Nondestructive testing (NDT), also known as nondestructive evaluation (NDE), includes various techniques to characterize materials without damage. This article focuses on the typical NDE techniques that may be considered when conducting a failure investigation. The article begins with discussion about the concept of the probability of detection (POD), on which the statistical reliability of crack detection is based. The coverage includes the various methods of surface inspection, including visual-examination tools, scanning technology in dimensional metrology, and the common methods of detecting surface discontinuities by magnetic-particle inspection, liquid penetrant inspection, and eddy-current testing. The major NDE methods for internal (volumetric) inspection in failure analysis also are described.

The master connecting rod of a reciprocating aircraft engine revealed cracks during routine inspection. The rods were forged from 4337 (AMS 6412) steel and heat treated to a specified hardness of 36 to 40 HRC. H-shaped cracks in the wall between the knuckle-pin flanges were revealed by visual examination. The cracks were originated as circumferential cracks and then propagated transversely into the bearing-bore wall. No inclusions in the master rod were detected by magnetic-particle and x-ray inspection. Three large inclusions lying approximately parallel to the grain direction and fatigue beach marks around two of the inclusions were revealed by macroscopic examination of the fracture surface. Large nonmetallic inclusions that consisted of heavy concentrations of aluminum oxide (Al2O3) were revealed by microscopic examination of a section through the fracture origin. The forging vendors were notified about the excess size of the nonmetallic inclusions in the master connecting rods and a nondestructive-testing procedure for detection of large nonmetallic inclusions was established.

A hoist lift hose on a loader failed catastrophically. The hoses were a 100R13 type (as classified in AS3791-1991) with 50.8 mm nominal internal diameter. They consisted of six alternating spirals of heavy wire around a synthetic rubber inner tube with a synthetic rubber outer sheath. Failure of the lift hose was approximately 50 to 100 mm away from the "upper" end of the hose, with the straight coupling that attaches to the hydraulic system. The return hose was in much better condition, with no apparent deformation and only small areas of mechanical damage to the outer sheath. There were two modes of failure of the wire: tensile and corrosion related. The predominant corrosion mechanism appeared to be crevice corrosion related, with the corrosion being driven by the retention of water by the cover material around the wire strands. In this case study (and in most wire-reinforced hydraulic hoses), the wire reinforcing strands were a medium-carbon steel in the cold drawn condition. Radiographic nondestructive testing (NDT) was recommended to determine when a hydraulic hose should be replaced.

Two leaks were discovered at a sulfur recovery unit in a refinery. The leaks were at pipe-to-elbow welds in a 152 mm (6 in.) (NPS 6) diam line, operating in lean amine service at 50 deg C (120 deg F) and 2.9 MPa (425 psig). Thickness measurements indicated negligible loss of metal, and the leaks were clamped. A year later, 15 additional leaks were discovered, again at pipe-to-elbow welds in lean amine lines. Further nondestructive testing located other cracks, giving a total of 35. These lines had been in service for approximately eight years. Investigation (visual inspection, hardness testing, and micrographic cross-sections) supported the conclusion that the failure was caused by lean amine SCC. It was considered unlikely that these pipe welds had received such a postweld heat treatment, although it is industry practice to postweld stress relieve piping and pressure vessels in lean amine service if the temperature is expected to be above 95 deg C (200 deg F). Recommendations included inspecting all welds using shear wave ultrasonic testing and postweld heat treating all welds in lean amine service.

Six cracked A36 steel gusset plates that formed part of the roof trusses of a large manufacturing facility were discovered during a routine final inspection of a new building construction. Two different-size plates from different locations in the building were removed and tested. It was determined that the gusset plates failed in the heat-affected zone via an intergranular microcracking mode due to hydrogen-assisted underbead and toe-weld cracking. Proper nondestructive testing techniques for magnetic particle and radiographic inspection of the plate-weld gusset areas were recommended.

The source of cracking in the circumferential weld seam in a JIS-SM50B carbon-manganese steel pipe used in a CO2 absorber was investigated, the absorber had been in service for 18 years. The seam had been weld-repaired twice, and the repair welds had been locally stress relieved. Longitudinal seams in the same vessel, which had been stress relieved in a furnace, showed no tendency toward cracking. The solution passing through the vessel contained CO2-CO-H20, KHCO, and Cl− ions. Nondestructive testing revealed that the cracks originated in the heat-affected zone and propagated into the base metal and weld. Severe branching of the cracks characteristic of stress-corrosion cracking was observed. