NTSB Factual Report

HISTORY OF FLIGHT

On June 26, 2003, at 1315 Hawaiian standard time, a Hughes 369D single-engine helicopter, N633WA, was substantially damaged during an autorotational landing following a loss of engine power while maneuvering near Kualapuu, Hawaii. The commercial helicopter pilot, the sole occupant, sustained serious injuries. The helicopter was registered to, and operated by, Windward Aviation Inc., Kahului, Hawaii, under the provisions of CFR Part 133, external load operations. Visual meteorological conditions prevailed at the time of the accident and a flight plan had not been filed. The local fire fighting flight originated from the Molokai Airport, near Kualapuu, approximately 1300.

According to the operator, and the pilot's written statement, the helicopter was utilized to suppress fire by dumping water from an external bucket. The pilot was maneuvering the helicopter over a pond in order to pick up water at 150 feet above ground level (agl) and 40 knots, when he heard a loud bang. He looked outside over his left shoulder and observed white smoke coming from the engine exhaust and noted a loss of engine power. The pilot lowered the collective and began an autorotation; however, with the low airspeed and altitude, he "could not properly flare." He began his flare at 50 feet agl with a "high rate of descent." The helicopter impacted the ground, and after landing, the pilot rolled off the throttle, pulled the rotor brake, and exited the aircraft after he heard a crackling sound.

A fire ignited around the helicopter burning the aft end of the fuselage and forward section of the tail boom.

PERSONNEL INFORMATION

The commercial helicopter pilot accumulated a total of 11,568 hours in helicopters, of which 8,055 hours were logged in the accident helicopter make and model. The pilot also held a flight instructor certificate for helicopters. He was issued a second-class medical certificate on June 6, 2003, with a limitation to wear corrective lenses. His last flight review took place on March 7, 2003, in the same make and model as the accident helicopter.

AIRCRAFT INFORMATION

The accident helicopter was equipped with an Allison 250-C20B turboshaft engine (serial number CAE 830351). It should be noted that manufacturing responsibilities for the Allison 250 engines were acquired by Rolls-Royce, who is mentioned herein as the manufacturer. A review of the maintenance records revealed the engine underwent a turbine assembly modification on April 11, 2002, during a turbine assembly overhaul. The modification was part of a commercial engine bulletin (CEB 1365R1), which allowed certain engines to be converted to obtain better performance "from the improved power turbine air flow characteristics." The conversion included a replacement of the 3rd and 4th stage nozzles and wheels. According to the maintenance records, the modified engine was installed on the accident helicopter on June 8, 2002, at an aircraft total time of 11,761.6 hours. On May 2, 2003, at an aircraft total time of 12,659.3 hours, the helicopter underwent an annual inspection, and the records were endorsed to the effect that the helicopter was in an airworthy condition. On June 6, 2003, at an aircraft total time of 12,752.3 hours, the helicopter underwent a 100-hour inspection in accordance with the manufacturer's 100-hour inspection guides. On the morning of the accident, the helicopter had accumulated a total of 12,804.6 hours.

The manufacturer's recommended time between overhaul for the modified turbine section was 1,750 hours.

The Allison 250-C20B is a two-shaft turboshaft engine with a combination compressor, which consists of a six-stage axial compressor attached to a one-stage centrifugal compressor. The engine incorporates a reverse-flow annular combustor, a two-stage high-pressure turbine (also referred to as the gas producer turbine or N1 turbine), and a two-stage low-pressure turbine (also referred to as the power turbine or N2 turbine). The gas path along the Allison 250 engine flows into the inlet, through the compressor's axial and centrifugal stages, into two external air transfer tubes and to the combustor, which is located at the very rear of the engine. The gases then turn 180 degrees toward the front of the engine and proceed through the two-stage compressor turbine (N1) and a two-stage power turbine (N2). Finally, the gases are directed out of the exhaust duct and upward through two exhaust outlets.

The gas producer (GP) turbine, consisting of turbine wheels and nozzles number 1 and number 2, drives the compressor section of the engine through an inner shaft, while the power turbine (PT), consisting of turbine wheels and nozzles number 3 and number 4, drives the power output gear (to the main rotor transmission) and the accessory gearbox through an outer shaft. The inner shaft rotates independently within the outer shaft.

WRECKAGE AND IMPACT INFORMATION

A Federal Aviation Administration inspector and the operator examined the wreckage at the accident site and noted the 4th stage power turbine wheel shroud track was protruding through the turbine case. In addition, numerous 4th stage turbine wheel airfoils were found separated. Holes were punctured from the inside out in the turbine case and exhaust ducting.

