An LNG tanker experienced a fracture of the solid tail shaft, which is a section of the main drive shaft. The tail shaft was made of a forged low-carbon steel. In spite of two ultrasonic inspections, a large defect the size of a football in the center of the shaft was missed. During heat treating following forging, it was surmised that the defect led to the propagation of an internal brittle crack, or clink. A fatigue crack propagated from this origin to the outer surface of the shaft after about a year of service. Finally a last ligament of a few square inches held the shaft together and broke, leading to the separation of the shaft. The cause of failure was fatigue crack initiation and crack growth under reverse bending cyclic stresses. There was no indication that misalignment existed because there was no indication of fretting at the bolt holes in the flange at the end of the shaft. In the case of this shaft, a solution would have been to machine the core of the shaft to remove the brittle material or to use a tubular shaft.

An ASTM A193-83a grade B7 (AISI 4140) steel turbine impeller shaft fractured after 2 months of service. Failure had initiated at three separate points around the periphery of the shaft, each associated with one of three keyways. SEM fractography, metallography, and chemical analysis indicated that the mechanism of fracture initiation was torsional fatigue. Intermittent deceleration and acceleration resulting from power surges during operation of the turbine caused torsional vibration and was considered the most probable source of the required cyclic stress. Final failure took place by torsional shear.

The splined shaft (1040 steel, heat treated to a hardness of 44 to 46 HRC and a tensile strength of approximately 1448 MPa, or 210 ksi) from a front-end loader used in a salt-handling area broke after being in service approximately two weeks while operating at temperatures near -18 deg C (0 deg F). During the summer, similar shafts had a service life of 5 to eight months. Examination of the fracture surface showed brittle fatigue cracks, and visual examination of the splines disclosed heavy chatter marks at the root of the spline, with burrs and tears at the fillet area. Evidence found supports the conclusion that the shaft failed as the result of stress in the sharp fillets and rough surfaces at the root of the splines. Cold weather failure occurred sooner than in hot weather because ductile-to-brittle transition temperature of the 1040 steel shaft was too high. Recommendations include redesign of the fillet radius to a minimum of 1.6 mm (0.06 in.) and a maximum surface finish in the spline area of 0.8 microns. Material for the shafts should be modified to a nickel alloy steel, heat treated to a hardness of 28 to 32 HRC before machining.

A steel eyebolt which attached a rear lift strut to the right wing of a helicopter failed by fatigue. As a contributing factor, thread cutting produced sharp notches at thread roots, reducing fatigue life. Also, design fatigue life may have been exceeded as the part was in use about 10,000 h. Cumulative damage resulting from a previous accident could have occurred too. Because of this accident, inspectors were instructed to examine threaded zones of eyebolts by magnetic particle inspection after every 100 h in service. A maraging steel drive shaft of a helicopter also failed because of corrosion (pits), and continuous abnormal misalignment as well. Corrosion probably developed from moisture and water droplets on shaft diaphragm profiles. Improved diaphragm pack seals and coatings made by an electron-coat process (such as a Sermetal finish) are now used in new shafts.

In a shaft subjected to reversed torsional stresses, failure resulted from the gradual development of fatigue cracks from opposite sides of the shaft. These broke out from origins located adjacent to the fillets at the start of the square section. The remaining uncracked material which fractured at the time of the mishap was in the form of a narrow strip, situated slightly to one side of the center of the shaft. The material was a mild steel in the normalized or annealed condition, having a carbon content of approximately 0.3%. The cracking was characteristic of that resulting from torsional fatigue. Because it occurred on two different planes at 45 deg to the axis of the shaft it was due to reversals of torsional stress rather than fluctuations of unidirectional torque. Following this failure, the shafts of six other similar cranes were tested ultrasonically. Cracks to varying degree were found in all the shafts. Timely replacement was possible and the likelihood of serious accidents removed.

Routine magnetic-particle inspection revealed crack indications in a number of shafts produced from hot-rolled 4130 steel bar. A pronounced indication of this size is cause for rejection if the defect is not eliminated during subsequent machining. A microstructural analysis of the shaft cross section revealed that the crack was approximately 0.5 mm (0.020 in.) deep and oriented in a radial direction. Furthermore, no stringer-type nonmetallic inclusions were observed in the vicinity of the flaw, which did not display the intergranular characteristics of a quench crack. The defect did, however, contain substantial amounts of oxide, which evidently resulted from the hot-working operation. This evidence supports the conclusion that the appearance of this discontinuity, with the long axis parallel to the working direction and radial orientation with regard to depth, strongly suggests a seam produced during rolling. Use of components with surface-defect indications as small as 0.5 mm (0.02 in.) can be risky in certain circumstances. Depending on the orientation of the flaw with respect to applied loads, the nature of the applied forces (for example, cyclic), and the operating environment, such a surface flaw can become the initiating site for a fatigue crack or a corrosion-related failure.

