure.  The fuel fed fire eventually burned an area approximately 100 m by 50 m.  The fire was extinguished by 4 Wing fire trucks and the Municipal District of Bonnyville Fire Department.

Explosive Devices

1.14.2  The emergency flap and gear squibs used to initiate emergency selection of these aircraft systems were not recovered and were likely consumed in the post-crash fire.  All pressurized vessels found intact following aircraft breakup were made safe.

The Ejection System

1.14.3  The command ejection lever was selected to ON and the ejection was initiated by P2 from the rear cockpit.  All cartridge-actuated devices (i.e. ejection handle, MDC striker) and pressure-actuated devices (i.e. rocket motor, timing charges, MDC, inertial reels) in the ejection system were accounted for.  The ejection systems in both cockpits functioned normally.

Munitions

1.14.4  No munitions were carried on this aircraft.

1.15 Survival Aspects

Ejection

1.15.1  Both P1/P2 suspended weights were within the published weight envelope for the ejection seat.  The vertical descent rate of both P1/P2 under the parachute was calculated to be 26 FPS and the combined velocity was calculated to be 32 FPS in still air.  P1/P2 landed in a swampy area.  Had they landed on firmer ground, P1 would most likely have broken one or both legs due his inability to deploy his seat pack.

1.15.2  Both P1/P2 pulled their legs back against the seat front just prior to ejection and received minor injuries to their legs where the leg restraint garter wrapped around the leg.

1.15.3  P1 reported a large triangular void in his harness above his shoulders following parachute opening and during P1’s parachute opening, the front edge of his foam helmet liner was dented and the right bayonet receiver cover of his helmet was cracked.  These marks, together with an associated injury, are indicative of a riser slap to the head.  An examination of the Hawk 2nd generation simplified combined harness (SCH) revealed a broken left side ¼ inch cross-strap bungee.  These features indicated high lateral loading of the harness during the initial parachute deployment.  High lateral loads to the same parachute and similar version of the SCH contributed to a fatality during a high speed ejection from CF188732 on 26 May 2003.

1.15.4  Subsequent to this accident, a 5th generation SCH completed testing by Martin Baker and implementation is scheduled to be completed on CF Hawk aircraft by 1 Oct 2015.  The 5th generation SCH incorporates features that mitigate a triangular void occurring in the shoulder harness and reduce parachute opening shock, thus reducing high lateral loading.

Aircrew Life Support Equipment

1.15.5  P1 was wearing a short-sleeved shirt under his CF-issued flight suit.  The lack of dual layer clothing on his arms and a partially retracted left sleeve resulted in MDC splatter injury.  The pilot was not wearing liners under his CF-issued gloves.  Examination of his Beaufort MK30 life preserver survival vest (LPSV) revealed that MDC shrapnel had penetrated the bladder cover and the orally inflated chamber rendering it unserviceable.  The automatically inflated chamber remained functional and intact.

1.15.6  P2 received minor injuries to his upper body from the MDC splatter even though he was wearing dual layer clothing.

1.15.7  Both LPSVs contained the MS2000 strobe light which was to have been replaced with the Firefly 3 strobe light IAW CF message DCOMD FG 055 211544Z Aug 08.  Both PSPs contained Tylenol© with an expiry date of 2009.  BMAT-supplied content lists for the PSP contained conflicting information concerning the requirement to pack a CF-issued survival knife; neither PSP contained this knife.

Survivor Locator Beacon

1.15.8  In Apr 2011, the PSPs for the Hawk fleet were equipped with the Cobham or ProFind SLB 2000-102A.  This SLB transmits on 121.5, 243.0 and 406 megahertz (MHz) with a GPS position embedded in the 406 MHz signal to help locate downed aircrew.  The SLB was modified locally with an external antenna held in place on the outside of the PSP so that radio signals would be broadcast immediately following seat-man separation.  The modification was completed to rectify unsatisfactory reception of bailout indications following Hawk ejections in 2004 and 2008.

1.15.9  The ejection occurred approximately two miles from the CYOD ATC facility where no signal was received on 121.5 or 243.0 MHz at the time of the ejection.

1.15.10  The secondary 121.5 and 243 MHz signal is transmitted at no more than 0.1 Watts and is designed for final local homing within short range (possibly as short as ¼ to ½ mile).  The primary aid to locate the beacon/survivor is the 5 Watt 406 MHz signal.  GPS coordinates are attached to the 406 MHz transmission as part of this signal.

1.15.11  The main transmitting power of the 406 MHz beacon is directed at reaching the SARSAT network that uses geo-stationary search and rescue (GEOSAR) and low-earth orbit search and rescue (LEOSAR) satellites.  When activated, 5 Watts of a composite signal, comprised of the individual beacon’s GPS coordinates, are transmitted to the satellite network and actioned through worldwide rescue coordination centres.

1.15.12  The 4 Wing Search and Rescue helicopter was able to home a 121.5 MHz signal coming from the front seat SLB 15 minutes post-ejection.  The signal was described as faint but was received at one mile, which was better than the manufacturer’s specifications.

1.15.13  A CF188 orbiting directly overhead the accident site at 3,000’ AGL received two successive tones on 243.0 MHz, lasting one second each, before transmissions ceased.

1.15.14  Both SLB’s 406 MHz signals were first detected by LEOSAR satellite on 10 June at 13:04 (LT), which was one hour and 17 minutes post-ejection occurred.  The delay in receiving the signal was primarily due to a gap in satellite coverage and due to “battery passivation.”

1.15.15  Over North America, particularly at higher latitudes, there is a well-known LEOSAR gap of 1-2 hours that occurs every 24 hours.  It was determined that the accident occurred right at the beginning of this gap in coverage and caused a delay of approximately one hour and 12 minutes until the next satellite came into view and detected the 406 MHz signal.  Once the satellite received the signal, a data package was produced and transmitted to the appropriate search and rescue centre.  This process takes between two and six minutes.

1.15.16  Battery passivation is a brief time period of voltage delay that occurs after a load is placed on the battery.  Battery passivation is used to extend battery shelf life.  Battery passivation on these SLBs has been resolved with a hardware update.

1.15.17  Under ideal conditions both SLB signals should be received by the geo-stationary satellites positioned over the equator within five minutes of activation; however, these signals were never received due to antenna water immersion when P1/P2 landed in a swampy area.  SLB signals are based upon line of sight travel to the satellite system where water, vegetation and state of antenna deployment and orientation may attenuate any transmitted signal.

1.15.18  Subsequent to this accident, a personal locator beacon (PLB) is now carried by front seat aircrew in the left LPSV pocket as an added security measure to aid in search and rescue efforts.  The PLB transmits on 121.5 and 406 MHz with 6 Watts power and will be carried until all Hawk aircraft PSPs are adapted with an ADU-triggering deployment of the SLB antenna.

Emergency Response

1.15.19  The 4 Wing Emergency Response Plan was initiated when ATC was notified by P1 of the gravity of the inflight emergency and the possibility of an ejection.

1.15.20  As soon as was possible, civilian paramedics, military police and ambulance crews coordinated immediate care of P1/P2 onsite and during their transfer to the Cold Lake Medical Center.

1.16 Test and Research Activities

PFL Profiles Flown in the Flight Training Device

1.16.1  PFL profiles were flown in the Hawk flight training device (FTD) to ascertain altitudes, sight picture and verify mathematical glide path calculations.  Also, the FTD was flown to examine the sequence of cautions and warnings presented to aircrew during an engine shutdown scenario.  In-flight relight procedures and use of stand-by flight instruments were examined for context to this accident.  The placement of leg garters was also assessed.

1.16.2  An assessment of aircraft climb performance and engine characteristics was carried out with a serviceable aircraft under flight conditions similar to those of the accident aircraft at Moose Jaw, Saskatchewan.  With speed brake in and landing gear and flaps up, it was found that the rate of climb at 85% RPM was approximately 800’/minute; engine RPM remained stable and was not notably affected by changes in altitude or airspeed.  An assessment of throttle movement for RPM change was also conducted, revealing that an increase of 1.0% RPM would equate to approximately 3.5 mm of throttle movement.

1.17 Organizational and Management Information

NATO Flying Training in Canada

1.17.1  NFTC, a government and industry partnership announced in 1997, was setup to achieve cost-effective pilot training and to change the way the CF purchases and operates training aircraft.  Under NFTC, aircraft are owned by Milit-Air Inc. and are managed by Bombardier Corporation’s subsidiary, Bombardier Military Aviation Training (BMAT).  The CF operates the Hawk IAW CF regulations and orders to provide flight training for CF and foreign aircrew. 

Bombardier Aerospace - Military Aviation Training

1.17.2  The Hawk fleet is contracted by the Department of National Defence (DND) for military operations.  As such, DND is responsible for the technical and operational airworthiness clearances.  Technical airworthiness for the aircraft is conducted IAW the Technical Airworthiness Manual (TAM).  IAW the TAM, DTAES has accredited the BMAT-supplied aircraft fleet management organization (AFMO) to maintain the aircraft.  The AFMO is composed of two main organizations:  an accredited maintenance organization to conduct aircraft maintenance and an accredited technical organization (ATO).  The ATO is led by the Senior Design Engineer to undertake airworthiness engineering and certification activities concerning the fleet.  Continuing airworthiness of the aircraft is ensured by compliance with the BMAT-supplied and DTAES-approved Engineering Control Manual as well as the TAM.

1.17.3  The DND Type Certificate is issued by the Technical Airworthiness Authority.  Type certification for the aircraft is held by Bombardier Aerospace.  RR remains the Design Authority for the Adour turbofan engine used in the Hawk and Martin Baker is the Design Authority for the ejection seats.

419 Tactical Fighter (Training) Squadron

1.17.4  419 Sqn is a lodger unit based at 4 Wing Cold Lake AB; however, the Squadron Commander reports to the 15 Wing Commander located in Moose Jaw.  The squadron conducts Phase IV flying training, part of the NFTC program, to prepare post-wing aircrew for fighter aircraft training.  The squadron is commanded by a CF pilot while qualified flying instructors and trainees come from many international air forces in addition to the CF.

Directorate of Air Contracted Force Generation

1.17.5  The Directorate of Air Contracted Force Generation (D Air CFG), now the Directorate of Air Simulation and Training, was the directorate within the Chief of Air Force Staff that provided technical, strategic and corporate-level coordination for force generation activities that were delivered to the RCAF using Alternative Service Delivery, including the NFTC program.  D Air CFG coordinated changes and additions to the NFTC contract based on operational requirements to meet air force pilot production needs and guided interactions with the contractor to provide the service necessary to conduct pilot training.

1.18 Additional Information

Hawk 205 LPT Blade Loss Occurrence

1.18.1  On 4 Feb 2009, Hawk 205, with engine 7818 installed, was operating 55 NM northeast of CYOD at low altitude when it suffered a LPT blade failure and loss.  During a break turn manoeuvre, the aircrew heard a loud machine gun-like noise and a “T6NL, T6NL” audio alert.  The engine had entered a locked-in surge and the pilot shut the engine down and successfully performed an immediate relight.

1.18.2  While turning the aircraft back towards CYOD, the pilot slowly advanced the throttle; however, at approximately 88% RPM, significant vibration was experienced in the cockpit.  The pilot elected to reduce throttle to 85% RPM which reduced the vibrations while still allowing the aircraft to climb slowly between 1-2 degrees at 210 KIAS (approximately 800’/minute).  The aircrew reported an engine-like smell enroute and both aircrew selected 100% oxygen.

1.18.3  The aircraft reached approximately 10,000’ MSL by the time it was in a position for the aircrew to carry out at straight-in PFL to runway 22.  The pilot selected the throttle to idle and maintained 5.5 units AOA until the threshold of the runway was situated at 15 degrees dive angle in the HUD.  At that point, the pilot selected the landing gear down and full flaps and performed an uneventful straight-in forced landing (FL) to runway 22.  The engine ran for 12 minutes from the LPT blade failure until the aircraft touched down.

1.18.4  The investigation determined that a single LPT blade had been lost, Figure 10.  Both the HP and LP spools were free to rotate with no signs of distress, and although the oil tank was full it was contaminated.  There was no visual damage to the LP compressor stage 1 blades; however, there was a significant rub on the casing inline with the blade track over the complete circumference (See Figure 10)

Figure 10:  Hawk 205 LPT Disk.  Note a single blade loss with very little collateral damage to the rest of the LPT blades.  The mass imbalance was equivalent to a single 64 gram blade.

Hawk Forced Landing Procedures

1.18.5  The Manual of Flying Training (MFT) contains a chapter on FL.  FL patterns are designed to keep the pilot within a safe ejection envelope until a safe landing can be made.  It states the gliding profile of the Hawk is optimised at 5.5 units AOA (approximately 190 KIAS, depending on weight) and that it should glide about 2 NM/1000’ of altitude loss in still air.

1.18.6  The MFT states FL’s are to be practiced with both the HUD available and not available.  With an actual engine failure, the HUD will not be available.

1.18.7  For engine malfunctions outside the traffic pattern, the MFT directs the pilot to carry out the appropriate emergency response, climb the aircraft (trading airspeed for altitude), and establish the aircraft in a glide at 5.5 units AOA.  The pilot should then turn towards the nearest suitable airfield and assess the glide profile to determine if a forced landing is possible.

1.18.8  To determine the maximum gliding distance, the MFT states to calculate the distance to fly using the best available means of distance measuring equipment (DME) or ATC radar assistance, and divide this number by two, multiply the result by 1000, then add the appropriate key in feet MSL.

