N258JBEmbraer ERJ 190-100 IGW · 2006

On September 3, 2009, a JetBlue Airways Embraer ERJ-190 airplane, N258JB, powered by two General Electric (GE) CF34-10E6 turbofan engines, experienced an engine fire after landing at the Nassau International Airport, Nassau, Bahamas (NAS). According to flight crew and digital flight data recorder information, a No. 1 engine fire master warning occurred during the landing rollout. The flight crew completed the engine shutdown, pulled the fire handle, and discharged one fire bottle. The fire warning light extinguished two minutes and two seconds after the warning. The aircraft was taxied off the active runway and all 84 passengers were evacuated using the cabin door emergency slides on the right side of the aircraft. Eight minor injuries were reported. The flight was operating as flight 1781, scheduled revenue service from Orlando, Florida to NAS under the provisions of 14 Code of Federal Regulations Part 121. There was no major airplane damage.

Damage to the airplane was limited to fuel varnishing on the flap canoe aft of the No. 1 engine exhaust.

The ERJ-190, serial number (S/N) 190.0047, is equipped with two General Electric (GE) CF34-10E6 turbofan engines. The airplane was manufactured by Embraer in September 2006 and delivered to JetBlue in October 2006.

An inspection of the airplane at NAS found evidence of a No. 1 engine under cowl fire. The engine fuel system was pressurized using the aircraft’s boost pumps, and a heavy fuel leak, spraying vertically upward, was observed at the active clearance control (ACC) fuel rod tube coupling on top of the No. 1 engine high pressure turbine (HPT)ACC (HPTACC) valve, which is located at 2 0'clock, viewed from the rear of the engine. The B-nut was tight and properly secured to the valve (male) fitting with a Safety Cable, but appeared to be cocked and cross-threaded on the fitting. There were 3 to 4 fitting threads exposed below the B-nut, which is not consistent with the number of threads exposed when the B-nut is properly installed. The engine was removed and shipped to a maintenance hangar in Orlando, Florida for further investigation.

A review of the airplane’s maintenance records found that there had been no engine removals, no maintenance entries involving the HPTACC valve or the HPTACC valve tubing, and no reports of fuel leakage in the area of the HPTACC valve. JetBlue’s engine health monitoring vibration trend data for the No. 1 engine showed that engine vibration levels had remained within field limits since the engine’s entry into service. The No. 1 engine had accumulated 8,705 flight hours and 5,073 flight cycles at the time of the incident. A review of the engine manufacturing records for the No. 1 engine found that the engine was assembled with the same serial number HPTACC valve.

The engine exhibited thermal damage axially in line with the HPTACC valve and 360 degrees around the core. The damage was most pronounced between 11 and 2 o’clock (clockwise), and included melted or destroyed electrical harness overbraid, blistered flexible fuel line fire sleeves, and fuel varnishing on the external surfaces of the fuel injectors and rigid tubes. One core fire detection loop isolator showed melting sufficient to trigger the fire detection system.

The HPTACC fuel rod coupling is a dual-seal type coupling; in addition to the metal-to-metal seal between the tube ferrule and the valve male fitting, an O-ring between the tube ferrule and the valve fitting serves as a secondary seal. The coupling ferrule is constrained by a standard B-nut fitting threaded to a male fitting installed in the valve.

X-ray photographs were taken of the HPTACC fuel rod tube coupling, viewed 45 degrees aft and 45 degrees forward of the fitting. In both views, a gap could be seen between the ferrule’s conical surface and the valve fitting’s spherical end, showing that the primary seal was not engaged. In both views, the ferrule shoulder appeared to be in contact with valve fitting O-ring gland edge on the inboard side, but a gap was evident opposite this location on the outboard side. A scaled measurement of the outboard gap was taken; this measurement was approximately 0.015 inch. A gap was also evident between the B-nut and top of the ferrule. A properly installed B-nut should have no gap in this area, or at the seal locations. The X-rays showed one-thread engagement on the outboard side of the B-nut.

