N31VABell 407 · 2000

On August 12, 2017, about 1649 eastern daylight time, a Bell 407 helicopter, N31VA, was destroyed when it was involved in an accident in Charlottesville, Virginia. The pilot and the observer were fatally injured. The helicopter was operated as a public aerial observation flight.

According to the Virginia State Police (VSP), the purpose of the flight was to provide the VSP command center with a continuous video downlink of the public demonstrations that were occurring in Charlottesville. The helicopter departed Charlottesville Albemarle Airport (CHO) about 1600. The helicopter arrived over the area of the demonstrations at 1604 and remained there until 1642 when the flight crew was tasked to provide overwatch for the Governor of Virginia's motorcade. At 1643, the flight crew advised the VSP command center that the helicopter was heading directly to the motorcade and was about 30 seconds away.

Radar data provided by the Federal Aviation Administration (FAA) indicated that, at 1648, the helicopter was flying at an altitude of about 2,200 ft mean sea level (msl) in the area of the motorcade. At that time, the helicopter was traveling north-northwest before it began to turn to the right and descend rapidly. Radar data indicated that, at 1648:30, the helicopter was descending at a rate of 6,800ft/min through 1,450 ft msl at a groundspeed of 30 knots. The helicopter then descended below the area of radar coverage, and radar contact was lost.

About 1649, a crewmember aboard a Fairfax County Police Department (FCPD) helicopter observed the accident helicopter descending upright into trees at a high rate of descent and then observed a "stirring" of debris. The crewmember advised the pilot, who immediately contacted the VSP command center to report that a helicopter had crashed. The pilot of the FCPD helicopter attempted to contact the accident helicopter but was unable to make contact with the flight crew. The FCPD helicopter pilot then landed near the accident site to render aid. The other two crewmembers exited the helicopter and proceeded to the accident site. Upon reaching the accident site, the crewmembers encountered heavy black smoke and fire.

The VSP interviewed 47 witnesses to the accident. Although their descriptions of the helicopter's altitude, direction of flight, and velocity varied, most witnesses reported that the helicopter, after initially hovering, entered a rolling oscillation, began to spin about its vertical axis, and descended in a 45° nose-down attitude while continuing to spin. Witnesses reported that they lost sight of the helicopter below the tops of the surrounding trees and then observed a plume of smoke rising from the area.

Video from a security camera located about 1.2 miles from the accident site captured images of the helicopter in a vertical descent with increasing vertical speed as the helicopter continued to descend toward the ground. Still photographs taken by a witness showed that the helicopter was spinning in a clockwise direction (when viewed from above the helicopter).

The pilot had been employed with the VSP aviation unit since 1999 and became the unit commander in December 2012. The observer had been employed with the VSP aviation unit since July 2017.

During this mission, the accident helicopter was configured with single main controls at the pilot's station and locked out pedals at the copilot (observer) station.

The accident helicopter's turbine engine had a full authority digital engine control (FADEC) system. The engine control unit (ECU) would continuously monitor the FADEC system for faults and would alert the pilot of any faults that could significantly impact engine performance.

The accident helicopter was also equipped with an airspeed-actuated pedal restrictor control system (PRCS), which reduces total left pedal travel at higher airspeeds by automatically adjusting the left pedal's forward stop. When the helicopter accelerates above 55 knots indicated airspeed (KIAS), the PRCS solenoid energizes, engaging a cam that limits forward travel of the left pedal by 25%, which reduces tail rotor blade angle from 25° to 17° When the helicopter decelerates below 50 KIAS, the PRCS solenoid de-energizes, which disengages the cam and enables full forward travel of the left pedal.

A sounding (a high-resolution rapid refresh model) for the accident site and time depicted a light and variable wind of 3 knots with clear skies over the accident site. No significant turbulence or wind shear was detected.

During this mission, the accident helicopter was configured with single main controls at the pilot's station and locked out pedals at the copilot (observer) station.

The accident helicopter's turbine engine had a full authority digital engine control (FADEC) system. The engine control unit (ECU) would continuously monitor the FADEC system for faults and would alert the pilot of any faults that could significantly impact engine performance.

The accident helicopter was also equipped with an airspeed-actuated pedal restrictor control system (PRCS), which reduces total left pedal travel at higher airspeeds by automatically adjusting the left pedal's forward stop. When the helicopter accelerates above 55 knots indicated airspeed (KIAS), the PRCS solenoid energizes, engaging a cam that limits forward travel of the left pedal by 25%, which reduces tail rotor blade angle from 25° to 17° When the helicopter decelerates below 50 KIAS, the PRCS solenoid de-energizes, which disengages the cam and enables full forward travel of the left pedal.