1.18.9  If unable to accurately determine distance, the MFT directs the pilot to visually assess whether the aircraft can glide to the airport.  This method is based upon judging movement of the runway relative to a fixed position in the windscreen or to the velocity vector if available.  Over a short period of time it should become apparent if the runway appears to move up or down from the position, allowing the pilot to determine if the runway can be reached.

1.18.10  Arriving overhead the runway of intended landing the pilot will fly a descending racetrack pattern to “keys,” of which there are three (high, low and final key).  Keys allow the pilot to judge if he is within parameters to make a safe landing from the FL.  For example, to complete the FL pattern from high key the aircraft must have sufficient altitude to glide approximately 10 NM beyond the airfield.  If this distance is not available, the aircraft must be able to glide at least six NM beyond the airfield to carry out the FL directly from low key.  If this distance is not available, the pilot should assess whether final key can be reached directly.

1.18.11  The MFT states aircrew should initiate the forced landing check prior to reaching high key, if time permits.  If altitude is insufficient and reaching high key is not assured, aircrew can proceed directly to low key, final key, or eject if a safe landing is not assured (Figure 11).

Figure 11:  Forced Landing Procedure.  Taken from C-12-HWK-000/MC-000 Hawk Pilot Checklist, Revision Basic Change 6.

History of Damage to MK30 LPSV

1.18.12  The LPSV is supplied by DND for use on the aircraft.  To date, six LPSVs have been worn during Hawk ejections.  During the two ejections from aircraft 155202 in May 2004, one of the LPSVs was rendered unserviceable.  In the Apr 2008 ejection, one of the LPSVs worn by an aircrew member sustained a punctured bladder.  Prior to this accident the LPSV was modified to use a ballistic nylon bladder covering.  Subsequent to this modification, one of the LPSVs worn during this accident received a puncture to the orally inflated chamber.

1.19 Useful or Effective Investigation Techniques

1.19.1  A civilian bystander uploaded three videos to the social media website YouTube, capturing P1/P2 post-landing activities including the arrival of emergency medical services.  The videos provided the investigation with factual events of P1/P2’s post-ejection recovery.

Top of page

2. ANALYSIS

2.1 General

2.1.1  The analysis examines the LPT blade history and failure, engine condition and malfunction, emergency airfield selection, engine relight attempts, human performance during emergency handling, visual assessment the FL glide profile, P1/P2 ejection altitude, P1/P2 medical condition, and aircrew life support equipment (ALSE).

2.2 Management

LPT Blade History

2.2.1  There have been a number of LPT blade failures in the NFTC Hawk fleet beginning in Nov 2005.  Since that time, four LPT blade failures were identified and proactively addressed.  Three failures were resolved prior to this accident.  The three resolved failures were mid-span blade failures as a result of LPT blade manufacturing inconsistencies and blade tip rubbing.  Since resolving these issues, no similar failures have occurred.

2.2.2  The failure mode of the fourth LPT blade was believed to have been resolved after 205 lost its LPT blade in Feb 2009.  In that instance, manufacturing issues resulted in quality control escapes where the root radius was not conformant to specifications.  Significant OEM quality control changes were made to the manufacturing process, and a joint CF and BMAT on-site audit of the OEM’s processes re-established confidence that there would not be any further occurrences of this nature.  The most recent root failure of a conforming LPT blade at 617 hours on 201 indicates that while a conforming root radius is a necessary condition for achieving the blade design life of 2,000 hours, it alone is not sufficient to achieve this life.

2.2.3  In spite of the proactive approach to managing this issue by all parties involved, the subtlety of the numerous factors contributing to LPT blade root failures prevented a full understanding of the nature of the problem.  After considering all of the issues regarding this blade root failure, the investigation found that the CF management of the LPT blade root failure issue was not a factor in this accident.

2.3 Technical

LPT Blade Failure

2.3.1  Previous analyses have shown the peak stress within the LPT blade to be located within the under-platform area at the rear pressure side shank fillet.  The nominal stress in this area of the blade with a conforming radius is 1610 mega Pascal (MPa); the current RR design practice of the under platform features is not to exceed 1000 MPa.  RCA revealed that there were a number of possible sources of increased stress in the area in question.  While any one source of increased stress might not have resulted in fatigue developing, a high nominal stress in the presence of some undetermined collection of these sources of increased stress likely reduced the inherent safety margins to zero, resulting in fatigue developing in some blades.

2.3.2  While it might not be possible to eliminate all the sources of elevated stress identified by the RCA, the most effective solution is to redesign the LPT blade so that it conforms to updated RR design practices.  This will re-introduce sufficient safety margins to accommodate variations in stress from other sources.

Engine Condition

2.3.3  Overall, engine 7825 (201) was found to be in good condition.  Bearings and seals showed no signs of distress (similar to engine 7818 (205)).  Engine 7825 LPT disk was found in a similar condition to engine 7818 and the mass imbalance was less than 7818.  Therefore, imbalanced forces acting on the LPT disk bearing were less in engine 7825.  Engine 7818 produced sufficient thrust to climb at 85% RPM for approximately nine minutes and then ran at idle for an additional three minutes during the straight-in PFL before being shut down after landing.  Given that engine 7825 was in similar condition to engine 7818 and was producing correct thrust for the operating RPM, it is reasonable to assume that engine 7825 would have produced 85% climb thrust for at least nine minutes.

Vibration Characteristics

2.3.4  The mass imbalance of the rotating LPT disk would have resulted in a dynamic interaction between the engine and the airframe through the engine mounts as the rotating mass imbalance excited natural airframe response characteristics.  The vibration characteristics would have varied with RPM since both the rotational frequency and the centripetal force would have also changed with RPM.  The vibrations felt by the P1/P2 in the cockpit were the airframe’s natural response to the rotating engine mass imbalance and not a direct reflection of the engine’s health.

Smoke Source Prior Engine Shut-Down

2.3.5  The LPT disk mass imbalance would have caused a slight deflection of the LPT shaft aft of the LPT bearing.  As the engine RPM increased, the centripetal force increased and therefore the LPT shaft deflection increased.  A point was reached where the deflection was likely sufficient to cause a slight leak in the fore and aft seals in the aft bearing housing (LPT and HPT bearings).  Oil escaping from the front oil seal would have entered the anti-icing tube and moved forward to be ejected from the engine’s bullet nose.  From there, some of the oily air would have been bled off for cabin pressurization, explaining P2’s report of oily smell in the cockpit.  Oil leaking out the aft seal would have mixed with hot LP cooling air where it would have partially burned before being dumped out the LP turbine cooling outlet vent mounted to the lower right side of the aircraft.  The smoke would have then moved aft and merged with the engine’s exhaust and would appear to Zulu 22 as if the smoke was coming out of the engine.  The presence of smoke was likely the direct response of the mass imbalance causing a leak in the oil seals as engine RPM increased; not a reflection of deteriorating engine health.

Smoke Source Post-Engine Shut-Down

2.3.6  As described previously, the original source of smoke was likely due to oil leaking from the HPT/LPT bearing housing being partially burned by hot LP turbine cooling air before being dumped overboard.  When Zulu 22 reported that the smoke was returning two minutes and two seconds after the engine was shut down, the RPM was low below 15%, and the engine had cooled considerably.  Therefore, the conditions for producing oil smoke were not present at that time.  Since the return of smoke coincided with the attempted engine relight, and since there is no evidence to suggest another source of smoke, the investigation concluded the most likely source of the return of smoke was unburned fuel being released out the jet pipe.

2.3.7  Given that there was no evidence to suggest a fuel leak, the LP fuel cock must have been ON and the throttle must have been at or above idle to release fuel from the jet pipe.  Therefore, the return of smoke at this point in time indicates that the LP cock was ON and that the throttle was at least at idle (see paragraph 2.4.13).

Increasing RPM

2.3.8  Following the loss of the LPT blade, the crew set the throttle to 86% RPM.  Over the next 160 seconds, the engine RPM increased at a rate of about 0.019 %/second until 89% RPM.  Then over the next 35 seconds it increased more rapidly at a rate of 0.11 %/second until 93% RPM.  Flight tests indicated that the increase in RPM was not related to the changing altitude and airspeed during the climb.

2.3.9  Furthermore, measurements of throttle position versus RPM under similar environmental conditions identified that the top of the throttle (position of the left hand) would need to have been advanced at a rate of about 0.07 mm/second (15 seconds to move 1 mm) for the first 160 seconds and at a rate of 0.4 mm/second (3 seconds to move 1 mm) for the next 35 seconds, until the engine was shut down.  The investigation considered it to be improbable that a pilot could consistently apply such a slow throttle advance over the long period in question; therefore, a mechanical explanation for the RPM increase was considered likely.

2.3.10  The rigid throttle linkage runs from the cockpit along the left side of the aircraft to the point where it first aligns with the FCU before turning inboard and connecting to it.  With the throttle friction not set high enough to overcome all movement, it was probable that airframe/engine vibrations caused, by the imbalanced LPT disk moving the engine left/right (when viewed forward) as the engine rotated, could have induced very small fore and aft inputs to the throttle through the linkage.  It was unlikely that there would have been a perfect balance between movements of the linkage in both directions.  Therefore, a slight imbalance of friction/freeplay in the throttle linkage between the fore and aft directions could have resulted in the throttle advancing very slowly in a consistent direction.  This would have resulted in an overall accumulation of movement in one direction, in this case in the direction of advancing RPM.  A throttle advancing of its own accord at the rates recorded by the DAU would have been physically imperceptible to the pilot who would have had to notice the RPM increase on the EMP.

2.3.11  The increasing RPM was likely due to a mechanical interaction of the airframe/engine with the throttle linkage in the presence of airframe/engine vibrations that resulted from the LPT mass imbalance caused by the blade failure and departure.  It is likely that after setting the RPM to 86% RPM, the crew did not notice the very slowly increasing RPM as they were pre-occupied managing the emergency situation.

Engine Summary

2.3.12  The investigation determined that engine 7825 (201) was found in good condition, similar to 7818 (205).  Engine 7818 produced climb thrust at 85% for nine minutes, therefore, it is likely that 7825 would have also produced normal 85% thrust for about nine minutes had it not been shut down three minutes and 46 seconds after the LPT blade failure event.

2.3.13  The crew felt increasing vibrations in the cockpit.  These were the result of the undetected slowly increasing RPM.  P1/P2 associated the increasing vibration with worsening engine conditions.  When the wingman reported the engine starting to trail smoke, the vibration also markedly increased.  This likely confirmed their belief that the engine condition was deteriorating rapidly, which the investigation believed was not the case.

2.3.13.1  Unable to verify the actual condition of the engine, other than the engine instruments and CWP displays that indicated normal engine behaviour, the crew mentally extrapolated the perception of worsening conditions to the point where they believed the engine would fail catastrophically.

2.3.13.2  The investigation determined the crew perceived a catastrophic engine failure was about to occur.  As a result, they felt it necessary to shut the engine down.

Glide Profile

2.3.14  The stated glide ratio of the aircraft is 2 NM/1,000’ of altitude loss [12], though the accident aircraft actually achieved a slightly better glide ratio of 2.12 NM/1,000’.  In paragraph 2.3.12 it was found that the engine would likely have produced normal 85% RPM thrust for at least nine minutes.  Since the engine was shut down three minutes and 46 seconds after failure of the LPT blade, it is likely that the engine would have produced usable thrust for an additional five minutes and fourteen seconds.  Therefore, given that final key was 2,500’ AGL, from the point where the engine was shut down (30 NM from CYOD) Zulu 21 could have intercepted the glide profile to reach final key, from which a safe landing was possible, had the engine shut-down been delayed by two minutes.

2.4 Hawk Emergency Handling and Checklists

2.4.1  The Hawk Aircraft Operating Instructions (AOI) - C-12-HWK-000/MB-000 and the Pilot Checklist C-12-HWK-000/MC-000 differ in their definitions of land as soon as possible and land as soon as practical, as depicted in Table 5.

 

 Land as Soon as PossibleLand as Soon as Practical
Aircraft Operating Instructions

 

“a landing at the nearest suitable landing area considering severity of the emergency, weather conditions, airfield facilities...”

“emergency conditions are less urgent…an immediate landing may not be necessary.”

Pilot Checklist

 

: “land at the first site at which a safe landing can be made.”

“The landing site and duration of flight are at the discretion of the pilot.”

 

Table 5:  Definitions of Land as Soon as Possible and Land as Soon as Practical.

2.4.2  Many, but not all, airborne red page emergency checklist procedures end with a statement of “Land as soon as possible,” or “Land as soon as practical.”  P1 thought the engine problem was a liberated LPT blade for which there is no dedicated emergency response.  The most likely checklist response for LPT blade liberation is the engine malfunction checklist.  This checklist response specifies neither land as soon as possible, nor land as soon as practical.  P2 thought the engine malfunction was an engine surge, which is a land as soon as practical checklist response.

2.4.3  Following this accident the Engine Malfunction checklist was amended and an Engine Mechanical Damage checklist was added.  Both checklists provide direction specifying criteria for Land as Soon as Possible and Land as Soon as Practical, depending on the nature of the engine malfunction.  The Engine Mechanical Damage checklist now provides a checklist response for engine blade liberations.

Emergency Airfield Selection CYOD/CYLL

2.4.4  CYOD provided three runways (12,600’, 10,000’, and 8,270’), while CYLL had a single 5,579’ runway (Annex C).  CYOD also provided a known airport with a practiced forced landing pattern that was well known by the crew in addition to crash, fire and rescue services.  Bonnyville airport (CYBF), located 16 nautical miles southwest of CYOD, provided the least desirable runway alternative, a single runway 4,433’ long by 75’ wide.  P1/P2 briefly considered it as a possible divert airfield but then assessed it to be too dangerous due to the available runway length given the required landing distance.