The HPTACC rod tube retaining clamp was tight and in the specified location on the tube (5.5 inches from the HPTACC valve coupling).

Breakaway torque measured in the tightening direction was 493 lb-ft. The GE production assembly torque is 500 +/- 10 percent. An estimated 314 lb-in of torque was applied in the loosening direction to break the B-nut free. Less than a one-quarter turn was required to fully disengage the B-nut from the fitting. When the B-nut was moved away from the ferrule, a small amount of debris and a piece of material consistent in appearance and dimension with the secondary seal O-ring was found on the fitting, above the threads. The tube clamps were removed and the ferrule was disengaged from the valve fitting. The O-ring was found inside the coupling O-ring gland, with one radial separation, and some debris.

Inspection of the ACC fuel rod tube found that the first B-nut thread was severely damaged and appeared to be deformed (bulged) at the leading edge, approximately 90 degrees past the beginning of the first thread. The ferrule had some circumferential marks on the inside surface near the back chamfer. Inspection of the HPTACC valve fitting found no signs of wear or deformation to the sealing surfaces. The entire first thread was damaged; the remaining threads were intact and showed a dull surface finish. Inspection of the O-ring found that one side of the separation was a relatively clean radial surface, with a small amount of material missing from the outside surface and what appeared to be some small circumferential splits in the material near the missing material. The opposing side of the break had a radial surface through about half the cross-sectional diameter, which changed to a surface angle greater than 45 degrees from radial. There was a split in the material extending from this surface approximately 20 degrees around the circumference of the O-ring. The piece of material found between the ferrule and the valve fitting was viewed under 10x magnification. This piece had a surface angle similar to the O-ring separation surface and showed a round outer surface, sweeping approximately 180°, consistent with the O-ring geometry.

GE performed a modal analysis of the coupling, based on the assumptions that the B-nut was secured to the valve and not participating in motion, that the tube and the ferrule were free to move, and that the tube clamp was in the specified location. No vibration excitation modes within the normal engine operating ranges were found. Forced vibration at 2.0 inches per second, which corresponds to vibration levels recorded in the engine S/N 994207 vibration trend data, showed a possible motion of approximately 0.004 inch. No data were available to determine if this motion would overcome the O-ring static friction (whether the O-ring would wear, or just twist with the o-ring gland). However, there was no evidence of wear on the O-ring, ferrule, or fitting surfaces that would suggest relative motion between the two parts.

An extrusion calculation based on the B-nut stack up measurements and using Figure 3-2, Limits for Extrusion, from the Parker O-ring Handbook, found that an O-ring of 75 durometer hardness at a pressure of 1000 psi, with a total diametric clearance of 0.010 inch should last 100,000 cycles.

A review of GE’s production installation procedure for the HPTACC rod tube at the time of engine manufacture found that the tube was attached at the fuel control during module buildup, and that the clamps that secure the tubes were installed to brackets and tightened at that time. Subsequently, during final engine assembly, the HPTACC valve was installed on the engine and the HPTACC tubes were installed on the valve fittings. As a result of this finding, GE has revised the CF34-10 production build procedure to install the HPTACC tubes on the fittings before tightening the HPTACC tube clamps. In addition, caution notes have been added to the production build procedure to ensure fuel tube B-nuts are fully seated, with no more than two full threads exposed, prior to final torque.

During the procedure review, it was noted that the CF34-10E engine manual (EM) does not call out the installation of the O-rings during installation of the HPTACC valve. Incremental Changes to the CF34-10E EM were issued on October 15 2009 to correct this omission, and

the production assembly procedures were carried over to the EM, to ensure that the fuel tube fittings are correctly engaged and fully seated before applying torque to the B-nuts, and that a maximum of two full threads can remain exposed after final torque is applied.

All references to position are aft looking forward unless otherwise noted. O’clock refers to approximate circumferential locations in a clockwise direction, viewed from the rear of the engine looking forward.

The Government of the Bahamas delegated this investigation to the United States as the State of Registration and as the State of Design and Manufacture of the engine. The Government of Brazil participated in this investigation as the State of Design and Manufacture of the airplane.