Accident Site

The main wreckage came to rest in an upright position along a magnetic heading of 333° in heavily wooded terrain that was adjacent to a residence. The main wreckage comprised the main fuselage (cockpit and cabin), aft fuselage, forward section of the tailboom, midsection of the tailboom (including the horizontal stabilizer), main rotor system, and engine. The main wreckage showed damage consistent with impact with trees and the ground. The main fuselage, aft fuselage, main rotor system, and engine were thermally damaged from the postcrash fire. The landing gear exhibited multiple fractures and a flattened appearance.

The aft section of the tailboom, containing the tail rotor gearbox, tail rotor, and vertical stabilizer, was found about 40 ft above the ground in a tree and about 100 to 150 ft south-southwest of the main wreckage. Debris from the fragmented tailboom was found in a debris field that spanned about 300 ft in length west of the main wreckage. Examination of the fragmented tailboom sections revealed multiple angled cuts consistent with main rotor blade contact.

Main Rotor System

The main rotor hub remained attached to the main rotor mast, and the four main rotor blades remained attached to their respective hub locations. For one of the main rotor blades, the pitch horn lug (for the pitch control link upper rod end) was fractured from its pitch horn. The pitch horn lug fracture surface exhibited signatures consistent with overload and thermal damage. Neither the fractured pitch horn lug nor the pitch control link upper rod end were found. The remainder of the major components of the main rotor system were found within or near the main wreckage site and exhibited fragmentation from impact and thermal damage.

The main rotor gearbox remained attached to the airframe. Drive continuity was established within the main rotor gearbox. The engine-to-transmission driveshaft was present, but its aft coupling was fractured.

Tail Rotor and Tail Rotor Drive System

Sections of the tail rotor drive system, from the steel tail rotor drive shaft at the forward end to the tail rotor gearbox input at the aft end, were recovered from the main wreckage, the debris field, and the aft tailboom section. Reconstruction of the tail rotor drive system revealed that most of the components were present except for the No. 3 tail rotor drive shaft, the forward portion of the No. 4 tail rotor drive shaft, and the hanger bearing between the Nos. 3 and 4 tail rotor drive shafts, which were not found. Fractures observed on the Nos. 1, 2, and 4 tail rotor drive shaft tubes were consistent with main rotor blade contact and were co-located with the angled cuts observed on the tailboom.

The tail rotor gearbox remained installed on the tailboom, and drive continuity within the gearbox was established. Residual oil was present within the gearbox, and the magnetic chip detector revealed no evidence of debris. The tail rotor remained installed on the tail rotor gearbox output shaft. Both tail rotor blades remained installed and were intact. One tail rotor blade displayed damage to its tip end, consistent with contacting the left side of the tailboom. The tail rotor blade leading edge also displayed a damaged area about 3 inches wide and about 15.5 inches inboard from the tip. The other tail rotor blade exhibited no anomalous damage.

Engine

The engine was found in the main wreckage lying on its right side near its installed location. All engine mounts had fractured in overload. The engine exhibited impact damage and was bent at an angle of about 30° at the junction of the turbine and gearbox modules. All major components for the engine were found at the main wreckage site.

The oil and pneumatic lines were manually checked and none showed evidence of looseness. The leading edges of the compressor impeller blades exhibited evidence of hard-body foreign object debris ingestion.

The ECU was found in the main wreckage near its installed location with one of its electrical connectors still attached. The ECU exhibited thermal damage due to the postcrash fire.

Flight Controls Systems

The three main rotor actuators were found in the main wreckage near their installed locations and exhibited impact and thermal damage. The main rotor controls, from the cyclic and collective to the swashplate, and the tail rotor controls, from the pedals to the forward section of the tailboom, sustained multiple fractures due to impact forces, and portions were consumed by the postcrash fire. Pieces of the tail rotor control tube from the midsection of the tailboom were recovered in the debris field. The tail rotor control system remained intact within the aft section of the tailboom (which was found in a tree, as previously discussed) except for slight bending of the tail rotor pitch control rods near the rod ends. The recovered main rotor and tail rotor controls showed no evidence of disconnection.