2.4.5  The investigation determined that had the P1/P2 diverted to CYLL and flown the identical CYOD glide profile, they would have had enough altitude to reach high, low, or final key. 

2.4.6  The investigation computed Hawk landing roll data for CYLL based aircraft configuration, aircraft weight, a 135 knot landing speed, maximum braking, and weather for the day of accident.  Calculations determined the aircraft required more runway than was available at CYLL with or without a functioning drag chute.  Runway overrun data varied from 400’ to 2,000’, based on drag chute functionality and how far beyond the runway threshold P1/P2 would have touched down, up to and including the Hawk MFT 1/3rd of the distance of the runway available [13].

2.4.7  Landing on a short runway at CYLL would have exposed P1/P2 to ejecting while at zero altitude and slow speed, thereby increasing the risk of injury or death to one or both pilots.  The investigation determined P1/P2’s decision to divert to CYOD was reasonable.

Engine Relight Attempts

2.4.8  The investigation determined that most likely two relight attempts were made after the engine was shut down, the last of which led the wingman to report “flames, flames.”

2.4.9  P1 carried out the engine mechanical failure drill and selected the LP cock off (Figure 2, paragraph 1.1.9).  The DAU confirmed the engine igniters were then activated 27 seconds after shutdown.  The investigation determined igniter activation occurred during a cold relight attempt due to sufficient RPM (Figure 12).  The cold relight was completed by pressing the relight button and selecting the throttle to idle.

Figure 12:  Cold Relight Procedure.  Taken from C-12-HWK-000/MC-000 Hawk Pilot Checklist, Revision Basic Change 6.

2.4.10  This course of action was unknown to P2 who had begun calling out the assisted relight checklist procedure.  P2 was unaware a cold relight was attempted due to a lack of communication which is analysed in detail in the following section (2.4.17 – 2.4.24).

2.4.11  The cold relight attempt was unsuccessful because P1 had previously selected the LP cock to OFF, per the engine mechanical failure checklist procedure.  P1 completed the initial steps of the cold relight procedure, but most likely did not complete the entire cold relight procedure, which directs placing the throttle back to cut-off if no relight occurs.

2.4.12  Subsequently, P1 identified the LP cock was OFF and selected it to ON.  The throttle was most likely still at idle, and now with the LP cock ON, fuel started to flow and vaporized out the jet pipe.  At this time the wingman reported “smoke returning” because the fuel vapour was assessed incorrectly as smoke.

2.4.13  P2 was advised by P1 the LP cock had been selected OFF, and it was selected back to ON.  P2 re-started reading the assisted relight procedure (Figure 13) from the beginning with P1 carrying out the actions.  During this assisted relight procedure, when P1 pressed the relight button, excess fuel was ignited in synch with the igniters producing a pulsating flame exiting the jet pipe.  Immediately, the wingman reported “flames, flames.”

Figure 13:  Assisted Relight Procedure.  Taken from C-12-HWK-000/MC-000 Hawk Pilot Checklist, Revision Basic Change 6.

2.4.14   P1/P2 without delay terminated the restart attempt with P1 carrying out the engine mechanical failure checklist procedure due to both crew members believing that the engine was critically damaged.

2.4.15  The investigation determined that given the relatively good condition of the engine, there was no evidence to suggest that the engine would not have started had either of the relight procedures been correctly executed.

Engine Malfunction Response - HPMA

2.4.16  The investigation examined the emergency responses and procedures undertaken by P1/P2 in order to identify possible Human Performance in Military Aviation (HPMA) issues associated with those responses.  Crew communication, also known as cockpit resource management was examined to determine its impact on the emergency handling by P1 and P2.

2.4.17  P1 was certain that an LPT blade failure had occurred while P2 thought that an engine surge had occurred.  Neither crew member conferred with the other to ascertain what they believed to be the engine problem or what checklist response should be completed.

2.4.18  P1 completed electrical load-shedding without advising P2 of his actions.  P2 notified P1 of a yaw damper caution light appearing on the CWP which was a result of P1 load-shedding.  P1 then informed P2 of his actions regarding load-shedding.

2.4.19  When the decision was made to shut down the engine, P1 carried out the engine mechanical failure procedure without checklist challenge or response by P2.  P2 was unaware the LP fuel cock had been selected OFF.

2.4.20  P1/P2 next decided to restart the engine.  Neither crew member communicated to the other which checklist procedure to follow.  P1 completed a cold relight procedure while P2 began calling out the assisted relight procedure.  The engine, however, did not relight due to the LP fuel cock having been selected off.

2.4.21  P2 was not aware that the cold relight procedure had been executed.  P2 was on a different checklist page and therefore, could not provide oversight of the cold relight checklist procedure.  When the cold relight attempt failed to start the engine, P1 most likely forgot to select the throttle to OFF and P2 could not correct this checklist error.

2.4.22  P1 then realized the LP cock was OFF and selected it to ON, informing P2 of his actions.  P2 restarted the assisted relight procedure from the beginning with P1 carrying out the actions.  With the LP cock ON and throttle at idle, fuel began to flow out the jet pipe, which led the wingman to report “smoke returning.”  Once the relight button was pressed, excess fuel partially ignited, which led to the wingman to report “flames, flames” and caused the engine start to be aborted by P1/P2.  No further relight attempts were made.

2.4.23  The investigation team determined that the breakdown in crew communication that occurred during P1/P2 emergency handling resulted in the unsuccessful engine relight attempts.

Visual Assessment of Glide Profile

2.4.24  Although DME and ATC radar were available to P1/P2, the crew reported that they did not use either of these references to assess their glide profile to attempt the FL, but instead performed a visual assessment.

2.4.25  As discussed in detail in section 1.18.5 to 1.18.11, the Hawk MFT states that to determine the maximum gliding distance, calculate distance to fly using the best available means (DME, radar or visually).  If using DME or radar calculations, divide distance to go by 2, multiply the result by 1000 and add the key altitude in ft MSL.

2.4.26  The MFT does not highlight a number of significant errors when estimating the glide profile visually.  The MFT describes some considerations when assessing the glide profile visually, but does not explain how various flight parameters can affect calculating glide profile visually.

2.4.27  It states the visual method is based on judging movement of the runway relative to a fixed point in the windscreen and that over a short period of time, it should become apparent that the runway appears to either move up or down from this position.  It also states that to reach a key, the aircraft must be able to glide a distance beyond the airfield by an amount appropriate for that key.

2.4.28  To reach high key, the MFT states that the aircraft must be able to glide approximately 10 NM beyond the airfield and to reach low key the MFT states that the aircraft must be able to reach 6 NM beyond the airfield.  The MFT does not indicate how far beyond the airfield the aircraft needs to be able to glide to in order to reach final key, which the investigation determined to be approximately 3 NM.

Visual Assessment – Engine Failed

2.4.29  Factors affecting visual assessment with the engine failed are:

a.  Until the aircraft is established on a steady gliding attitude and airspeed, the pilot will have a false sense of where the aircraft is capable of gliding.  In this accident the aircraft was in a slight climb and at a higher speed than the recommended 190 knot glide.  Therefore P1/P2 would have had an overly optimistic visual assessment of reaching CYOD;

b.  When established on glide parameters, any pitch inputs will change the visual aim point, either short or long, and the error induced will be exacerbated at increased range; and

c.  be able to glide to the point beyond the airfield appropriate for that key.  The point beyond the airfield must be visually estimated, adding additional uncertainty to this method of assessment.

2.4.30  The investigation determined that the visual method of estimating glide profile for a FL has inherent problems:  the degree of uncertainty identifying a fixed point in the windscreen, the impact of small pitch inputs on the apparent distance the aircraft can glide and the requirement for the pilot to estimate to a point beyond the airfield.  These highlight that it is very difficult for a pilot to accurately assess, through visual means alone, whether or not the aircraft can reach a given key.  (Additional analysis of the visual assessment is found in Annex E).

Visual Assessment – Engine Operating

2.4.31  Factors affecting visual assessment with the engine operating are:

a.  When the engine is operating, the HUD is available and, therefore, a velocity vector and a pitch ladder are available for reference; and

b.  The use of the HUD velocity vector reduces errors when visually identifying the point on the ground that the aircraft can reach, however, visually estimating the gliding distance beyond the airfield is required in order to determine whether the aircraft can reach a particular key.  (Additional Analysis of Visual Assessment is found Annex E).

Visual Assessment Contribution to Engine Shut Down

2.4.32  In this accident, P1/P2 determined their glide profile by visual means and believed they were close to a glide profile to reach the runway.  In reality the aircraft was approximately 4,000’ below the calculated profile to reach final key, without considering manoeuvring or the selection of landing gear and flaps.

2.4.33  The investigation determined that had P1/P2 maintained the climb for approximately two more minutes, the aircraft would have intercepted the glide profile for final key.  Had P1/P2 calculated their glide profile using DME instead of visually assessing it, they would have been aware of what minimum altitude was needed in order to reach a key position for a successful FL.  It is possible they would have considered maintaining engine operation until they reached a suitable altitude for a FL.

Visual Assessment – Summary

2.4.34  The investigation determined that visual assessment introduces significant errors when calculating the aircraft’s glide profile and likely contributed to P1/P2 prematurely shutting the engine down.  It is an inferior method when compared to using DME or radar distance to calculate the FL glide profile.  Given the importance of determining the glide profile, if actual distance information is available to the pilot, that information must be used to calculate the aircraft’s glide profile.  Visual means should only be used as a last resort when more accurate means are not available.

Ejection

2.4.35  Descending through 10,000’ MSL, P1/P2 were aware that there existed insufficient altitude to make any runway at CYOD and that they would, therefore, eject; nevertheless, the controlled ejection checklist was not completed even though sufficient time was available.  Upon passing 5,000’ MSL (3,000’ AGL), P1/P2 had 33 seconds remaining prior to reaching their minimum ejection altitude of 4,000’ MSL (2,000’ AGL) [14].  However, it became clear to the investigation that P1/P2 were focussed on non-checklist items rather than ejecting at or above the minimum ejection altitude.

Figure 14:  Controlled Ejection Procedure.  Taken from C-12-HWK-000/MC-000 Hawk Pilot Checklist, Revision Basic Change 6.

2.4.36  P1 removed his wrist watch and stowed it in his flight suit pocket.  This action negatively affected his scan outside the aircraft and his cockpit scan, including the standby altimeter.  Also, P1 mistook 2,000’ MSL on the standby altimeter as their minimum ejection altitude of 2,000’ AGL, which added to the delayed decision to eject.

2.4.37  During this critical time period, P1 requested P2 to eject the cockpit VCR tape depicting the audio and HUD information and store it prior to ejecting.  P2 then became occupied with ejecting the tape from the VCR housing and stowing it into his right-leg lower G-suit pocket, which was then sealed by the Velcro fastener.  This action negatively impacted P2’s exterior and interior scan as his attention was directed towards the right-rear instrument consol where the VCR housing was located.  Valuable seconds were lost unlatching the VCR door, ejecting the tape, latching the VCR door, and then stowing the tape in the G-suit pocket.

2.4.37.1  There is neither a requirement, nor a checklist procedure that directs aircrew to eject data storage devices prior to ejection.  Additionally, the investigation could not identify any historical case in which such a data storage device had been collected prior to ejection.  It was, therefore, not clear to the investigation why P1 and P2 diverted their attention away from preparing for a safe ejection by following the ejection checklist, when ample time existed to do so.

2.4.37.2  Prior to ejection, P2 secured the VCR tape in his G-suit pocket.  The investigation discovered a YouTube video in which, post-ejection and while standing on the road awaiting 4 Wing response vehicles, P1 asked P2 if he had the VCR tape.  While seen patting his G-suit pocket that contained the tape, P2 replied that he had the tape.  At some point the ALSE equipment was loaded into a support vehicle by unknown person(s); after its arrival at the 4 Wing ALSE shop the investigation requested the VCR tape.  The tape was missing and was never found.  The VCR tape’s disappearance impeded the investigation’s ability to analyze the P1/P2 interaction during the handling of this emergency.

2.4.38  Lastly, P2 was reluctant to eject, as a “loss of control” would occur, whereby he would not have control of the outcome of the ejection, but would be at the mercy of “the ejection system operating automatically.”  This reluctance may also have contributed to delaying the decision to eject until below 2,000’ AGL.

2.4.39  The investigation determined that sufficient time was available to carry out the appropriate checklist items when it became obvious a controlled ejection was going to be necessary.  P1/P2 should have focused their attention on the impending ejection and not concerned themselves with personal non-checklist items.  This distraction resulted in P1/P2 ejecting at an altitude between 2,350 and 2,550’ MSL, or based on the AETE crew systems investigation report between 550’ and 750’ AGL, which was well below the 2,000’ AGL recommended minimum altitude.

2.5 Fitness to Fly

Fatigue

2.5.1  Multiple factors associated with both acute and chronic fatigue were assessed:  chronic sleep debt, recent sleep issues, extended periods of wakefulness prior to the accident, multiple time zone crossings prior to the accident, or night time work.  However, the investigation determined that fatigue was not a factor for either P1 or P2 in this accident.