The skids were spread, the pilot seat was collapsed, and the Plexiglass windscreen was melted. The tail boom was separated from the helicopter approximately 1 foot forward of the tail rotor. One of the main rotor blades was separated from the main rotor hub approximately 8 inches outboard of its hub, and was found approximately 100 yards from the main wreckage.

The helicopter was transported to Maui where the engine was removed from the airframe. According to the operator, two fittings were found loose during the engine removal; one was the fuel line upstream of the fuel nozzle (between the fuel control unit and the fuel nozzle), and the other was the oil supply line. The engine was shipped to Rolls-Royce Corporation, Indianapolis, Indiana, for further examination.

TESTS AND RESEARCH

On July 14, 2003, the engine was examined by personnel from the FAA, the operator, the engine manufacturer, and overhaul facility. According to reports from the aforementioned personnel, the N1 and N2 turbine wheels would not rotate. The compressor case's plastic coating had melted during the postcrash fire, seizing the compressor section. Disassembly of the compressor section was not conducted. The upper and lower magnetic chip detectors were removed, revealing a light metallic fuzz. The combustion liner revealed heat and metal splatter damage.

The turbine section was removed from the accessory gearbox for disassembly. Upon disassembly, it was noted that the power turbine's (PT) number 3 wheel had all of its airfoils fractured, and the outer portions of the airfoils and shrouds were missing. The PT rotor was removed from the exhaust collector and it was noted that heavy scoring was present on the PT outer shaft surface. The PT number 3 nozzle displayed impact damage on the trailing edges of the airfoils, sections of which were missing. Heavy rub and impact damage was noted on the number 3 wheel side of the nozzle.

The PT number 4 nozzle displayed heavy impact damage on the airfoil leading edges. Sections of the airfoils' trailing edge were missing. The nozzle revealed heavy rub and impact damage on the PT number 3 and number 4 blade tracks. Review of the PT number 4 wheel revealed the outer portions of the blade airfoils and shrouds were fractured and missing over 225 degrees of the wheel circumference. The number 4 PT wheel also revealed impact damage on the airfoils' leading and trailing edges, and the remaining portion of shroud exhibited rub damage on the outer knife seals.

The gas producer turbine (GP) wheels rotated freely following their removal from the PT section and were not disassembled. The accessory box's gear trains turned freely after the compressor section was removed. The accessory gearbox was not disassembled and was found to be, along with the accessories, in "good condition."

The compressor-to-turbine shaft coupling sustained heavy deformation damage. The turbine end of the shaft was flared outward and revealed heavy rub damage on the outside diameter.

Various components from the power turbine section were examined in more detail at Rolls-Royce's Failure Analysis Laboratory. The following is an account of that examination.

Metallurgical Examination Summary

As previously mentioned, all of the number 3 turbine wheel airfoils were fractured and the outer portions of the blades and shrouds were not recovered. One of the airfoils appeared to exhibit fracture features consistent with fatigue cracking and was labeled airfoil number 1 for reference purposes. The remaining airfoils were numbered sequentially counterclockwise (when viewing the leading edge of the airfoil) for identification purposes. A detailed optical examination of the airfoil fracture surfaces revealed a number of the fractured blades displayed a smeared appearance near the trailing edge, which was attributed to secondary damage. Of note was the fact that a number of blades displayed fracture features similar to that of airfoil number 1; fatigue crack propagation. Airfoil number 1 displayed the most distinct fatigue features and exhibited the largest fatigue fracture. In addition, visual and fluorescent penetrant inspections revealed cracks near the root of every airfoil.

In addition to the damage sustained by the number 3 turbine wheel airfoils, the 3rd stage turbine wheel inner seal, which is an integral part to the wheel, sustained significant rub damage that was in line with the number 1 airfoil. The rub damage was significant enough to wear through the entire thickness of the inner seal.

Closer examination of airfoil number 1's fracture features revealed a smooth portion of the fracture extended approximately 0.5-inch forward from the trailing edge of the blade and exhibited fracture features indicative of fatigue progression. The forward portion of the fracture surface exhibited a rougher texture, indicative of overload. As mentioned previously, the trailing edge of the blade was smeared from secondary damage, which obscured the fatigue cracking origin. A scanning electron microscope (SEM) examination of the fracture surface revealed no evidence of a casting inclusion or metallurgical anomaly. SEM images of the fracture indicate it failed as a result of high-cycle fatigue cracking that progressed away from the trailing edge area.