The PRCS remained installed, but its solenoid exhibited impact damage. The PRCS cam was found in the engaged position. The PRCS emergency release cable (which enables manual disengagement of the PRCS pedal stop) was found in the cockpit and was thermally damaged, and the copper wire for the emergency release cable (which prevents the inadvertent disengagement of the PRCS pedal stop and provides an indication for when the emergency release has been pulled) was found unbroken. The plastic pull knob for the emergency release cable was not recovered, and the adjacent cable housing exhibited thermal damage.

A pitot-static test bench was used to functionally test the pedal restrictor control unit (a PRCS component). The unit responded normally in activating and extinguishing the PRCS engagement and solenoid activation lights when pitot-static pressure (to simulate airspeed) was increased and decreased, respectively. The solenoid functioned normally when power was applied to it. Functionality of the emergency release cable was confirmed.

Avian Material Examination

During the investigation, no evidence was observed to suggest that the accident was the result of a mid-air collision involving another aircraft, or object, and examination of samples taken from the main rotor, nose, windscreen, and cockpit areas were examined for microscopic avian material. No bird remains were found in any of the samples.

Vortex Ring State

According to the FAA's Helicopter Flying Handbook (FAA-H-8083-21B), a vortex ring state "describes an aerodynamic condition in which a helicopter may be in a vertical descent with 20 percent up to maximum power applied, and little or no climb performance." The handbook also states the following:

A fully developed vortex ring state is characterized by an unstable condition in which a helicopter experiences uncommanded pitch and roll oscillations, has little or no collective authority, and achieves a descent rate that may approach 6,000 feet per minute (fpm) if allowed to develop….

Situations that are conducive to a vortex ring state condition are attempting to hover OGE without maintaining precise altitude control, and approaches, especially steep approaches, with a tailwind component.

When recovering from a vortex ring state condition, the pilot tends first to try to stop the descent by increasing collective pitch. The traditional recovery is accomplished by increasing airspeed, and/or partially lowering collective to exit the vortex. In most helicopters, lateral cyclic thrust combined with an increase in power and lateral antitorque thrust will produce the quickest exit from the hazard.

Vortex Ring State Training

According to the FAA's Helicopter Instructor's Handbook (FAA-H-8083-4), vortex ring state (also known as settling with power) can safely be introduced and practiced at altitudes allowing distance to recover. The handbook also states the following:

Ensure the student understands that settling with power can occur as a result of attempting to descend at an excessively low airspeed in a downwind condition, or by attempting to hover OGE at a weight and density altitude greater than the helicopter's performance allows….

Recovery is accomplished by…if altitude allows, reducing collective and lowering the nose to increase forward speed. This moves a helicopter out of its downwash and into translational lift. When the helicopter is clear of the disturbed air, or downwash, confirm a forward speed indication and initiate a climb to regain the lost altitude.

Virginia State Police Aviation Unit Training Manual

The VSP aviation unit training manual required that its unit instructors refer to the current Federal Aviation Regulations and the FAA's practical test standards for standardization. Review of the practical test standards for rotorcraft revealed a required task for settling with power (vortex ring state) for which pilots were to (1) exhibit knowledge of the elements related to settling with power, (2) promptly recognize the onset of settling with power, and (3) use the appropriate recovery procedure.

Review of the VSP aviation unit training manual revealed that vortex ring state was not listed in any of the sample lesson plans for initial or recurrent training and that the associated maneuvers were considered to be optional. Anecdotal information indicated that the pilot had knowledge of vortex ring state, but review of the accident pilot's training records from 2001 to the accident found no record of him receiving settling with power or vortex ring state recognition and recovery training on the accident helicopter make and model.

The Virginia Department of Health, Office of the Chief Medical Examiner, Richmond, Virginia, performed autopsies of the pilot and the observer. The pilot's cause of death was blunt force injury to the head, torso and extremities, and the observer's cause of death was blunt force injuries to the head and torso. The autopsy also identified the pilot's moderate coronary artery disease with a 60% stenosis of the left anterior descending coronary artery. The remainder of the heart examination was unremarkable.

Toxicology testing at the FAA Forensic Sciences Laboratory were negative for the pilot for carbon monoxide, ethanol, and all drugs tested. The testing for the observer detected naproxen in his urine samples. Naproxen is a non-narcotic analgesic and anti-inflammatory agent that is available over the counter and as a prescription. Carbon monoxide and ethanol were not detected in the observer's specimens.