P1 Medical Fitness to Fly

2.5.2  P1 was grounded on two occasions in the months prior to this accident.  Post-grounding, there were no records of follow-up appointments for re-assessment and there was no documentation indicating that he was seen again or otherwise ungrounded.  P1 was not seen by a health provider in the weeks leading up to the accident, but he was seen four days after the accident for the same condition that he had previously been “grounded” for.

2.5.3  IAW B-GA-100-001/AA-000, Chapter 9, paragraph 1, aircrew who are medically unfit to fly must be assessed by a Flight Surgeon before they are allowed to resume flight duties; this is repeatedly reinforced to all pilot aircrew from Phase I training onwards.  Once either grounded or cleared to resume flying, a CF2018 CF Health Services Chit is provided to the member for their unit’s chain of command primarily so that operations schedulers and flying supervisors are aware.  In some exceptional cases, health providers may allow pilots to “self-unground,” but this needs to be clearly documented on both the health record and the CF2018 CF Health Services Chit.  Regardless, there was no documentation in P1’s health records that showed he was either ungrounded or able to resume flying duties or that he was able to self-unground.  There was no copy of a CF2018 CF Health Services Chit relevant to these groundings.

2.5.4  P1 utilized 100% oxygen during the accident flight in an effort to improve how he felt while flying.  B-GA-100-001/AA-000, Chapter 9, paragraph 1, states, “An aircrew member shall not fly when feeling unusually fatigued or suffering from any physical or psychological illness or injury (except minor cuts, scrapes, etc.) without the prior approval of a CF medical officer.”

2.5.5  P1’s toxicology sample tested positive for Diphenhydramine, a common over-the-counter antihistamine that is often taken at night.  There was no record of P1’s prescription by a medical officer to consume Diphenhydramine while conducting flying operations.  B-GA-100-001/AA-000, Chapter 9, paragraph 2, states, “An aircrew member shall take only those drugs (prescribed and over-the-counter), patent medicines and pharmaceutical preparations that are authorized by a medical officer and taken under supervision.”  Paragraph 3 is even more direct and clearly states, “Under no circumstances shall an aircrew member be permitted to fly while under the influence of any drug without the flight surgeon's prior approval.”  It was not known when P1 ingested the Diphenhydramine but, given its concentration level in the toxicology, it was probable that it was ingested 12 to 36 hours prior to the accident.

2.5.6  Together, the use of oxygen during the flight and the confirmation of Diphenhydramine in his toxicology results lead the investigation to conclude that P1 was likely suffering from a medical condition that he did not report to medical authorities.

2.5.7  IAW Flight Surgeon Guidelines 1900-01, “Medications and Aircrew,” Diphenhydramine is unsuitable for use in aircrew because of its anticholinergic and sedative effects.  Patients may be impaired with respect to attention, memory, vigilance and perception of speed even if they are not experiencing the effects of sedation.  These effects persist into the next day after drug ingestion.

2.5.8  Prior to and post-accident, P1 was prescribed an approved medication for his condition by a flight surgeon.  This medication is authorized for use by aircrew under the supervision of a qualified health provider.  P1’s toxicology was negative for this approved medication.

2.5.9  This investigation could not conclusively determine the extent to which P1’s performance during the accident flight was affected by the unauthorized use of Diphenhydramine.  Nonetheless, current orders prohibit its use by aircrew because it causes degradations to human performance that are known to be incompatible with safe and effective flying.  Even so, it can be said that the HPMA breakdowns during the accident flight (the decision to shut down the engine while 4000’ below the FL profile, the uncoordinated and unsuccessful attempts to relight the engine, the inaccurate FL profile determination, the carrying out of non-checklist pre-ejection procedures, the non-completion of the ejection procedure, and ejecting below ejection altitude criteria) were consistent with those actions of an individual under the influence of Diphenhydramine.

P2 Medical Fitness to Fly

2.5.10  Toxicological results for P2 were normal.  The investigation determined that P2 was fit to fly at the time of the accident.

Aeromedical Support

2.5.11  The Flight Surgeon who treated P1 prior to the accident did not document that P1 must return to be ungrounded before he could return to flying duties.  In addition, The Flight Surgeon chose to hand write his clinical notes instead of typing them into the electronic health record as required.

2.5.12  For this reason, no CF 2018 Health Services Chit documented the grounding and it could not be determined if one was printed and given to the member to take to his unit.  A hand written CF 2018 Health Services Chit with the grounding may have been given to P1 to take to his unit but this was not scanned into the health record.

Prevention of the Use of Substances Hazardous to Aviation

2.5.13  A statistical analysis of Directorate of Flight Safety (DFS) Flight Safety Occurrence Management System (FSOMS) data indicated that human factors play a role in approximately 45% of air occurrences and 80% of ground occurrences, or about 63% of all occurrences [15].  Over the years, the RCAF has spent considerable effort to mitigate the threat to mission accomplishment that human factors represent through the ongoing efforts by organizations like DFS, CF Environmental Medicine Establishment, and Director of Aerospace Engineering Program Management (FT6), amongst others, and by programs like Crew Resource Management, Flight Plan 97, and HPMA.  Yet, the contribution of the human element to RCAF flight safety occurrences has not appreciably changed in modern times; there is no one panacea.  Therefore, it remains very difficult to make significant one-time reductions to the human contribution to occurrences, necessitating a modest incremental approach that targets specifics within the spectrum of human performance.  Compliance with orders, regulations and standard operating procedures (SOPs) is one such incremental step that has a very broad-reaching impact.

2.5.14  Orders, regulations, and SOPs govern how personnel either conduct or support air operations.  For example, certain orders deal with training, qualifications, and currencies; these are enforced by periodic assessments of documentation by standards personnel.  Others deal with the performance of duties; these are enforced by periodic verifications like an annual proficiency check ride, a category upgrade check, or quality assurance programs.  Additionally, the materiel state of equipment is also validated against rigorous criteria and inspection cycles; aircraft fuel samples are taken daily, for example.  However, when it comes to assessing the lynchpin integrator of all the elements of any air operation, the individual, his or her physical and mental health may be assessed by a medical authority as infrequently as one to five years, depending on their trade.

2.5.15  Although aircraft and weather, amongst a myriad of other parameters, are checked daily prior to flying operations, there is no rigorous daily check of an individual’s human performance before a wrench is turned or an engine is started.  For example, flying supervisors are to ensure that aircrews “…are qualified and competent to accomplish the assigned duty…;” [16] however, in practice, flight authorization is often granted by virtue of a signed daily flying program that removes routine authorizer/crew interaction.

2.5.16  Given the materiel advances in modern RCAF equipment and given that the individual is so fundamental to training for, planning, supporting, and executing air operations, the collective weight of effort should focus on optimizing human performance, thereby resulting in operational efficiencies.  Although P1 contravened a well-known existing order by self-medicating before conducting flying operations, both evidence and research suggest that his behaviour is not isolated.

2.5.17  In addition to this accident, there have been several recent cases where a flight safety investigation identified that the consumption of substances hazardous to aviation was contrary to established orders, regulations, and, in some cases, laws:

a.  Griffon hard landing in Feb 11, FSOMS 146017;

b.  Griffon wire strike in Feb 12, FSOMS 151471;

c.  Griffon hard landing in Jul 12, FSOMS 153131;

d.  Sea King shipboard landing in Jan 14, FSOMS 159212; and

e.  Harvard near mid-air collision in Apr 14, FSOMS 160189.

2.5.18  From February to April 2009, a pan-CF blind drug testing (BDT) study was conducted across 22 CF bases by the Director General Military Personnel Research and Analysis.[17]  The study demonstrated that amongst the 1,442 samples collected, the prevalence of recent illicit drug use by full-time CF members was 4.2% (4.8% males and 1.1 % females).  The rate was 6.5% for junior NCMs (i.e. the thousands of technicians who maintain our aircraft) compared to 1.4% for officers.  The rate in the navy was higher (4.9%) compared to the air force’s (3.6 %) but the report stated that this was not statistically significant.  Recent data indicated that there are 3540 aircrew, 5763 maintenance, and 1124 aerospace controllers and firefighters within the RCAF.[18]  Therefore, applying BDT data results, 438 RCAF personnel with occupation classifications that are directly involved with air operations would test positive for illicit drugs only and likely be cognitively impaired on any given day.

2.5.19  The pan-CF BDT also identified that the most commonly found drug was marijuana.  This study did not test for anabolic steroids, alcohol, prescription medications or over-the-counter non-prescription medications, all of which can negatively affect human performance.  Therefore, knowing that the illicit drug user rate of 4.2% only reflected those that tested positive on a given date and that not all illicit drug users would have been identified, and knowing that the testing did not look for many other common drugs that also impair human performance, the investigation felt it reasonable to conclude that the true user rate of substances hazardous to aviation over any period of time is higher than 4.2% or, in absolute terms, higher than 438 RCAF personnel.

2.5.20  Similarly, in 2008 a health and lifestyle information survey (HLIS) of CF was conducted by Director General Military Personnel Research and Analysis.[19]  The HLIS determined that in the preceding 12 months, 20% of the 2,175 respondents reported consuming hazardous amounts of alcohol, 59% received prescription and non-prescription medications from a civilian pharmacist, 14% reported using performance enhancers, and 3% reported using marijuana; these results support the investigation’s conclusion above, that the true user rate of substances hazardous to aviation over any period of time is higher than 4.2%  In the RCAF context, these drugs may affect HPMA if not taken under appropriate CF medical supervision.

2.5.21  The cumulative result of the 2009 CF BDT and the 2008 CF HLIS indicates that more than just 4.2% of CF members who conduct and support air operations are influenced on any given day by substances hazardous to aviation.  The toxicological findings from this accident and other recent ones are the visible indicators of the use of these hazardous substances; however, they appear only after an accident has occurred.  The occurrence data above strongly indicates their use continues today in spite of the CF Drug Control Program, orders, education and awareness programs, and access to health care.  The Chief of Defence Staff has not designated RCAF personnel conducting or supporting air operations to be safety sensitive [20] and, therefore, safety-sensitive drug testing, or any other random drug testing for that matter, is not routinely carried out within the RCAF.

2.5.22  Recognizing the incidence of drug usage in civil aviation, the United States Department of Transportation implemented a randomized drug and alcohol screening program in 1994.  The study demonstrated that positive tests decreased from 1.76% to 0.82% over 14 years.[21]  In 2008, based on 97,277 tests, positive results also decreased from 2.15% on pre-employment testing to 0.82% once people were subjected to a randomized drug screening program.  Also in 2008, post-accident testing showed a drug usage rate of 1.24%, or 50% higher than the 0.82% baseline rate for randomized testing; this indicates that drug use of any kind appears to be a risk factor for accidents.  With respect to what randomized drug testing can detect within the CF context, mass spectrometry of urine can easily determine the presence of illicit drugs; over-the-counter medications, such as Diphenydramine; alcohol, banned substances, and any other substance deemed by the Aerospace Medical Authority to be hazardous to aviation.

2.5.23  In 2007, a discussion paper on illicit drug testing in the CF reviewed the literature available and provided an opinion on drug testing.[22]  The evidence on the effectiveness of drug testing was not definitive; however, the paper did not specifically consider the influence of human factors in aviation.  The interesting point that the investigation drew from this paper was that, apart from the American Department of Transport testing, there appears to be little scientific literature existing that directly correlates drug use to accident occurrence rates.  Most legal and scientific advice approach accident prevention in an outdated manner by looking for a direct cause and effect rather than see the importance of managing pre-conditions, such as personnel cognitive fitness.  Accidents occur because of the existence of several seemingly innocuous pre-conditions that, if they occurred in isolation, would not in themselves have caused the accident.  The removal of just one pre-condition, or even just making the hole in Dr James Reason’s Swiss Cheese model of accident causation [23] a bit smaller, can prevent the same accident from happening in the future.  Therefore, this necessitates focusing efforts to reduce the risk to safety that the consumption of substances hazardous to aviation poses.

2.5.24  The advantages of an active random drug testing program include:

a.  Prevention of injury to the individual as a result of cognitive impairment;

b.  Protection of the individual’s co-workers;

c.  Preservation of materiel and resources;

d.  Protection of the public; and

e.  Preservation of the individual’s health through effective diagnosis and treatment.

2.5.25  In summary, aviation and RCAF air operations are complex.  As technology advances, both the aircraft and the tools to maintain them become increasingly reliable.  However, the same cannot be said for humans, whose fallibility must be attacked from every possible approach as the human factor remains the largest single cause of aircraft accidents and incidents.  Combating degradations of cognitive performance is one such aspect that has not significantly advanced despite modern programs to control, educate, and regulate.  It is a fact that CF operations, including air, are being conducted by people with degraded human performance caused by the unauthorized use of substances hazardous to aviation.  These substances not only include illicit drugs, but also alcohol, anabolic steroids, and prescription and non-prescription medications.  Even subtle effects of these substances can be amplified by fatigue, mild hypoxia, mental and physical stress, and the environment, all of which are routinely present during RCAF air operations whether in the air or on the ground.

2.5.26  Human behaviour and human performance are not as easily quantifiable as other sciences.  The direct link between the consumption of substances hazardous to aviation and aviation occurrences is not well researched, though it is clear that it exists.  It is similarly clear to the investigation that randomized testing for hazardous substances mitigates this known risk because it modifies an individual’s behaviour and minimizes the risk of injury to the individual, his co-workers, and the public; preserves resources for future use; and improves the probability of mission success.

