On July 19, 1989, United Air Lines Flight 232, a McDonnell Douglas DC-10 in route from Denver, CO to Chicago, IL, with 285 passengers and 11 crew members aboard, suffered an uncontained failure of the Stage 1 fan disk in its Number 2 engine. Due to its location in the vertical stabilizer, immediately above the horizontal stabilizer and the elevator, the uncontained failure of the fan disk generated shrapnel that severely damaged hydraulic lines associated with all three of the aircraft’s hydraulic systems. Hydraulic fluid rapidly leaked out of all three systems. As a result, Captain Al Haynes and his crew lost all use of the aircraft’s flight control systems shortly after the engine failure. Despite all efforts from the pilots and crew, the aircraft crashed and broke apart during the landing attempt.
ESi actively participated in the investigations and litigation that occurred after the accident on behalf of one of the major parties. It was determined that the failure of the fan disk was due to a fatigue crack that originated on the surface of the “bore,” which is a term of art used to describe the hole in the middle of the fan disk. It was further determined that this fatigue crack had been present during the last maintenance inspection of the disk performed 760 cycles prior to the accident (one cycle = one takeoff and landing). The metallurgical investigation, along with stress and fracture mechanics analyses performed by ESi, allowed for the determination of an approximate size for the crack in the shot-peened titanium fan disk at the time of the last inspection.
After an estimate of the size of the crack at the last inspection was obtained, the technical issues hinged on the question of crack detectability. The manufacturer had specified the Fluorescent (dye) Penetrant Inspection (FPI) method for detection of cracks in the fan disk that had been used to clear the part was effective, but ESi theorized that FPI as an inspection technique might not be nearly as effective due to the tightness of the crack.
ESi’s subsequent work included crack growth experiments using shot-peened titanium test articles, crack detectability studies using different Non-Destructive Evaluation (NDE) techniques, and the development of Probability of Detection (POD) data for the various inspection methods. It was ESi’s conclusion that FPI was not the best method of inspection for the highly stressed bore region. In fact, in June of 1990, a revision was made to the engine shop manual requiring Eddy Current (EC) inspection of the bore area of the fan disk.