Performance Study

The performance study for the accident flight was conducted using three data sources: 1) data recovered from the ECU; 2) radar data from airport surveillance radar (ASR)-9, which was located about 3 nautical miles north of CHO; and 3) automatic dependent surveillance – broadcast (ADS-B) system data.

ASR-9 radar data showed that the helicopter left the downtown Charlottesville area about 1644:00 and flew to the southwest and then to the north. The terrain below the helicopter had an elevation from 300 to 600 ft. The helicopter's maximum groundspeed was above 100 knots early in the flight and then varied from 5 to 80 knots during the rest of the flight.

ADS-B data toward the end of the flight indicated that the helicopter was climbing and that its calculated forward airspeed was slowing until 1646:00, when the helicopter leveled off at 1,950 ft for about 1 minute. The helicopter then began climbing again, reaching an altitude of 2,250 ft, and its forward airspeed slowed from 30 to about 20 knots. At 1648:06, the helicopter's forward airspeed increased to 30 knots. Four seconds later, the helicopter climbed from 2,225 to 2,275 ft, and its forward airspeed slowed to about 10 knots. The helicopter's descent began at 1648:18.

The nonvolatile memory from the ECU was successfully downloaded. About 20 seconds of parametric data, which included rotor speed, torque, collective position, gas generator speed, and absolute ambient pressure, were recorded at the end of flight. The ECU data revealed an increase in torque, from 54% to 104%, immediately before the helicopter's descent. The ECU data also indicated that, between 1648:18 and 1648:20, the collective position decreased from 40% to 14% and that, during the next second, the collective position increased to about 30%. Even as the collective continued to increase to a peak of 68% by 1648:31, the helicopter's altitude decreased, indicating that the helicopter did not respond to the increase in collective.

A video study determined the motion of the helicopter based on the security camera video, which captured about 10 seconds of the helicopter's descent but not the beginning of the descent, and a sequence of four still photographs, which were taken during a 2-second period. The video study indicated that, according to the security camera video, the helicopter was descending with an estimated vertical acceleration of 12 ± 1.5 ft/s2. The photographs indicated that the estimated yaw rate of the helicopter about 20 seconds before impact was at least 92° ± 5° per second in the clockwise direction. The helicopter had already started descending at the time that this estimated yaw rate occurred. The helicopter tail structure appeared undamaged in the photographs.

After 1648:16, the helicopter's low forward speed while descending put it in or near a region conducive to a vortex ring state, which is an aerodynamic condition that occurs when the helicopter descends at the downward speed of its own vortex wake. The vortex system accumulates, building in strength and producing increased downwash through the main rotor. The rotor, operating in a high downwash field, is unable to arrest the helicopter's descent rate, even with increased collective. Even though the collective was raised after 1648:20, the helicopter's altitude did not increase. The security camera video and the photographs of the descent, which were determined to be after 1648:22, showed that the helicopter rolled to the left, between 30-57 degrees, as it was spinning to the right.

Pedal Restrictor Control System Calculations

To determine the effect on tail rotor authority if the PRCS were to remain engaged below 50 KIAS, the National Transportation Safety Board (NTSB) requested that Bell perform calculations to determine the left pedal control margin that would be available for different airspeed conditions with the PRCS cam engaged. The calculations used conditions similar to those on the day of the accident: a gross helicopter weight of 4,633 pounds, an ambient temperature of 86°F (30°C), and a pressure altitude of 2,200 ft.

Bell determined that the left pedal margin would increase with increasing airspeeds and that a hover out of ground effect (OGE) would be the most critical condition for restricted left pedal authority. The calculated tail rotor blade collective pitch angle that would be needed to maintain heading while hovering OGE was between 10° and 11°. If the PRCS were to remain engaged during a hover, a tail rotor blade collective pitch angle of 17° could be achieved with full left pedal travel restricted by the PRCS. Without the PRCS engaged, a tail rotor blade collective pitch angle of about 25° +/- 0.5 could be achieved with the left pedal at its unrestricted full forward position.

On May 11, 2010, about 1335 eastern daylight time, a Bell 407 helicopter, N31VA, operated by the Virginia State Police, was substantially damaged during an emergency landing following an engine failure and autorotation near Virginia Highlands Airport (VJI), Abingdon, Virginia. The certificated flight instructor (CFI) and the commercial pilot were not injured. Visual meteorological conditions prevailed, and no flight plan was filed for the public use instructional flight.