2.6 Aviation Life Support Equipment

Ejection

2.6.1  P1 reported a large triangular void in his parachute harness and experienced riser slap during the ejection.

2.6.2  Both P1 and P2 suspended weights were within the published weight envelope.  The vertical descent rate of P1 and P2 under the parachute was calculated to be 26 FPS and the combined velocity was calculated to be 32 FPS in still air.  Descent rates in excess of the Military Standard have been documented in all previous Hawk ejections with preventative measures recommending a replacement of the GQ1000 parachute.

2.6.3  P1/P2 landed in a relatively soft swampy area, which significantly reduced their impact forces with the ground.  Most noteworthy was that because PI was unable to release his PSP, his higher descent rate could have resulted in his sustaining lower limb injuries had he landed on firm ground.

2.6.4  The investigation determined that a mitigation plan is in effect to install an automatically-deployed PSP along with changing the 2nd generation SCH to a 5th generation SCH.  Both of these measures should mitigate the likelihood of aircrew injury during ejection due to riser slap, reduce or eliminate the large triangular void in the shoulder harness, and reduce ground impact on landing.  The risk of injury due to ejecting and high rates of descent for a parachute landing will most likely be reduced to an acceptable level of safety.

Personal Survival Pack Deployment

2.6.5  Both P1 and P2 had difficulty releasing their respective PSPs.  P2 deployed his PSP by locating one of the side connectors by feel with the hand and then squeezing two buttons.  P1 made several attempts to find his connectors but was unsuccessful in attempts to release the PSP.  The side connectors can be difficult to locate, particularly following a highly disorienting ejection, making it less likely that aircrew would be able to deploy the PSP.

2.6.6  The investigation determined that had the PSP deployment mechanism been a simple-to-find pull-to-deploy handle, or an automatic deployment mechanism, both P1/P2 would have had more time to prepare for the parachute landing.

2.6.7  Implementation of an automatically-deployed seat pack capability is scheduled to begin in 1 Sep 2014 and should be fully implemented by 1 Oct 2015 on all Hawk aircraft.

Survivor Locator Beacon - Antenna Deployment and Modification

2.6.8  BMAT has been delegated design authority for the aircraft, including the ejection seat.  The lack of an auto-activated emergency radio was a concern to the operational community and BMAT was directed to incorporate an emergency radio auto-activated upon seat-man separation.

2.6.9  While the Directorate Aerospace Engineering Project Management (DAEPM FT) 6 is available to provide technical input on ALSE issues, the coordination activity and technical airworthiness approval rests with the BMAT Senior Design Engineer (SDE).  DAEPM (FT) 6 manages a limited number of items of ALSE used on the CT155 fleet (helmets, LPSV, and G Suits).  All other ALSE items are managed by BMAT.  As such, the BMAT SDE identifies, coordinates, tracks and resolves all Hawk fleet ALSE design changes.  If design changes impact DAEPM (FT) 6-managed equipment, then DAEPM (FT) 6 is included as a stakeholder.

2.6.10  BMAT conducted all design work, testing and release to service for the SLB.  Early implementations had the antenna retained inside the seat pack; however, following an accident on 14 May 2004, it was found that the emergency radio signal was not received by ATC on 121.5 or 243.0 MHz.  In either case, ATC should not have been expected to receive the low power 121.5 MHz and 243 MHz signals due to low signal strength.  It was believed that while the radio was transmitting with the antenna retained inside the seat pack, radio signal propagation may not have been adequate; however the bulk of the power again was being reserved to transmit the 406 signal.

2.6.11  To ensure that sufficient antenna was exposed, the antenna length increased to 43 inches when unfurled.  It is now thought that even had the external antenna been exposed correctly, ATC would still not have heard a 121.5 MHz or 243 MHz signal due to their low signal strength on these frequencies.

2.6.12  As an added safety measure to the SLB in the seat pack, there is a personal locator beacon (PLB) carried by front seat aircrew in the left pocket of the LPSV to aid search and rescue efforts.  The PLB will be carried until the Hawk PSP is modified with an ADU.

2.6.13  Upon ejecting, P2 was whipped by the longer antenna flailing in the airstream.  Though no injury occurred, P2’s outer visor was cracked and layers of his G-suit fabric were cut by the flailing antenna.  P1 did not suffer any injuries from the flailing antenna.

2.6.14  A modification to the deployment of the 43 inch cable antenna was completed following this accident, whereby the deployment of the antenna would deploy only upon seat pack deployment.  This configuration change ensures that the antenna deployment happens later in the ejection sequence and that the antenna deploys further from the pilot, thereby reducing the possibility that it could strike him.

2.7 Other ALSE Equipment

2.7.1  During the investigation the following ALSE discrepancies were noted:

a.  P1/P2 PSPs contained Tylenol© with an out-of-date expiry of 2009;

b.  Both LPSVs contained the MS2000 strobe light, which was to have been replaced with the Firefly 3 strobe light IAW CF message DCOMD FG 055 211544Z Aug 08; and

c.  One Beaufort MK30 LPSV sustained damage from MDC shrapnel, which penetrated the bladder cover and the orally inflated chamber reducing aircrew floatation time to 70 minutes.

Top of page

3.  CONCLUSIONS

3.1 Findings

3.1.1  The Adour engine LPT blade has a known problem of fatigue cracking at the trailing edge rear acute corner. (1.6.4, 1.6.11, 1.18.1, 2.2.1)

3.1.2  LPT fatigue cracking has resulted in five LPT blade failures prior to the LPT blade reaching its original design life of 2,000 hours. (1.6.13, 2.2.2)

3.1.3  Adour engine component modifications have been implemented to address LPT blade cracking, but have been unsuccessful in resolving the problem, resulting in reduction of design life of the LPT blade to 500 hours. (1.6.3, 1.6.5, 2.2.3)

3.1.4  The Hawk AOI’s and Hawk Pilot checklist provide differing definitions to Land as Soon as Possible and Land as Soon as Practical.  (1.1.7, 2.4.1)

3.1.5  P1/P2 and the student pilot were qualified and current for the mission flown. (1.5)

3.1.6  No documentation existed in P1’s health records that showed he was either ungrounded or able to resume flying duties or that he was able to self-unground. (1.13.3, 2.5.3)

3.1.7  P1 conducted the mission without being cleared to fly from a CF medical officer while suffering from a physical illness, contrary to B-GA-100-001/AA-000, Chapter 9, paragraph 1. (1.13.3, 2.5.3)

3.1.8  P1’s toxicology was positive for Diphenhydramine. (1.13.2, 2.5.5)

3.1.9  Diphenhydramine is not authorized for use by aircrew because it causes sedation, decreased concentration, resulting in cognitive impairment with the faculties affecting attention, memory, vigilance and perception of speed even if one is not symptomatic of sedation (1.13.2, 2.5.7)

3.1.10  P1 conducted the mission under the influence of Diphenhydramine without the prior approval of a flight surgeon, contrary to B-GA-100-001/AA-000, Chapter 9, paragraph 2. (1.13.2, 2.5.5, 2.5.6)

3.1.11  The investigation could not conclusively determine the extent to which P1’s performance during the accident flight was affected by the unauthorized use of Diphenhydramine. (1.13, 2.5.9)

3.1.12  The HPMA breakdowns during the flight (the decision to shut down the engine while 4000’ below the FL profile, the uncoordinated and unsuccessful attempts to relight the engine, the inaccurate FL profile determination, the carrying out of non-checklist pre-ejection procedures, the non-completion of the ejection procedure, and ejecting below ejection altitude criteria) were nonetheless consistent with the actions of an individual under the influence of Diphenhydramine. (1.13.3, 2.5.4 - 2.5.6)

3.1.13  Mission planning, start, taxi, take-off, departure, and mission warm-up were uneventful. (1.1)

3.1.14  During the first BFM engagement, Hawk 155201 engine experienced a single LPT blade liberation due to material failure, causing minor damage to other LPT blades. (1.1.6, 2.3.1, 2.3.3)

3.1.15  After the initial indications of an engine problem, P1 turned the aircraft towards CYOD, climbed gently and initiated emergency response procedures. (1.1.7, 1.11.1, 2.3.14)

3.1.16  P1/P2 used a visual distance estimation technique rather than using DME or radar distance to more accurately calculate their FL glide profile to CYOD. (1.1.8, 2.4.24)

3.1.17  An LPT blade mass imbalance induced a vibration to the throttle linkage feedback that then caused the engine RPM to increase from 85% to 93%.  Increased engine RPM led to the increased engine vibrations that P1/P2 experienced. (1.1.9, 2.3.10, 2.3.11)

3.1.18  P1/P2 did not detect the engine RPM increase to 93%; due to perceived worsening engine conditions, they elected to shut down the engine below the FL glide profile. (1.1.9. 1.1.11, 1.1.12, 2.3.11, 2.3.13)

3.1.19  P1/P2 attempted two engine relight procedures, however, checklist errors resulted that prevented a successful relight. (1.1.11, 2.4.9, 2.4.16, 2.4.23)

3.1.20  P1/P2 did not complete the Hawk controlled ejection checklist procedure despite having sufficient time available to do so. (1.1.12, 2.4.35)

3.1.21  Prior to ejecting, PI requested P2 to remove from the cockpit video status and control unit the VCR tape that depicted the audio and HUD information.  P2 then secured the tape in a Velcro-enclosed lower-leg pocket of his g-suit.  This was not IAW the Hawk controlled ejection checklist. (2.4.37, 2.4.37.1)

3.1.22  P1/P2 ejected between 550’ and 750’ AGL, well below the minimum recommended altitude of 2,000’ AGL, due to distraction caused by stowing non-checklist personal items such as a wrist watch and the HUD tape. (1.1.12, 2.4.39)

3.1.23  P1 suffered minor injuries during the ejection due to the canopy-fracturing MDC, riser slap, a triangular void in the shoulder harness, and improper leg position against the seat front just prior to ejection. (1.1.14, 1.6.21, 1.13.1, 2.6)

3.1.24  P2 suffered minor injuries during the ejection due to improper leg position against the seat front just prior to ejection; his outer visor was cracked and layers of his G-suit fabric were cut by a flailing 43” SLB antenna. (1.15.2, 2.6.13)

3.1.25  Post-ejection, P1 was unable to find or release his PSP. (1.1.15, 2.6.5)

3.1.26  A YouTube video recorded P2 touching his g-suit pocket to confirm to P1 that he had the VCR tape.  The tape subsequently disappeared, likely while the ALSE equipment was stowed in the back of a parked support vehicle near the crash site. (2.4.37.2)

3.1.27  The disappearance of the VCR tape impeded the investigation’s analysis of the crew’s response to the emergency. (2.4.37.2)

3.1.28  There is no random drug testing program in place for the RCAF. (2.5.21)

3.1.29  Post-accident drug testing is not an effective means to deter drug use. (2.5.21, 2.5.22)

3.1.30  Illicit drug use is present in at least 4.2% of the CF population.  When factoring the use of common drugs, such as alcohol, steroids, over-the-counter medications, and medications not used under the supervision of CF medical authorities, the percentage of the CF population potentially cognitively impaired at any given time is higher than 4.2% or, relative to the RCAF, more than 438 personnel. (2.5.18, 2.5.19, 2.5.20)

3.1.31  Random drug testing has been shown to effectively lower illicit drug use to 0.82% of a given population. (2.5.22)

3.2 Cause Factors

Unsafe Acts

3.2.1  P1/P2 did not detect the engine RPM increase to 93%; due to perceived worsening engine conditions, they elected to shut down the engine below the FL glide profile and without reference to available engine performance instruments. (1.1.9. 1.1.11, 2.3.11, 2.3.13.2)

3.2.2  P1/P2 attempted two engine relight procedures, however, due to a breakdown in internal cockpit communications (HPMA), checklist errors resulted that prevented its successful relight. (1.1.11, 2.4.9, 2.4.24)

Latent Cause Factors

3.2.3  During the first BFM engagement Hawk 155201 engine experienced a single LPT blade liberation due to material failure, causing minor damage to other LPT blades. (1.1.6, 2.3)

3.3 Other Findings

3.3.1  The RARM for the GQ1000 main parachute will likely be mitigated to an acceptable level of safety with the implementation of an automatically-deployed PSP and the 5th generation SCH. (1.15.3, 1.15.4, 2.6.4, 2.6.7)

3.3.2  The signal transmitted from the SLBs on 406 MHz was sufficiently strong, but a one hour and 12 minute delay in receipt of the transmission was caused by a gap in LEOSAR satellite coverage. (1.15.10 - 1.15.14, 2.6.10, 2.6.11)

3.3.3  A PLB is carried by front seat aircrew in the left pocket of the LPSV as an added security measure to aid in search and rescue of aircrew. (1.15.18, 2.6.12)

3.3.4  The Hawk AOI’s and the Hawk Pilot Checklist currently provide a checklist response for an Engine Malfunction or an Engine Mechanical Damage, including direction to land as soon as possible or practical, depending on the nature of the engine problem. (1.1.7, 2.4.3)

3.3.5  Both PSPs contained Tylenol© with an out-of-date expiry of 2009. (1.15.7, 2.7.1)

3.3.6  Both LPSVs contained the MS2000 strobe light, which was to have been replaced with the Firefly 3 strobe light IAW CF message DCOMD FG 055 211544Z Aug 08. (1.15.7, 2.7.1)

3.3.7  One Beaufort MK30 LPSV sustained damage from MDC shrapnel, which penetrated the bladder cover and the orally-inflated chamber, rendering it unserviceable. (1.15.5, 2.7.1)

Top of page

4.  PREVENTIVE MEASURES

4.1 Preventive Measures Taken

4.1.1  LP Turbine Blade Remedial Actions

4.1.1.1  To reduce the stress in the LPT blade rear root neck, corner below the platform, the current standard blade AX72453 has been re-designed.  The re-designed LP turbine blade AX73185 introduces a buttress feature to the rear shank/retaining plate groove area and also negates the necessity for hand dressing.