The broken end of a shaft from a centrifugal pump had a smooth fracture surface characteristic of failure from fatigue. Failure occurred in the plane of the keyway end and followed a slightly helical path, indicating that combined bending and torsional stresses were responsible. The material was a Cr-Mo-Ni alloy steel of the En 19 type in the hardened and tempered condition and of satisfactory quality. The assembly also included a copper sleeve attached by a circumferential braze behind the plane of fracture. The cracks were examined for the presence of copper, thinking that penetration by molten copper may have played a role, but no evidence was seen. An absence of chromium plating at the region of the heat-affected zone was also observed but could not be explained. Unfortunately, the end portion of the shaft was not available for examination.

Shaft fracture of a 10 hp squirrel cage motor took place at the driving end just outside the roller bearing and not at an abrupt change of section behind the bearing where it might be expected to occur. A portion of shaft to the right of the fracture was deeply grooved. About a year prior to failure the inner race of the roller bearing became slack on the shaft and the seating was built up by the metal-spray process. The shaft was machined to form a rough thread to provide the requisite mechanical key for the sprayed-on metal. Part of this sprayed-on layer became detached after the fatigue failure occurred. The quality of the welding was poor. Slag inclusions were present adjacent to the sides of the keyway, which had been re-cut shorter than the original one after the welding repair. Failure at the unusual location was caused by the presence of the weld deposit.

A steering knuckle used on an earthmover failed in service. The component fractured into a flange portion and a shaft portion. The flange was 27.9 cm (11 in.) in diam around which there were 12 evenly spaced 16 mm diam bolt holes. The shaft was hollow with a 10.5 cm (4 in.) OD and a wall thickness of 17 mm. The steering knuckle was made of 4340 steel and heat treated to a hardness of about 415 HRB (yield strength of about 1069 MPa, or 155 ksi). The vehicle had been involved in a field accident six months before the steering knuckle failed. Several components, including portions of the frame, had been damaged and replaced, but there was no observed damage to the steering. Analysis supported the conclusion that the fracture was the result of the prior accident, the most likely explanation being that the shaft was bent and that continued use caused a crack to initiate and propagate to fracture. No evidence of a defective design, improper microstructure, high inclusion count, or other stress-raising condition was observed. No recommendations were made.

A shaft which carried the diverter sheave wheel of an electric goods lift failed, resulting in the cage failing to the bottom of the well. Failure had taken place at a reduction in diam at which no filet radius existed. Metallurgical examination did not disclose any abnormal features. The material was a mild steel in the normalized condition. The appearance of the fracture indicated failure was due to bending stresses. The absence of any fillet radius at the reduction in diam provided a region of stress concentration from which fatigue cracks developed.

An amusement ride failed when a component in the ride parted, permitting it to fly apart. The ride consisted of a central shaft supporting a spider of three arms, each of which was equipped with an AISI 1040 steel secondary shaft about which a circular platform rotated. The main shaft rotated at about 12 rpm and the platforms at a speed of 20 rpm. The accident occurred when one of the secondary shafts on the amusement ride broke. The point of fracture was adjacent to a weld that attached the shaft to a 16 mm thick plate, which in turn bore the platform support arms. Investigation (visual inspection, 0.4x magnification, and stress analysis) supported the conclusion that a likely cause for the fatigue failure was the combination of residual stresses generated in welding and centrifugal service stresses from operation that were accentuated by areas of stress concentration at the undercut locations. Without the excessive residual stress, the shaft dimensions appeared ample for the service load. Recommendations included applying the fillet weld with more care to avoid undercutting. The residual stresses could be minimized by pre-weld and post-weld heat application.

A 4340 steel shaft, the driving member of a large rotor subject to cyclic loading and frequent overloads, broke after three weeks of operation. The driving shaft contained a shear groove at which the shaft should break if a sudden high overload occurred, thus preventing damage to an expensive gear mechanism. The rotor was subjected to severe chatter, which was an abnormal condition resulting from a series of continuous small overloads occurring at a frequency of around three per second. Investigation (visual inspection, hardness testing, and hot acid etch images) supported the conclusion that the basic failure mechanism was fracture by torsional fatigue, which started at numerous surface shear cracks, both longitudinal and transverse, that developed in the periphery of the root of the shear groove. These shear cracks resulted from high peak loads caused by chatter. The shear groove in the shaft had performed its function, but at a lower overload level than intended. Recommendations included increasing the fatigue strength of the shaft by shot peening the shear groove to minimize chatter.

Plunger shafts machined from 4150 steel bar stock were involved in a series of fatigue failures. The fractures consistently occurred at two locations on the shafts: the shaft fillet and either side of a machined notch. The material specification for the shafts required 41xx series steel with a carbon content of 0.38 to 0.53%, a hardness of 35 to 40 HRC for the shaft, and a hardness of 50 to 55 HRC for the notch (which was case hardened). Analysis (visual inspection, chemical analysis, hardness testing, and magnetic particle inspection) supported the conclusions that all the fractures were fatigue-induced failures due to sharp radii in the fillets. The stress-concentrating effects of the fillets caused fatigue cracks to initiate and grow under cyclic loading until the crack depth was critical, causing the shaft to fail and rendering the assembly inoperative. Recommendations included increasing the radii of the notch and shaft fillets. If fatigue cracking had continued to be a problem with this component, shot peening of the subject radii would be appropriate. This process produces residual compressive stresses in the surface of the part, thereby retarding initiation of fatigue cracks.