According to the crew, who were both Virginia State Police (VSP) officers, they and the helicopter were based at VJI. The purpose of the flight was to provide aircraft orientation training for the pilot. The pilot conducted one takeoff and landing at VJI, and then proceeded to the northwest to practice confined area operations at a field about 3 miles from VJI. When the flight was enroute to the practice field, the crew heard a noise that they described as a "very low growl" coming from the engine compartment. As they approached the field at an altitude of about 250 feet above ground level (agl) and a speed of 80 knots, they heard a "very loud growl," and the engine "surged" twice. The "FADEC DEGRADE" caution light illuminated, and an aural "ENGINE OUT" alert sounded. The CFI informed the pilot that he was taking control of the helicopter, and then initiated a 180 degree right turn, and an autorotation to the field. At about 50 feet agl, the CFI flared the helicopter, and then landed it. The helicopter bounced one time and came to rest on a "slight slope" in the field, with the right skid on the uphill side. The crew shut down the helicopter and exited normally. The pilot reported that the flight duration was seven minutes.

Two days after the accident, the helicopter was recovered to a Bell Helicopter completion and maintenance facility. The engine, including the electronic control unit (ECU) and fuel control, was removed and shipped to the Rolls-Royce facility in Indianapolis, Indiana for detailed examination and testing. Since the helicopter was a public use aircraft, a week after the accident, the VSP formally requested that the Federal Aviation Administration (FAA) "conduct a formal investigation" into the accident.

Pilot and FAA records indicated that the CFI held an airline transport pilot certificate, with several ratings, including rotorcraft-helicopter, and a flight instructor certificate with rotorcraft-helicopter and instrument helicopter ratings. The CFI's most recent FAA second-class medical certificate was issued in July 2009, and his most recent flight review was completed in August 2008. He reported that he had 3,278 total hours of flight experience, which included 2,316 hours in helicopters, of which 1,100 hours were in the accident helicopter make and model.

Pilot and FAA records indicated that the pilot held a commercial certificate, with several ratings, including rotorcraft-helicopter, and instrument helicopter. The pilot's most recent FAA second-class medical certificate was issued in April 2010. He reported that he had accumulated approximately 4,328 total hours of flight experience, which included 4,007 hours in helicopters, of which 147 hours were in the accident helicopter make and model.

According to FAA records, the helicopter was manufactured in 2000, and was first registered to the Commonwealth of Virginia in January 2001. Examination of the maintenance records revealed that the records system utilized three separate hour-tracking categories, as well as an engine "cycles" value. The three hour-tracking categories were "Hobbs," "Aircraft TT (total time)," and "Engine TT." Examination of the records from October 2009 to the date of the accident indicated a constant difference between the aircraft and engine TT values; the aircraft TT value was 198.0 hours more than the engine TT value. In contrast, the Hobbs value did not maintain a constant difference from those values, but was about 43 hours more than the aircraft value.

The helicopter was equipped with a Rolls-Royce 250-C47B engine. According to the maintenance records, the most recent annual inspection was completed in October 2009. The most recent 50 hour/3 month airframe inspection was completed on April 20, 2010, and the most recent 150 hour engine inspection was completed the following day. As of those two latter inspections, the aircraft TT was about 3,887 hours, the engine TT was about 3,689 hours, and the engine had accumulated 6,631 cycles. The helicopter and engine accumulated about 14.5 hours between those inspections and the accident.

The VJI 1343 recorded weather observation included wind from 200 degrees at 10 knots, with gusts to 18 knots, visibility 10 miles, scattered clouds at 3,700 feet agl, broken cloud layers at 4,200 and 5,000 feet agl, temperature 21 degrees C, dew point 12 degrees C, and an altimeter setting of 30.20 inches of mercury.

Representatives from the FAA and Rolls-Royce arrived at the accident scene the day after the accident. They reported that both landing skids were splayed in the outboard direction, and that the right skid exhibited more deformation than the left skid. The upper 3 inches of the left vertical stabilizer was missing, and one main rotor blade had paint transfer marks consistent with stabilizer contact. The tail skid and tail rotor blades were intact. The forward-looking infrared (FLIR) turret that was mounted on the underside of the fuselage below the left rear seat was pushed up, and penetrated the cabin floor. The "Night Sun" lamp that was mounted on the underside of the fuselage, below the left front seat, was damaged, but did not penetrate the cabin. The remainder of the airframe, main rotor and tail rotor were otherwise intact. Movement of cockpit controls confirmed continuity to all control surfaces.