4.1.1.2  Following this accident, the blade life limit was set to 400 hours based on statistical analysis.  Sampling of a number of low and high time engines allowed the legacy LPT blade life limit to be raised to 500 hours.  Two sets of blades have been sampled at 500hrs with no anomalies found.

4.1.1.3  CT-155 Hawk Adour engines have been upgraded with newly designed LPT blades.  The LPT blade upgrade program was completed 31 Mar 2013.

4.1.1.4  A sampling program of new LPT blades was initiated to assess LPT blade every 500 hours up until 2,000 hours.  It is expected that the sampling program and LPT blade certification to 2,000 hours will be completed by 1 Mar 2016.

4.1.1.5  The Engine Malfunction checklist was amended and the Engine Mechanical Damage checklist was added, providing direction specifying Land as Soon as Possible or Land as Soon as Practical, depending on the nature of the malfunction.

4.1.1.6  The SLBs are transmitting as per the OEM specifications with 406 MHz being the primary signal for search and rescue.  Front seat Hawk pilots carry a PLB as added safety measure until the Hawk PSP is modified with an ADU.

4.1.1.7  BMAT developed and implemented an improved process for tracking expired PSP contents.  BMAT also produced Work Instruction AFM WIN 5407 Hawk Personal Survival Pack Repacking and Replacement of Contents released 31 Jan 2012.

4.2 Preventive Measures Recommended

4.2.1  2 Cdn Air Div AF Trg:  Amend the MFT chapter 7, paragraph 7.06, to add a warning about visually assessing range to determine PFL or FL glide profiles.  This method should be identified as the least desirable to be used.

4.2.2  2 Cdn Air Div AF Training:  Amend the Hawk AOIs IAW the Hawk Pilot Checklist to standardize the definitions of Land as Soon as Possible and Land as Soon as Practical between both documents.

4.2.3  2 Cdn Air Div AF Trg:  Review this accident in order to develop HPMA practices when two pilots crew an aircraft that is typically flown single-pilot.

4.2.4  1 Cdn Air Div A3 Fighter:  Update 1 Cdn Air Div Order 3-304 to provide direction to aircrews to not remove or attempt to remove aircraft recording device memory storage units when egressing aircraft in emergency situations.  Attempts to do so may either compromise aircrew safety or result in the loss of these memory storage devices and impede investigation.

4.2.5  Chief of Military Personnel (CMP):  Implement a randomized substance testing program to detect and deter people from consuming substances hazardous to aviation safety.

4.2.6  CO CFEME:  Implement a recurrent training program for aircrew health providers to ensure they have the knowledge required to deliver safe and effective aeromedical support.

4.2.7  1 CAD Flight Surgeon:  Implement a quality assurance review of aircrew health records that gives feed back to the health provider about delivering safe and effective aeromedical support.

4.2.8  CMP Director of Medical Policy:  Develop a system to transmit Medical Employment Limitations (including groundings) directly to the unit on a regular basis.

4.3 Other Safety Measures Recommended

4.3.1  2 Cdn Air Div AF Trg:  Expedite implementation of the 5th generation SCH and the PSP automatic deployment capability for the Hawk fleet.

4.3.2  2 Cdn Air Div AF Trg:  Ensure ATC personnel and CF SAR aircrew are aware of the Hawk’s 121.5 and 243.0 MHz weak transmission signal capability and that the primary use of 406 MHz homing is preferred due to greater signal strength and detection range.

4.3.3  BMAT:  Ensure all HAWK CT155LPSVs are equipped with the Firefly 3 strobe light IAW CF message DCOMD FG 055 211544Z Aug 08.

4.3.4  DAEPM (FT) 6:  Improve the Beaufort MK30 LPSV bladder stowing material with a more durable material to prevent LP damage from MDC shrapnel.

4.4 DFS Remarks

Military flying is a dynamic, challenging environment, in which all aircrew are called upon to exercise sound judgement, excellent flying skills, and at times, quick decision-making during highly stressful situations.  In order to complete the mission safely and effectively, aircrew must be mentally and physically prepared before they climb into the aircraft.

Giving us the structure to be both mentally and physically prepared are our orders, regulations, and SOPs, some of which prohibit the use of over the counter medications and illegal substances as they adversely affect human physiological performance.  The science is clear with respect to the implications of degraded human performance in aviation.  Our regulations are written in blood, so the message here is simple:  follow the rules.

Another issue I would like to address is the application of HPMA principles within our flight operations.  Whether maintaining or flying our aircraft, including drones, effective and timely HPMA is critical to safe and effective mission accomplishment.  Ineffective crew coordination enables and can accelerate the accident chain, as was seen on 10 Jan 1992 when two experienced pilots of a CF-5 were killed while attempting to verify a pod had jettisoned safely from their aircraft while flying at 500’ AGL.  How could this happen?  Neither pilot communicated effectively with the other to confirm who was responsible for flying and avoiding the ground and who was responsible for verifying the pod’s safe jettison.  Instead, both Pilots focussed their attention on the pod and consequently flew into the ground.

In this Hawk accident it is apparent that an HPMA breakdown occurred between both pilots during their handling of the emergency.  Although we cannot irrefutably state that the aircraft could have safely recovered at CYOD, it is apparent that effective HPMA between the two pilots would have improved their chances for a safe recovery.

S. Charpentier
Colonel
Director of Flight Safety


Top of page

Annex A:  Engine data vs running time

Annex A, Figure 1:  DAU Engine Data.  The full flight is presented until RPM drops below 15%.  Time is engine running time from start up.  Note that from the LPT blade loss event until the engine was secured was 3 minutes 46 seconds.

Annex A, Figure 2:  DAU Engine Data.  Data is presented from just prior to the LPT blade loss event until RPM drops below 15%.  Time is engine running time from start up.  Note that the throttle was initially retarded to approximately 82% RPM and then advanced in two distinct steps to about 86% RPM.  Also note that the igniters were activated (solid blue region) at local time 11:39:46, 27 seconds after engine shut down.

Annex A, Figure 3:  DAU Engine Data.  Note that after the pilot set the throttle, RPM continued to slowly increase for approximately 2 minutes and 40 seconds from about 86% RPM to about 89% RPM and then increased more rapidly for the next 35 seconds until the engine was shut down at about 93% RPM.  Additionally, a close up of the RPM data during this period revealed noise in the signal (+/- 0.1%) when compared with RPM data prior to the LPT blade loss event.  Included is a plot of the expected NL (NL Model).  The NL Model was based on engine performance data from this flight prior to the LPT blade failure event.  Note that after 11:36:04, the recorded NL does not follow the expected NL profile and is extremely noisy but appears to return to normal behaviour after the engine is shut down (ie: exponential decay).

Annex A, Figure 4:  DAU Engine Data.  Note that the point where Zulu 22 reports Zulu 21 is trailing smoke begins when the slope of RPM data increases.  Also note that the report of smoke intensifying occurs just prior to the engine being secured.

Annex A, Figure 5:  DAU Engine Data.  Post-engine shut down, engine parameters are seen to follow an exponential decay profile.  Engine ignition was activated when RPM reached 21% RPM (27 seconds after the engine is shut down).  The DAU does not record throttle position so it was not possible to directly ascertain from this data that a relight was attempted.  Recording ceased 20 seconds later as RPM dropped below 15%.

Annex A, Figure 6:  DAU Engine Data.  The RPM data was curve-fit and extended to a point beyond where the failed relight attempt occurred.  When Zulu 22 radioed that the smoke was returning (1 minute 35 seconds after the engine ignition is activated), RPM was estimated to be about 8%.  When P1 told Zulu 22 to watch for smoke (1 minute 6 seconds after the smoke returning call), RPM was estimated to be about 5%.  Although these values are only estimated, it is clear that at 11:42:27 LT the engine RPM was well below the minimum required to perform a cold relight (13%).


Top of page

Annex B:  Flight profile data

Annex B, Figure 7:  ACMI Data.  From the ACMI data, the actual glide profile of Zulu 21 was determined.  The flight parameters were fairly close to the prescribed engine out glide parameters (185 KIAS and 5.5 units AOA).  Over the period that the ACMI recorded data, the aircraft achieved a 2.12 NM/1,000’ glide ratio.  This is very close to the AOI-prescribed value of 2 NM/1,000’.

Annex B, Figure 8:  ACMI Data.  The magenta line is the calculated Zulu 21 glide profile assuming the crew of Zulu 21 maintained the prescribed engine out glide (185 KIAS and 5.5 units AOA) between the position and altitude when the ACMI stopped recording and the position and altitude that the aircraft impacted the ground.  Ground impact data indicates that the aircraft struck the ground in a near level pitch attitude, left wing down.  Since the engine out flight path angle is 4.7 degree down, at 5.5 units AOA the aircraft would have impacted the ground with of about 1° nose up attitude, essentially level pitch.  From this data, the engine out glide ratio was found to be 2.11 NM/1,000’, in essence a continuation of the glide ratio from the ACMI data.  Plotted alongside Zulu 21 data, Zulu 22 data identifies the relative positioning of the two aircraft.  Based on the known ejection time, it was calculated that the ejection occurred at 750’ AGL.  Allowing the glide ratio to vary from 2.01 NM/1,000’ to 2.21 NM/1,000’, the ejection altitude was found to vary +/- 250’ AGL so the ejection altitude was most likely between 500 and 1,000’ AGL.


Top of page

Annex C:  Cold Lake and Lloydminster aerodrome data

Annex C, Figure 1:  CYOD Airport Information

Annex C, Figure 2:  CYLL Airport Information


Top of page 

Annex D:  Additional analysis – Ardour Engine and LPT blade

Examination of Blade #62

Binocular examination of the fractured blade remnant (blade #62), after it had undergone a light cleaning, revealed that the crack had propagated in mechanical fatigue, which had initiated at approximately 1.0 millimetre (mm) along the rear face from the rear acute corner.  The fracture had occurred approximately 18.5mm from the base of the blade.  The fatigue system developed 8.0mm along the rear face and 13.0mm along the length of the long face on the concave side of the blade.  The remaining area of the fracture surface displayed large crystallographic facets, evident of fast fracture.  Scanning electron microscope (SEM) examination revealed two pores, measuring ~110 microns in length and ~20 microns respectively (acceptable to material specification), adjacent to the origin.  The root had suffered damage at the origin itself, which may have obscured another feature.  The wear patterns on the loaded flanks of the root serrations were consistent with patterns noted on the remaining blades in the engine set.

A fractographic analysis of the fracture surface of blade #62 revealed multiple levels of banding, working from coarse to medium, fine and extra fine as increasing levels of magnification were used.  Although the presence of highly irregular banding is indicative of high cycle fatigue (HCF) modes, it was thought likely that the coarse (macro) bands correlated with engine cycles (low cycle fatigue (LCF)) or sub-LCFs.  The macro banding was at its finest on approach to the crack tip, when HCF was believed to be the dominant propagation mode.  An alternate straw – dark straw/purple colouration was noted and was most likely indicative of differing propagation rates (potentially two or more resonant modes).

To understand the rate of crack propagation, an attempt was made to quantify the fatigue banding and identify groups of banding compared to flight profiles.  After extensive analysis of the fracture surface and the associated fatigue banding, it was not possible to correlate the banding with any part of the respective flight profiles.

Inspection of the rear face of the root fragment in the SEM found numerous small, secondary cracks to be present immediately below the origin with cracks running parallel to fracture surface.  Further from the origin, numerous ‘micro cracks’ were also found to be present, often associated with features on the surface.  These small cracks or fissures were subsequently identified as fine cracking within a surface layer.

After polishing into the origin of the primary fracture of blade #62, the presence of the features noted above were confirmed.  It should be noted that damage to the origin had resulted in loss of surface layer, so no measurement could be obtained.  However, the surface layer was measured as 14 micron, just below the origin on the rear face.  A section taken through the centre of the shank gave a surface layer thickness of 9 micron.

Remaining Blades

The fluorescent penetrant inspection (FPI) indication on the rear acute corner of blade #10 (S/N 3548A 18) was opened and confirmed as a crack measuring 2.3mm (rear face) by 1.8mm (long face) with a crack depth of 1.4mm.  The crack was noted to have a pore (~60 microns in length) located adjacent to the origin.  A further eight blades containing FPI indications were ‘opened’ and confirmed as containing fatigue cracks.  All of these blades were noted to have multiple fatigue systems, with multiple initiation points, present to a depth of no greater than 250 microns.  It should be noted that porosity, present as a result of the casting process, had served to locate the origin of a number of the fatigue systems observed; however, there was no direct correlation apparent between pore size and crack size, although it should be noted that blade #62 had by far the largest pore observed.

The fatigue system present on all cracked blades was highly complex, with a large number of propagation ‘events’ present across the full extent of the fracture surface, and most likely indicative of complex operating profiles.  The cause of such complex banding could include LCF cycles, sub-LCF cycles, transient passages through resonant conditions, throttle movements and minor perturbations at peak resonant conditions.

The presence of such complexity in the fracture surface, especially the presence of numerous coarse bands, indicates that crack growth was unlikely to have occurred over the duration of a single flight.  Attempts made at a detailed band count did not result in any obvious correlation with engine usage.

A significant number of cracked blades were opened from a number of engine sets across a number of operators.  At no stage did any opened crack extend past the coarse banded (likely underlying LCF) region of crack growth, further supporting the theory that the final portion of crack growth was HCF dominated and likely occurred over only a few flights.