Airfoil number 15 was examined and it was noted that a crack had developed on the trailing edge of the blade, near its root, and had progressed inboard (toward the turbine wheel's center), then outboard (toward the outer turbine casing) as it traveled toward the airfoil's leading edge. This crack was inboard of the fracture that liberated the outboard section of airfoil. The remaining portion of fractured blade was lab fractured so the inboard crack could be examined closer. The fracture surface was dark and appeared oxidized, which according to Rolls-Royce laboratory personnel, indicated the crack "was present during engine operation." SEM examination of the crack surface revealed it too was failing as a result of high-cycle fatigue cracking that originated on the pressure side of the airfoil near its trailing edge. The fatigue portion of the crack measured 0.046 inches. The remaining portion of the crack displayed signatures indicative of overload and was believed to have "propagated during the failure sequence."

Airfoil number 17 and number 22 displayed cracks with signatures similar to Airfoil number 15. These cracks were also lab fractured to allow closer examination of the crack surfaces, which displayed signatures of high-cycle fatigue cracking that originated at the pressure side of the airfoil's trailing edge, and overload-type cracking as it progressed toward the blade's leading edge. The overload portion of each crack was attributed to propagation that occurred during the failure sequence.

The fracture surface for airfoil number 28 was examined, and did not display any evidence of fatigue cracking. Airfoil number 29 displayed a crack inboard of the failure surface. The remaining portion of blade was lab fractured in order to examine the inboard crack. It was noted that the crack morphology was indicative of overload. During close visual examination of airfoil number 29's inboard crack surface, another crack was noted in the fillet radius, between the blade and the turbine wheel rim. A longitudinal cross-section was cut perpendicular to the crack. It was noted that the crack was relatively flat and contained oxides, which according to Rolls-Royce, indicated "the crack was open during engine operation."

Semi-quantitative X-ray Despersive Analysis determined that the 3rd stage turbine wheel was composed of the material required by the manufacturer's engineering drawings, and the microstructure observed in the airfoil sections were "indicative of fully heat treated material."

Examination of the PT inner shaft revealed both ends of the shaft were damaged, and an area near the middle of the shaft exhibited evidence of rub and thermal damage. Circumferential cracks were observed at the forward and aft ends of the damaged area. In addition, axial cracks were noted on the inner diameter of the shaft at the edge of the rub damage. On the aft end of the shaft, deformed areas were observed that resembled the shape of the number 6 roller bearings. The Rolls-Royce metallurgists believed the rub and deformation damage was considered secondary and was likely due to contact with the turbine-to-compressor coupling during the failure sequence. An undamaged area of the PT inner shaft was examined to establish coating thickness. The coating on the outside diameter of the shaft measured 0.0018-inch thick and was slightly under the engineering requirements; however, the coating thickness was not attributed to being a factor in the failure event. The nickel-plating, on both the inner and outer diameters of the shaft (below the external protective coating), was within the manufacturer's specifications.

The torque required to remove the inner shaft torque nut was not measured; however, the mechanic who removed the nut indicated it was tight. The flange area was dimpled in two places approximately 180 degrees apart.

The PT outer shaft displayed scoring on the outer surface and no fretting was noted on the curvic teeth. The outer shaft torque nut displayed no evidence of wear or damage.

The rotating labyrinth seal displayed no evidence of wear in the inner flange, but witness marks were noted on the forward and aft faces that indicated it was in contact with the 3rd stage wheel and PT inner shaft flange. The inner surface of the seal displayed localized rub and wear damage that was aligned with the wear found on the outer seal but on the opposite side of the diameter and was "likely caused by contact with the hub of the PT inner support." All of the localized wear noted on the seal was considered secondary damage that "likely occurred during the failure sequence."

It should be noted that the cause of the airfoil fatigue cracking and subsequent failure was not determined.

ADDITIONAL INFORMATION

On July 29, 2003, Rolls-Royce issued a "Book Fax" to Allison 250 model authorized maintenance centers (AMC) and military customer support field representatives. The Book Fax (03AMC022) advised the recipients that the aforementioned failure was being investigated, and it was the first known failure of the enhanced power turbine configuration 3rd stage turbine wheel. The Book Fax requested that the AMCs and military field representatives fluorescent penetrant inspect the turbine wheels when the engine and/or turbine sections enter their facilities for any reason. The inspection results, along with the wheel serial number, time and cycles since new, and operator information is to be sent to Rolls-Royce's customer support division. As of this report's writing, no similar failures have been identified.

The engine com

NTSB Probable Cause

the total loss of engine power resulting from a fatigue failure of the 3rd stage turbine wheel airfoils. The reason for the airfoil's fatigue failure could not be determined.