The helicopter had approximately 790 pounds of fuel on board at the time of the event. All fuel, lubrication, and pneumatic lines were checked for damage, continuity and security; all were intact. The engine was found securely in position, with all attaching hardware in place and secure. A visual inspection of the engine exterior did not reveal any damage.

Checks were then made of the N1 and N2 drive trains. Motoring of the engine to approximately 10 percent rpm resulted in smooth and continuous rotation from the starter generator to the compressor. Rotation of the main rotor head showed resultant smooth and continuous rotation to the No.4 power turbine wheel. No attempt to start the engine was made.

Engine Data Recorder

According to the Rolls-Royce representative, the ECU was equipped two separate non-volatile memory (NVM) units, known as the "maintenance terminal" (MT), and the "incident recorder" (IR). The MT recorded discrete events relevant for maintenance purposes, and the IR recorded time history data of engine parameters. The IR recording was designed to start whenever a "trigger" (parameter exceedance) was detected; the recording would capture data from 12 seconds prior to the trigger, and continue after the trigger.

Visual examination of the ECU found it to be securely in position, with its data connectors in place. The ECU NVM data were downloaded. Examination of the data revealed that two "engine surge" events were captured in the "Last Engine Run Fault" section of the recording. Also, the "Accumulated Faults" data revealed only one temperature exceedance. That value was a gas temperature exceedance of 1.25 seconds duration, with a maximum temperature of 1,712.9 degrees F. The exceedance did not have a time-of-occurrence associated with it, but it was the opinion of the Rolls-Royce representative that it most likely occurred during the engine anomaly/fault event.

The Rolls-Royce Maintenance Manual (MM, Sec 72-00-00 p 20) required the following:

• During engine starts, gas temperatures between 1,700 and 1,830 degrees require an inspection of the turbine, and entries in the engine maintenance records (including temperature and duration)

• During power transients, any gas temperatures above 1,661 degrees requires that the turbine be removed for "heavy [maintenance] or overhaul"

No previous engine overspeed or overtemp exceedance events were noted in the engine maintenance records.

Detailed Engine Examination

On June 10, 2010 an engine investigation was conducted at the Rolls-Royce facility. In attendance was an FAA inspector, and representatives of Rolls Royce, VSP, and Bell Helicopter.

Visual examination of the compressor module exterior revealed no damage. When rotated manually, the compressor exhibited smooth operation, both before and after separation from the accessory gear box. Disassembly and inspection of the compressor front support, compressor rear support, impeller, and compressor shroud revealed no damage.

Prior to separation of the engine modules, manual rotation of both the N1 and N2 drive trains at the tachometer generator pads revealed smooth and continuous rotation of the N1 and N2 gear trains through the accessory gear box. Visual examination of the gearbox interior revealed that it contained clean oil, and no damage was noted.

The outer combustion case and both air discharge tubes were properly positioned, and no external damage was noted. Removal of the outer combustion case revealed a metal strip, approximately 2 inches long, lodged between the basket and the inner wall of the outer combustion case. A second, similar metal strip was observed bent around, and lodged in, one of the dilution holes of the combustion liner. Visual examination of the interior of the combustion liner revealed no unusual streaking, or other evidence of thermal damage. The support plate (the "-6" component in the manufacturer's Illustrated Parts Catalog (IPC), commonly referred to as the "deflector plate"), which was normally located at the aft end of the combustion liner, was absent.

Examination of the turbine module revealed heavy metallic spatter across the aft face of the No.1 nozzle shield, and a smearing of a yellow substance around the outer rim of that nozzle shield. A foreign strip of metal was found bent around a first stage nozzle vane saddle. The No.1 nozzle exhibited discoloration of the vane surfaces consistent with excessive thermal exposure. Several turbine vanes and the No.1 turbine wheel of the gas producer section exhibited foreign object impact damage. All the turbine blades were damaged, and four blades were missing the majority of their airfoil length.

The No.2 turbine wheel exhibited metal spatter across the blade surfaces with foreign object impact damage to many blade leading edges. The No.2 turbine nozzle exhibited nicks across the leading edges of several vanes. The trailing edges of approximately 20 percent of the vanes exhibited thermal damage consistent with over-temperature. Metal spatter was noted across the vane surfaces. The Nos.3 and 4 turbine wheels all exhibited nicking across the blade leading edges, and light spatter across the blade surfaces.

