NTSB Factual Report

HISTORY OF FLIGHT

On September 18, 2014, at approximately 0930 Pacific standard time (PST), a jetBlue Airways Airbus A320-232, registration number N656JB, flight number 1416, powered by two International Aero Engines (IAE) V2527-A5 turbofan engines, experienced a No. 2 (right) engine failure and subsequent undercowl fire during initial climb after departing Long Beach Airport (LGB), Long Beach, California. According to the flightcrew, just prior to reaching 10,000 feet above ground level, they received several cockpit Electronic Centralized Aircraft Monitor (ECAM) messages relating to the No. 2 engine including an "ENG 2 FIRE WARNING" and were informed of smoke in the cabin. The flightcrew shutdown the No. 2 engine, discharged both fire bottles, and performed an air turnback to Long Beach. The airplane made a successful and uneventful single-engine landing at LBG and Aircraft Rescue and Fire Fighting (ARFF) personnel met the aircraft and observed no damage. Of the 142 passengers and 5 crewmembers on board the flight, no injuries were reported. The incident flight was a 14 Code of Federal Regulations (CFR) Part 121 domestic passenger flight from LGB to Austin-Bergstrom International Airport (AUS) Austin, Texas. Day visual meteorological conditions prevailed at the time, and an instrument flight rules flight plan was filed.

ENGINE DAMAGE EXAMINATION

On-site examination of the airplane revealed that the No. 2 engine (right) thrust reverser had considerable heat distress and some delamination, and small impact marks (no skin penetrations) on the right hand side aft fuselage near the rear cargo door and to the right hand horizontal stabilizer. Examination of the No. 2 engine revealed considerable low pressure turbine (LPT) blade damage and a fractured fuel pressure line to the station 2.5 low pressure compressor bleed valve slave actuator.

The engine was removed from the airplane and shipped to MTU in Germany for detailed examination. Examination of the outside of the engine revealed evidence of thermal distress such as consumed, partially-consumed or oxidized insulation blankets, loop clamps cushions, wiring harness sheathing, and sooting of various components and cases. No case breaches or penetrations were noted although the LPT case did exhibited a localized outward bulge. Disassembly of the engine revealed that all high pressure turbine (HPT) stage 2 blades were present except for two that were full length releases which included the root. A single fir tree blade retaining lug from the HPT stage 2 disk had fractured between the inner and middle attachment teeth of the fir tree slot and released two HPT stage 2 blades on either side of that fractured disk lug. Turbine hardware upstream of the HPT stage 2 disk did not exhibit any damage as a result of the HPT stage 2 blade releases; however, the remaining HPT stage 2 blades, along with downstream turbine hardware, all exhibited varying degrees of heavy secondary impact damage, tears, and material loss.

TEST AND RESEARCH

Metallurgical examination of the fractured HPT stage 2 disk lug by IAE revealed evidence of fatigue from multiple origins that propagated from the pressure side (PS) of the middle (No. 2) fillet towards the suction side (SS) almost through the entire width of the lug before finally fracturing due to progressive tensile overload. The fractured disk lug was sectioned from the rest of the disk via wire electrical discharge machining to facilitate examination of the fracture surface. Closer examination of the fractured lug revealed a concave 'divot' in the PS No. 2 fillet, immediately adjacent to the fracture surface. The depth of the 'divot' measured up to 0.0008 inches at the fracture origin site and the 'divot' was confirmed to run the entire length of the fillet. Visual examination of all the other remaining lugs revealed that same 'divot' on PS No. 2 fillet and based on this IAE concluded that the groove appeared to be a tool mark resulting from the original machining (broach) operation. Visual inspection using a shadowgraph revealed that the groove/tool mark created an irregular profile and appeared as 'divots' at three locations within the compound radius. Bulk microstructure appeared typical of properly processed IN-100 powder metal.

ADDITIONAL INFORMATION

Since IAE identified the possible source of the 'divot' defect to be attributed to the broaching operation, the HPT stage 2 disk broached before (referred to as disk 7.1) and after (referred to as disk 7.3) the failed disk (referred to as disk 7.2) were initially considered suspect because they were on the same reconditioning/sharpening cycle as the failed disk, meaning that the broaching tool was not removed and sharpened between machining of the three disks. The broaching tool can finish machine 3 disks or 216 slots before it is removed and reconditioned (sharpened) and it can be reconditioned 12 times before it is discarded. Disks 7.1 and 7.3 were removed from service and evaluated by IAE in March 2015. Disk 7.1 had the first 52 blade slots free of defects; however, the last 20 blade slots exhibited the same tool marks, 'divots', observed on the failed disk. Disk 7.3 had all 72 disk slots with the same tool marks that were observed on the failed disk.

Since disk 7.3 had the 'divot' in all the blade slots, IAE had the next sequential disk broached (referred to as disk 8.1), the first disk broached after the broaching tool was reconditioned, removed from service and inspected to determine if the reconditioning of the broaching tool would eliminate what was creating the 'divot' in the blade slots. Disk 8.1 was evaluated by IAE in June 2015 and the examination revealed that all 72 slots exhibited the tool marks on the PS fillet No. 2 as did the failed disk; however, an additional unique tool mark located on the PS No. 3 fillet was found that was not initially found on the failed disk or the other previously examined HPT stage 2 disks. IAE reexamined the previously inspected disks and found traces on the PS No. 3 fillet tool mark on all the disk; qualitatively, the tool mark was more prevalent on Disk 8.1 than on any of the others. According to IAE, the PS No. 3 fillet radius tool mark observed on all the inspected disks, except for Disk 8.1 would not have been a rejectable anomaly.

Reconditioning of the broaching tool did not correct the 'divot' problem, so an audit team made up of IAE, Avio Aero (performed the finished machining/broaching operation), and General Electric (owner of Avio Aero) evaluated the entire manufacturing process with an emphasis on the broaching operation. The evaluation of the Avio disk machining process revealed the following primary contributing factors: 1) cutter tool draft angle design leading to scuffing/sliding along the relief surfaces with associated side loading/deflection and rapid tool wear, 2) a non-optimized tool redressing process resulting in uneven material removal and non-uniform cutter tool profiles, and 3) procedural issues with inspection of tooling, set-up and final parts. Based on these findings, the best practices from GE and IAE have been implemented to address these manufacturing deficiencies.

Based on the findings from disk 7.3 and 8.1, IAE proposed a fleet management plan that would include the issuance of a Non-Modification Service Bulletin (NMSB), anticipated in the first quarter of 2016, for a once-through the fleet inspection of all HPT stage 1 and 2 disks manufactured by Avio at the next engine HPT overhaul. According to IAE, Avio manufactured over 4,000 HPT stage 1 and 2 disks. Discussions with the Federal Aviation Administration indicated that they intend to issue an Airworthiness Directive (AD) mandating the inspection of Avio manufactured V2500 HPT stage 1 and 2 disk based on the IAE NMSB.

NTSB Probable Cause

The probable cause of the engine failure and subsequent undercowl engine fire was due to the fatigue fracture of a high pressure turbine stage 2 disk blade retaining lug that released two blades which impacted the low pressure turbine case causing a fuel line to fracture spraying fuel on the hot engine cases where it ignited. During a machining operation of the disk lug, a tool mark was introduced that set up the area for fatigue cracks to initiate.