The failure of HP turbine blades in a low bypass turbofan engine was analyzed to determine the root cause. Forensic and metallurgical investigations were conducted on all failed blades as well as failed downstream components. It was found that one of the blades fractured in the dovetail region, causing extensive damage throughout the turbine. Remedial measures were suggested to prevent such failures in the future.

Aluminide-coated and uncoated IN-713 turbine blades were returned for evaluation after service in a marine environment because of severe corrosion. Based on service time, failure of these blades by corrosive deterioration was considered to be premature. Analysis (visual inspection, 2.7x micrographic examination on sections etched with ferric chloride and hydrochloric acid in methanol) supported the conclusions that the blades failed by hot-corrosion attack. Variation in rate of attack on coated blades was attributed to variation in integrity of the aluminide coating, which had been applied in 1966, when these coatings were relatively new. It is evident that maintaining the integrity of a protective coating could significantly increase the life of a nickel-base alloy blade operating in a hot and corrosive environment.

Two 20 MW turbines suffered damage to second-stage blades prematurely. The alloy was determined to be a precipitation-hardening nickel-base superalloy comparable to Udimet 500, Udimet 710, or Rene 77. Typical protective coatings were not found. Test results further showed that the fuel used was not adequate to guarantee the operating life of the blades due to excess sulfur trioxide, carbon, and sodium in the combustion gases, which caused pitting. A molten salt environmental cracking mechanism was also a factor and was enhanced by the working stresses and by the presence of silicon, vanadium, lead, and zinc. A change of fuel was recommended.

In an electric power station, seven turbine blades out of 112 broke or cracked within 8 to 14 months after commencement of operation. The blades in question were all located on the last running wheel in the low pressure section of a 35,000 kW high pressure condensing turbine. They were milled blades without binding wires and cover band. They did not fracture at the fastening, i.e. the location of highest bending stress, but in a central region which was 165 to 225 mm away from the gripped end. The blades were fabricated from a stainless heat-treatable chromium steel containing 0.2C and 13.9Cr. Microstructural examination showed the blades were destroyed by flexural vibrations which evidently reached their maximum amplitude at the location of fracture. Erosion of the inlet edge, possibly in connection with vibration-induced corrosion cracking, contributed to fracture.

Cracking in gas turbine blades was found to initiate from a mechanism of low-cycle fatigue (LCF). LCF is induced during thermal loading cycles in gas turbines. However, metallography of two cracked blades revealed a change in microstructure at as-cast surfaces for depths up to 0.41 mm (0.016 in.). Evaluation by SEM confirmed the difference in structure was associated with a lack of formation of coarse gamma prime structure in the matrix. Microhardness and miniature tensile test results indicated lower strength consistent with the absence of the coarse gamma prime constituent. The blade vendor found that the lot of hot isostatically pressed (HIP) blade castings had been exposed to an improper atmosphere during the HIP process, resulting in the weakened structure. Because subsequent failures were found in blades that did not come from the suspect HIP lot, the scope of the problem was considered generic, and the conclusion was that the primary failure mechanism was LCF. Material imperfections were a secondary deficiency that had the effect of causing the blades from the bad HIP lot to crack first.

When a steam turbine was put out of service, cracks were noticed on many of the blades in the low pressure section round the stabilization bolts and perpendicular to the blade axis. The blades were made from chrome alloy steel X20-Cr13 (Material No. 1.402). When the bolts were brazed into the blades inadmissible localized overheating of the steel must have occurred, which resulted in transformation stresses and hence reduced deformability. The cracks arose as a consequence of careless brazing. Whether the cracks should be considered as stress cracks over their entire extent or partially as fatigue cracks produced by vibration in the operation of the turbine as a result of steplike growing of microcracks could not be deduced from the fracture surfaces. Microfractography showed that the cracks developed in stages.

Three blades from 45,000 kW, 3,000 rpm turbine were received for examination, comprising the root of blade 28, blade 89 showing a crack in one of the root teeth, and blade 106 which was free from defects. Microscopic examination of the blade material showed it to be a ferritic stainless steel of the type commonly used for turbine blades. A number of non-metallic inclusions were present which had been drawn into threads in rolling; these appeared to consist largely of duplex silicates. The failure of blade 28 was the result of the development of a creeping crack. Magnetic crack examination of blade 89 revealed a crack in a tooth in an identical position to the start of the crack in blade 28 but on the opposite, i.e., steam inlet, side of the blade. Similar examination of blade 106 did not reveal any cracks. Cracking was associated with unsatisfactory bedding of the blade teeth on the faces of the wheel grooves. It was concluded that the blade failures were due primarily to over-loading of the individual blade teeth due to incorrect fitting in the wheel. Vibration was an important contributory factor, as it resulted in the imposition of fluctuating stresses on the overloaded teeth. Non-metallic inclusions in the blade material playing a minor part.

The assignment of financial liability for turbine blade failures in steam turbines rests on the ability to determine the damage mechanism or mechanisms responsible for the failure. A discussion is presented outlining various items to look for in a post-turbine blade failure investigation. The discussion centers around the question of how to determine whether the failure was a fatigue induced failure, occurring in accordance with normal life cycle estimates, or whether outside influences could have initiated or hastened the failure.