The lower chip detector was clear of any metallic particles; the upper chip detector exhibited light metallic particles or slivers. All other engine components appeared normal and undamaged.

Combustion Liner History

The combustion section consisted of an outer combustion case and an inner combustion liner. The liner was supported at the forward end by the gas producer nozzle vane assembly, and at the aft end by the fuel nozzle, which was mounted in the aft end of the outer combustion case.

Review of the helicopter and engine maintenance records indicated that the accident combustion liner (part number 23064570, serial number PHI-0020) was originally manufactured by Rolls-Royce. In 2008, the liner was sent to Cadorath Aerospace Lafayette LLC (CAL LLC) for inspection, with the possibility for overhaul if required. At that time, CAL LLC was a Rolls-Royce designated "authorized repair facility."

According to the repair facility's work order "traveler" document, the liner was received and visually inspected. Subsequent detailed inspection revealed that the liner did not conform to the inspection criteria, and was therefore rejected, which denoted that it was no longer an airworthy component. The document indicated that cracks were present in the liner and the deflector plate, and that attempts to weld-repair those cracks were unsuccessful. The document also indicated that the liner was partially disassembled, a new deflector plate and associated spacers were installed, and the liner was reassembled. The liner disassembly and re-assembly process included cutting, machining, brazing and welding. In October 2009, when the engine had a TT of 3,544.9 hours, its combustion liner was removed, and the overhauled combustion liner was installed.

Combustion Liner Repair Details

As noted above, the combustion liner deflector plate, which normally surrounded the fuel nozzle boss, was found absent from its normal position. Failure analysis of the remaining liner revealed that the required circumferential fillet weld between the liner and the deflector plate was not performed during the repair at overhaul; only the plug (positioning) welds were present to affix the deflector plate to the liner. The combustion liner was approved for return to service on October 24, 2008. The TT on the combustion liner could not be determined, but at the time of its failure, it had accumulated 158.4 hours since overhaul and installation in the accident engine. The manufacturer's MM-specified "recommended time between overhaul" (TBO) for the combustion liner was "On Condition," which the MM explained as the component "May remain in service provided operation and condition are satisfactory."

The engine manufacturer's maintenance documentation, included the Overhaul Manual (OHM), the Overhaul Procedures (OHP) manual, the Parts Repair Procedures Letters (PRPL) and the Illustrated Parts Catalog (IPC). The OHM provided top-level repair information, and specific guidance was contained in the OHP and PRPL. Examination of those documents revealed that while the engine manufacturer permitted the deflector plate to be replaced, there was no specific guidance for that procedure in the OHP or PRPL, and therefore the repair facility was not authorized to conduct that procedure. In addition, the repair facility did not possess the applicable guidance for replacement of the deflector plate. The investigation was unable to determine the specifics of how or why the repair facility replaced the deflector plate, and then inspected and approved that replacement.

Although the combustion liner was approved for return to service by the repair facility about 19 months prior to its failure, the investigation did not locate any information that indicated that either the repair facility or the FAA principal maintenance inspector (PMI) for the repair facility was aware that the repair facility had accomplished a procedure that it was not authorized to conduct. In addition, there was no evidence to indicate that the FAA or the repair facility attempted to remove the subject combustion liner from service prior to its failure.

On December 21, 2010, after the repair facility was advised of the deflector plate failure mode due to the improper repair, the repair facility identified other assemblies which had a known or suspected improper repair, and recalled those assemblies from their customers. That action was accomplished by means of a repair-facility-issued "Urgent Stop Use and Product Recall Notice," which listed a total of 19 units.

On March 24, 2011, the engine manufacture sent a "letter of finding" to the repair facility. The letter formally advised the facility that the deflector plate replacement was not an authorized procedure per OHP 72-40-14-01, and provided details regarding the deficiencies of the repair facility's procedures as executed. The letter instructed the facility not to conduct any such repairs in the future, to identify and recall any previously-affected combustion liners, and to notify the manufacturer once all suspected liners were successfully recalled. The FAA PMI was notified of those findings and actions. On April 18, 2011, the repair facility informed the engine manufacturer in writing that all 19 suspected combustion liners had been successfully recalled, and that no additional in-service failures had occurred.

In September 2011, the engine manufacturer modified section 72-40-00 of its OHM to more clearly state that replacement of the deflector plate by a repair station could only be accomplished by replacement of the next-higher assembly, the pre-ignition sub-assembly.