A crane long-travel worm drive shaft was found to be chipped during unpacking after delivery. Chemical analysis showed that the steel (EN36A with a case depth of 1 mm, or 0.04 inch did not meet specifications. Magnetic particle inspection revealed a crack on the side of the shaft opposite the chip. Metallographic examination indicated that the case depth was approximately 2 mm (0.08 in.) and that a repair weld of an earlier chip had been made in the cracked area. The chipping was attributed to excessive case depth and rough handling. It was recommended that the shaft be returned to the manufacturer and a replacement requested.

Field failures, traced to internal cracks that were initiated from gross nonmetallics, were encountered in the upset portion of forged 4118 steel shafts. Ultrasonic inspection was thought to be the best method for detection from the location of these cracks, their orientation, and the size of the shaft. A longitudinal beam was sent in from the end of the shaft. The shaft was observed to have a radially drilled oil hole 9 mm in diam. Since there was a variation in flaw orientation, testing of the shaft was desired from both the long and short end. The rejection level was set at 20% of full screen and was based on the size of flaws observed when the shafts were cut up. The inclusions were considered to be rejectable if the size was larger than 20 mm diam. Similar flaws were observed in larger shafts, but no flaws were observed once the shafts were sectioned. It was interpreted that the flaw signals were false and had happened when a portion of the beam struck the oily surface of the longitudinal oil hole. The problem was solved by removing the oil film from the longitudinal oil hole.

A shaft can crack twice before it fails. A Detroit electric plant had this experience with one in a coal pulverizer. Because the first crack rewelded partially (by friction) in service, the pulverizer remained serviceable until the second crack developed.

A large fan assembly deformed and broke at multiple locations. The user wanted to know whether the bearing pillow block fracture caused the fan blade assembly to crack, or whether a fan blade assembly fracture caused the pillow block to crack. Close inspection of the entire length of the crack showed the crack probably grew quite a while before it was large enough to cause the final catastrophic event. No evidence of fatigue cracks was visible on the broken pillow blocks. In the absence of some other contradictory information, the usual conclusion would be to presume that the fatigue crack predated the single overload crack.

A hollow, splined alloy steel aircraft shaft (machined from an AMS 6415 steel forging – approximately the same composition as 4340 steel – then quenched and tempered to a hardness of 44.5 to 49 HRC) cracked in service after more than 10,000 h of flight time. The inner surface of the hollow shaft was exposed to hydraulic oil at temperatures of 0 to 80 deg C (30 to 180 deg F). Analysis (visual inspection, 15-30x low magnification examination, 4x light fractograph, chemical analysis, hardness testing) supported the conclusions that the shaft cracked in a region subjected to severe static radial, cyclic torsional, and cyclic bending loads. Cracking originated at corrosion pits on the smoothly finished surface and propagated as multiple small corrosion-fatigue cracks from separate nuclei. The originally noncorrosive environment (hydraulic oil) became corrosive in service because of the introduction of water into the oil. Recommendations included taking additional precautions in operation and maintenance to prevent the use of oil containing any water through filling spouts or air vents. Also, polishing to remove pitting corrosion (but staying within specified dimensional tolerances) was recommended as a standard maintenance procedure for shafts with long service lives.

A type 410 stainless steel circulating water pump shaft used in a fossil power steam generation plant failed after more than 7 years of service. Visual examination showed the fracture surface to be coated with a thick, spalling, rust-colored scale, along with evidence of pitting. Samples for SEM fractography, EDS analysis, and metallography were taken at the crack initiation site. Hardness testing produced a value of approximately 27 HRC. The examinations clearly established that the shaft failed by fatigue. The fatigue crack originated at a localized region on the outside surface where pitting and intergranular cracking had occurred. The localized nature of the initial damage indicated that a corrosive medium had concentrated on the surface, probably due to a leaky seal. Reduction of hardness to 22 HRC or lower and inspection of seals were recommended to prevent future failures.

A 3 in. diam shaft was found to have suffered excessive wear on one of the journals and was built up by welding. While it was in the lathe prior to turning down the built-up region, a crack was discovered in the root of the oil-seal groove and subsequently the end of the shaft was broken off with hammer blows. The fracture surface was duplex in nature, there being an annular region surrounding a central zone, which suggests that the fracture developed in two stages. Microscopic examination confirmed that the fracture was of the brittle type. The shaft material showed a microstructure typical of a medium-carbon steel (carbon approximately 0.4%) in the normalized condition, a material not weldable by ordinary methods. It was concluded that the post-welding crack arose primarily from the thermal contraction which developed in the weld metal on cooling. It is probable that if the built-up zone had extended beyond the oil seal groove, failure in the manner would not have occurred. Experience indicated however, that failure from fatigue cracking would still have been likely to occur.