It should also be noted that even if a successful band count and comparison had been obtained, it would likely only apply to the conditions of the particular blade in question.  Comparison of the initial 1.5mm of fatigue crack growth has been compared for three blades from two engine sets, and no consistency in banded features was observed.

Retaining Plate Groove

Visual examination of the blade retaining plate groove identified witness marks caused by loading of the retaining plates as a result of centrifugal force.  Four distinct witness patterns on the LPT blades from engine 7825 were observed:

d.  Centrally loaded (60 – 80% of the groove length);

e.  Tending to convex side of the groove;

f.  Tending to concave side of the groove; and

g.  Centrally loaded (80 – 100%of the groove length).

In some instances, where the loading extended to the concave side of the retaining plate groove, two distinct contact witness marks were evident, which implied that the blade had experienced different loading conditions against the retaining plate throughout its service life.

Surface Layer

The presence of a surface layer was noted on the external surface of the blade root adjacent to the fracture, which was evident in both the fatigue and the overloaded areas.  The surface layers in both areas were similar in appearance, implying that overload was the likely cause of the fracture of the surface layer, with the layer associated with the fatigue being more oxidised in appearance.  The fatigue appeared to originate from the surface layer interface with the substrate.

A transverse section was taken through the root of crack-free blade #29 (S/N 3739A7) at a height consistent with the known crack location, to determine the extent of the surface layer present around the shank of the root.  The sample underwent an electrolytic etch and the presence of a surface layer was confirmed.

A transverse section was taken just below the crack present in blade #14 (S/N 3699 A25), with an additional longitudinal section taken to enable the origin of the crack to be polished through.  The transverse section confirmed the findings from blade #29, with respect to surface layer.  The longitudinal section taken through the origin confirmed the presence of the surface layer adjacent to the fatigue crack, which measured approximately 15 microns thick.  Also of note was the presence of a number of oxidised fissures through the surface layer.  The sample was then lightly re-polished to enable compositional analysis to be performed.  A line scan was taken across the surface layer, which indicated that the surface layer was depleted in chromium and titanium when compared to the substrate, and revealed it to be rich in aluminum.

A longitudinal section taken through blade #42 also revealed that fissures were present in the surface layer.  When viewed under a higher magnification, a number of fissures were found to have started to propagate into the ‘base’ material of the alloy.  Carburisation was also found to a depth of up to 110 microns behind the surface layer with a further layer of additional carbides of approximately 60 microns in depth.

The surface layer was considered to have been formed due to chemistry exchange between the blade root and the masking compound used during the blade aluminising process.

Whole blade fatigue testing has been carried out to measure the impact that this surface layer has on the fatigue behaviour of the blade.  The testing showed that the presence of the surface layer on the shank corner of the blade reduced the fatigue life of the turbine blade by approximately a factor of two when the blade was excited in the first edgewise mode.  The engine does not operate within the speed range at which the first edgewise mode will occur and so it contributed in part to the cracking seen, but it was not considered to be the sole root cause.

Temperature Assessment

One blade was randomly chosen for temperature assessment.  A number of transverse sections were taken through the aerofoil of the blade.  The assessment identified only slight micro-structural modification, in the form of rafting.  This was in line with previous experience of in-service blades and indicates that no over-temperature event was experienced by this blade set.

Semi-quantitative analysis of the bulk material in all blades assessed gave compositions consistent with the drawing requirements.

Retaining Plates

The retaining plates are located in a groove in the LPT blade and an annular groove in the LPT disc.  The purpose of the retaining plates is to maintain the axial position of the blade in the disc.  There are forty-six retaining plates and one pre-bent lock plate in the set.  The width of the pre-bent lock plate is adjusted to achieve the correct cold build circumferential clearance and is flattened in position after the retaining plates have been fitted.

It should be noted that any circumferential location described against the retaining plates is only consistent with how the retaining plates were found.  The pre-bent lock plate was located three retaining plates away from the location of the fractured blade.

A number of retaining plates were noted to have pips present on the outer circumference contact face.  The pips were present on a number of the retaining plates and appeared to be formed at the end of the blade retaining plate groove positions as they are one blade width apart.  Retaining plate #1, associated with fractured blade #62, was the only retaining plate to exhibit any signs of deformation.  When the condition of the retaining plates was mapped, the pip features did not correlate strongly with blade crack location.

LPT Disk

The disc underwent an ultra high sensitivity penetrant inspection, with no crack indications observed.  No unusual rubs were noted to be present on the disc.  Contact witness with the retaining plates was found to be consistent across all disc posts, with only minor exceptions noted.  These exceptions were not associated with any cracked blade locations.

As the retaining ring had been machined off at the beginning of the investigation, it was possible to assess the contact patterns on faces of the disc retaining ring where it had come into contact with the inner diameter of the retaining plates.  The presence of a continuous fret mark around the circumference of the ring was noted; however, there were changes in the patterns observed indicated that the retaining plates had moved radially outwards at some stage during engine running.  These patterns were not found to correlate with location of cracked blades.

Root Cause Analysis

To fully understand the contributory factors into the cause of the LPT blade fracture, an Ishikawa or “Fish Bone” diagram was initially used to identify possible areas or influences that could detrimentally affect the LPT blade that would lead to its fracture.  Six main categories were identified; design, manufacture, configuration, assembly, quality, and operational / environmental.

In addition to the Ishikawa diagram, a fault tree was compiled.  Root Cause Analysis (RCA) is a technique that is used to evaluate all of the conceivable causes of a failure; in this case the fracture of the LPT blade.

The conceivable causes are collated from the failure modes and effects analysis (FMEA) and assessed by the relevant company experts considered to have a field of expertise that may help to understand the failure.  Once the conceivable causes have been identified these are then individually assessed by reference to the physical evidence, theoretical analysis and the judgement of the experts with the intention of eliminating as many of the causes as possible.  Once these have been eliminated the remaining cause is further tested to ensure that it remains as a potential cause.  In the case of more than one cause being identified, these can be evaluated against a balance of probabilities to decide on the most likely cause, from which the appropriate corrective actions can be implemented.  Should it not be possible to identify a single cause; corrective action for more than one cause must be considered.

The results of the RCA showed that there was no change evident during the review of the manufacturing history and process of the LPT blade that could have significantly affected the integrity of the blade.

Previous analyses have shown the peak stress within the LPT blade is located within the under-platform area at the rear pressure side shank fillet.  The nominal stress in this area of the blade with a conforming radius is 1610 MPa; the current design practice of the under platform features is not to exceed 1000 MPa.

The analyses also showed that the stresses in the fillet will vary with the radius formed.  A dimensional assessment of the fractured blade confirmed that the rear corner radius was within the required drawing limits.  The analysis undertaken as part of the RCA also showed that the stresses in the rear root neck corner of the blade can vary during engine running due to a number of factors.  The main factors that influenced the changes in stress were the retaining plate configuration, blade geometry, blade crystal orientation and the environmental conditions. 

Retaining Plate Load and Chocking

This was the first LPT blade fracture that had modification AO1692 embodied.  This modification replaced the triple span retaining plate manufactured in Nimonic 75 with a double span retaining plate manufactured in Nimonic 90.  Nimonic 90 has a significantly higher yield point than Nimonic 75; therefore to understand whether the stronger double span retaining plate influenced the fracture of the blade a review of the behaviour of the retaining plates during engine running was conducted.

Inspection of the bedding patterns on the LPT blades from the event engine and other blades fitted with Nimonic 90 retaining plates subsequently inspected showed that in the majority of the blades examined, the retaining plates were not loading across the full assumed contact area, as had been observed with the triple-span Nimonic 75 lock plates previously used.  Loading tended towards the centre of the blade retaining plate groove circumferentially.  Some axial variation was also noticed.

During the build of the LPT module a specified total gap is required between the retaining plates, which are achieved by altering the width of the lock plate.  This is to allow for the thermal growth of the retaining plates.  Based on the retaining plate proof stress, it was calculated that the deformation seen on the upper edge of the Nimonic 75 triple span retaining plates was greater than would be expected from centrifugal loading alone.  It was determined that as the retaining plates grow thermally the cold build gaps between them are consumed.  Constrained radially by the retaining plate groove in the blades, and circumferentially by the now-continuous ring of retaining plates, no further growth can take place.  This is known as “chocking.”  The Nimonic 90 retaining plate chocking could increase the nominal root shank acute corner stress by as much as 11% above the nominal root shank acute corner stress in blades fitted with Nimonic 75 retaining plates.  The reason for the increase in the nominal stress is that the Nimonic 90 retaining plates chock during engine running and increase the loading into the blade retaining plate groove as a result of their increased strength.  Nimonic 90 retaining plates do not plastically deform in the way that the Nimonic 75 plates do and so loading into the blade retaining plate groove will not be evenly distributed.

Having a lesser yield strength than the Nimonic 90 retaining plates, the triple span Nimonic 75 retaining plates would deform and bed in evenly under chocking loads.  Evidence suggests that the double-span retaining plate retained their shape and can load on a smaller proportion of the retaining plate groove in the blade.  Observed evidence also suggested that the retaining plates can load at a variety of positions along the retaining plate groove of the blade.

Modelling the blade by varying the load on the blade retaining plate groove showed that the peak stress in the blade retaining plate groove and in the rear pressure side shank fillet will increase with uneven retaining plate loading.  Blade modelling using observed evidence suggested that the retaining plates can load off-centre on the blade resulting in change in the nominal acute corner stress between -11% to +8%.

Dimensional Assessment

The single crystal LPT blade has been manufactured at two different facilities and has been machined at three different facilities.  Therefore, to understand whether there have been any changes to the blade geometry as a result of the changes in the different facilities, a number of blades were optically scanned.  The electronic image of each scanned blade was compared with the drawing requirements.

Over 70 blades were scanned, covering samples having been manufactured from the two casting facilities and the three machining facilities.  The samples also included blades with and without cracks in their root neck, and with pack and vapour aluminised aerofoils.

Some of the blade dimensions recorded during the optical scanning were compared with the drawing requirements; both the cast and machined dimensioned were compared.

The results of the analysis showed small differences in the geometry between blades cast at the Italian facility and blades cast at the United Kingdom facility; however, there was no significant difference between cracked and un-cracked blades.  There were also no significant geometric differences between the machining facilities.  The differences identified in the blade geometry were used to produce dimensional distributions, which formed inputs into the RCA validation and the steady state stress assessment.

Although conforming to the drawing requirements, the blade modelling identified small differences in the blade geometry that could cause an increase the stress in the rear root neck corner.  The differences in the blade geometry were the malformation in the retaining plate groove, which could occur as a result of a worn grinding wheel/cutting tool, the rear root corner radius and the shank offset.  The magnitude of the variation of stress that could be caused by the largest geometric variation found in corner radii, shank offset and a malformed blade retaining plate groove are; nominal stress +35%, +7% and +13% respectively.

Single Crystal Orientation

The effect of the blade crystal orientation was assessed and found that the range of stress variation using the extremes of crystal orientation was predicted nominal stress +10 to -33%; however, the predicted stress within the shank fillet radius of the fractured blade from engine 7825 (blade 3529A25) was 13.8% below the nominal stress.  Although the assessment found that the crystal orientation affects the nominal stress, it was also identified that the crystal orientation affected the strength of the blade and not necessarily in an adverse way. 

Therefore, an increase in the nominal stress due to the crystal orientation can also increase the strength of the blade and therefore mitigate the impact of the stress increase.

Environmental and Flight Conditions

An assessment was made because of the environmental conditions in which the NFTC Hawk aircraft operate, both on the ground and in the air.  The colder climatic conditions in which the aircraft operate affect the maximum LP spool speed; the LP spool speed of the Adour Mk 871 engine is limited by either the TGT or by the NL.  Engines running at full throttle in colder climatic conditions are typically maximum NL speed limited.  In colder climatic conditions the engine idle speed is typically lower.  It was, therefore, considered that the dynamic stresses in the LPT blades in engines operated by NFTC may be increased by running longer in known blade resonant modes that occur at the extremes of normal engine operating speeds.

Health Usage Monitoring System (HUMS) data recorded on all NFTC aircraft was used to analyse the engine usage and look for correlations between usage and blade cracking.  Over 30,000 flights were analysed and a comparison was made between engines that operated at the extremes of their operating range for longer, compared to engines with LPT blades which have known to have cracked.  A similar comparison was made between the types of mission sortie carried out, compared to engines with LPT blades which are known to have cracked.  In both comparisons there was no correlation between the operation of the engine and the occurrence of LPT blade cracking.

Using the extensive HUMS data the cyclic exchange rate for the LCF usage of the LPT blade was calculated and found to be 2.16 cycles/hour, which was a slight increase from the previous assessment of 2.09 cycles/hour.

To identify any ground handling anomalies that the NFTC Hawk fleet experienced, a questionnaire on the daily operating procedures of the Hawk aircraft was completed by Bombardier Aerospace, Military Aviation Training (BMAT).  Information received did not show anything that would affect the propensity to crack the LPT blade.

Dynamic Stress Addition

The resonant modes of the LPT blades were considered as contributors to the dynamic stress at the root corner.  The maximum increase in the nominal stress predicted is 4% for the worst case.


Top of page 

Annex E:  Additional analysis – visual assessment of Forced Landing profile

Visual Assessment - Engine Failed

When the engine has failed, the HUD is not available and, therefore, a velocity vector, an aircraft symbol and a pitch ladder are not available for pilot reference.  Without these displays, the pilot will have to visually estimate the flight path angle of the aircraft by identifying a fixed point in the windscreen where the terrain does not appear to be moving.  Unless the aircraft is at or very close to the prescribed engine out glide speed, the pilot will have a false sense of where the aircraft is actually gliding towards.

Furthermore, while established on the engine-out glide parameters, pitch inputs will momentarily change the pilot’s glide path perspective such that the aircraft will glide either short of or beyond the fixed point.  Small pitch inputs (+/- 1°) are natural as the pilot strives to maintain the glide parameters and turbulence can easily induce small momentary pitch changes.  This effect is magnified with altitude (ie: distance the aircraft can glide).  For instance, at 15,000 ft AGL (ie: the aircraft will glide 30 NM), a +1° pitch change from the ideal gliding flight path angle (ie. shallower attitude) will move the fixed point in the windscreen across the ground from 30 to 38 NM.  Likewise, a -1° pitch change (ie. steeper attitude) will move the fixed point in the windscreen from 30 to 24 NM.  Combining +/-1° pitch changes from the ideal gliding flight path angle can induce up to 14nm of combined fixed aim point movement, or nearly 50% the total distance the aircraft can glide at that altitude.  Similarly, at 10,000’ AGL( ie. the aircraft will glide 20 NM) a +/-1° pitch change from the ideal gliding flight path angle will move the fixed point across the ground from 20 to 25 NM and 16 NM respectively, a total movement of 9 NM.

As the aircraft approaches the ground the error induced by small pitch inputs decreases so that by the time the aircraft is at 2,500’ AGL (ie. the aircraft will glide 5 NM) a +/-1° pitch change from the ideal gliding flight path angle will induce +/- 1 NM of fixed aim point movement.  If an aircraft can reach a given point (based on altitude), visually assessing whether or not it can actually reach that point will not become clear until close to the ground.

To complicate matters, the fixed point is not actually marked in the windscreen.  The pilot must identify the point in the windscreen where the terrain does not appear to move at the same time that the pitch attitude is naturally fluctuating and causing the terrain against the windscreen to move.  Identification of the fixed point in the windscreen is a rough estimate at best.

To visually determine if the aircraft can reach a given key, the aircraft must be able to glide to the point beyond the airfield appropriate for that key.  The point beyond the airfield must also be estimated.

The degree of uncertainty identifying the fixed point in the windscreen, the impact of small pitch inputs on the apparent distance the aircraft can glide and the requirement to estimate to a point beyond the airfield highlights that it is very difficult for a pilot to accurately assess, through visual means alone, whether or not the aircraft can reach a given key.

Visual Assessment - Engine Operating

When the engine is operating, the HUD is available and, therefore, a velocity vector, an aircraft symbol, and a pitch ladder are available for reference.

With the throttle at idle and the aircraft established at the engine-out glide parameters, the velocity vector identifies the point on the ground where the aircraft will reach.  The pilot needs to estimate the distance beyond the airfield that the velocity vector is positioned in the HUD and use that distance estimate to identify which key that can be reached.  Although use of the HUD velocity vector reduces the errors when visually identifying the point on the ground that the aircraft can reach (ie: velocity vector vice estimating a fixed point in the windscreen), estimating the gliding distance beyond the airfield will at best be an estimate, whether or not the aircraft can reach a particular key.  Given the uncertainty inherent with this method, it is difficult for a pilot to accurately assess, through visual means alone, whether or not the aircraft can reach a given key.

With the throttle above idle, as in the accident situation, the velocity vector cannot be used to identify how far the aircraft will glide when the engine is at idle or off, because it will be positioned according to where the aircraft is currently going with the engine producing thrust, not where the aircraft will go when the engine ceases to produce thrust.  When established at the engine out glide parameters, the aircraft will glide at a flight path angle of -4.7 degrees.  The MFT makes no mention of the engine out glide flight path angle.  When the aircraft is not established in a glide instead of using the velocity vector the pilot can use the pitch ladder to assess where the aircraft will glide when the engine is at idle or off.  The pilot simply needs to identify the point beyond the airfield that is aligned 4.7 degrees below the horizon line, although this method is not described in the MFT.  Although use of the HUD pitch ladder reduces the errors when visually identifying the point on the ground that the aircraft can reach, estimating the gliding distance beyond the airfield will at best be an estimate, when determining whether or not the aircraft can reach a particular key.  Given the uncertainty inherent with this method, it is difficult for a pilot to accurately assess, through visual means alone, whether or not the aircraft can reach a given key.

Visual Assessment - Contribution to Engine Shut Down

In this accident, the crew stated they determined their glide profile by visual means and believed they were close to a glide profile to reach the runway.  When the engine was shut down, the aircraft was 30nm from Cold Lake established in a 260KIAS climb at 12,755ft MSL.  From this position, the MFT prescribed minimum altitude to reach final key was 19,250ft MSL.  Therefore, when the engine was shut down, the aircraft was 6,495 ft below the calculated profile to reach final key.  Further, the aircraft was 4,045ft below the calculated profile just to reach the runway without considering manoeuvring or the selection of landing gear and flaps.  In this accident, the pilots determined their glide profile by visual means and believed that they were close to a glide profile to reach the runway.  Contrary to the crew believing they were close to the glide profile, it is evident that the aircraft was not close to a glide profile that would allow safe recovery of the aircraft. 

Post engine shut down, a climb was initiated trading airspeed for altitude; the aircraft altitude peaked at 13,490’ at 27 NM from Cold Lake.  From this position, the MFT prescribed minimum altitude to reach final key was 18,000ft MSL.  Therefore, the aircraft was 4,310’ below the calculated profile to reach final key.  Further, the aircraft was 2,010’ below the calculated profile just to reach the runway without considering manoeuvring or the selection of landing gear and flaps. 

It is possible that the crew’s misperception of the aircraft’s proximity to the glide profile was due to a misunderstanding about how to visually determine it.  It is also possible that the errors inherent when making a visual estimate of the glide profile contributed to this misperception.  Believing they were close to a glide profile may have made the decision to shut the engine down less difficult.

The investigation determined that had the crew maintained the climb for approximately two more minutes, the aircraft would have intercepted the glide profile to final key.  Since distance was available to the crew, had they calculated their glide profile instead of visually assessing it, they would have been aware of what minimum altitude was needed in order to reach a key position for a successful FL.  It is possible they would have been more inclined to maintain engine operation until they reached a suitable altitude for a FL.


Top of page 

Annex F:  Abbreviations

°C - degrees Celsius

°M - degrees magnetic

AC - alternating current

ACMI - air combat manoeuvring instrumentation

ADS - air data sensor

AFMO - aircraft fleet management organization

AGL - above ground level

AIA - Airworthiness Investigative Authority

AIS - airborne instrumentation subsystem

ALSE - P1/P2 life support equipment

AOA - angle of attack

AOI - Aircraft Operating Instruction

ATC - air traffic control

ATO - Accredited Technical Organization

BFM - basic fighter manoeuvres

BMAT - Bombardier Aerospace, Military Aviation Training

CDA - climb/dive angle

CF - Canadian Forces

CMP - Chief of Military Personnel

CVR - cockpit voice recorder

CYOD - Cold Lake airport

CYLL - Lloydminster airport

CWP - caution and warning panel

CWS - central warning system

D Air CFG - Directorate of Air Contracted Force Generation

DAU - data acquisition unit

DC - direct current

DFS - Directorate of Flight Safety

DME - distance measuring equipment

DND - Department of National Defence

DTAES - Directorate of Technical Airworthiness and Engineering Support

DAEPM - Directorate Aerospace Engineering Project Management

ECA - engine control amplifier

EFH - engine flying hour

EMP - engine monitoring panel

FCU - fuel control unit

FDR - flight data recorder

FL - forced landing

FMEA - failure modes and effects analysis

FPI - fluorescent penetrant inspection

FSIR - Flight Safety Investigation Report

’ - feet

FSOMS - Flight Safety Occurrence Management System

G - normal acceleration in multiples of gravitational acceleration

GPS - global positioning system

GTS - gas turbine starter

HCF - high cycle fatigue

HE - high energy

HP - high pressure

HPC - high pressure compressor

HPMA - human performance in military aviation

HPT - high pressure turbine

HUD - head-up-display

HUMS - health usage monitoring system

IAW - in accordance with

ICAO - International Civil Aviation Organization

IFOV - instantaneous field of view

KIAS - knots indicated airspeed

KTAS - knots true air speed

Knots - knots

LCF - low cycle fatigue

LP - low pressure

LPC - low pressure compressor

LPSV - life preserver/survival vest

LPT - low pressure turbine

LT - local time

m - metres

MDC - miniature detonation cord

MFT - manual of flying training

MHz - megahertz

mm - millimetres

MND - Minister of National Defence

MPA - megapascal

MSL - mean sea level

NFTC - NATO Flying Training in Canada

NH / RPM - high pressure rotor speed / revolutions per minute

NL - low pressure rotor speed

NM - nautical mile

NMSB - Non-Mod Service Bulletin

NOTAM - notice to airmen

OEM - original equipment manufacturer

P1 - Zulu 21 front seat pilot

P2 - Zulu 21 rear seat pilot

PDI - parties of direct interest

PFL - precautionary forced landing

PSP - personal survival pack

QETE - Quality Engineering and Test Establishment

QFI - qualified flying instructor

QRF - quick release fitting

RARM - record of airworthiness risk management

RCA - root cause analysis

RPM - revolutions per minute

RR - Rolls-Royce PLC

SCH - simplified combined harness

SDE - senior design engineer

SEM - scanning electron microscope

SLB - survivor locator beacon

SOP - standard operating procedure

TACAN - tactical air navigation system

Tac F (T) Sqn -  Tactical Fighter Training Squadron

TAM - Technical Airworthiness Manual

TFOV - total field of view

TGT - turbine gas temperature

TSN - time since new

TSO - time since overhaul

TSPI - time-space-position-information

UHF - ultra high frequency

LT - universal coordinated time

UOD - Cold Lake TACAN

USN - United States Navy

LT - coordinate universal time

VCR - video cassette recorder

VHF - very high frequency

VMC- visual meteorological conditions


Top of page 

Notes

[1] From highest to lowest, there are four instructional categories:  A1, A2, B and C.  Upgrading from a lower category to a higher category is based on a formal flight evaluation of the quality of an individual’s flight instruction, their knowledge of the flying syllabus, their flying proficiency and their debriefing.

[2] RCAF Flight Operations Manual, Annex 3.7.1.D.  1 Cdn Air Div Orders, Vol 5-506, Annex A, Appendix 1.

[3] Frog-Bronson training area is restricted class F airspace.  See Figure 1 for an aerial map of Frog-Bronson training area and location from CYOD.

[4] Zulu 22 later described the smoke as a solid dark trail originating near the jet pipe opening.  The trail was thicker at the bottom and wispier at the top with small sporadic puffs.

[5] The engine-out glide parameters are an aircraft glide at 5.5 units angle of attack (AOA) and approximately 185 KIAS, depending on aircraft weight.

[6] Level 5 is a rating of performance, in this case the highest possible rating.

[7] RR was using a subjective measurement technique to measure blade conformance (RARM-155-2007-002 Revision 9).

[8] A test device attached to the engine to measure movement in microns.

[9] The AOA is displayed in relative “units” vice degrees.

[10] Environmental data consists of altitude in feet MSL, IAS in knots and outside air temperature in degrees C.  Engine data consists of RPM in %, NL in % and TGT in degrees C.

[11] NL is the abbreviation for low pressure rotor speed.

[12] Hawk MFT Ch 7, Forced Landing.

[13] Landing roll calculations from the Hawk Operating Data Manual, C 101B-44115-16, Section 8, Landing, pages 11 and 18.  From section 8, page 4, paragraph 16, add a ground run increase of 10% for every 5 knot increase in touchdown speed.

[14] Hawk MFT Ch 7 para 7.06, Rate of Descent engine-out is approximately 1800 fpm.

[15] 10 year averages from 2002 to 2012 as researched for and published in the 2013/2014 Annual DFS Presentation.

[16] B-GA-100-001/AA-000, National Defence Flying Orders, Book 1, Chapter 3.

[17] Pan-CF Blind Drug Testing 2009, Results, Methodology, and Lessons Learned; Defence R&D Canada, Director General Military Personnel Research and Analysis, Chief Military Personnel, DGMPRA TM 2001-019, September 2011.

[18] Annex B to 5555-1 (DPGR 5), Jun 14:  Trained Strength vs Trained Effective Establishment / Preferred Manning Level Fiscal Year 2014/2015

[19] Health and Lifestyle Information Survey of CF Personnel 2008/2009; Director General Military Personnel Research and Analysis, Chief Military Personnel; http://cmp-cpm.forces.mil.ca/health-sante/pub/hlis-sssv/pdf/20082009-rgf-res-eng.pdf.

[20] CDS Letter, Expansion:  Safety Sensitive Drug Testing, 30 May 08.

[21] US Department of Transportation, Federal Transit Administration, Drug and Alcohol Testing Results 2008, Page 5; http://transit-safety.volpe.dot.gov/publications/substance/damis08/pdf/damis08.pdf.

[22] Illicit Drug Testing in the CF – A Discussion Paper, Final Report, June 2007, Dr Brent W. Moloughney.

[23] Douglas A. Wiegmann and Scott A. Shappell (2003).  A Human Error Approach to Aviation Accident Analysis:  The Human Factors Analysis and Classification System.  Ashgate Publishing, Ltd. pp. 48–49.

Top of page

Date modified: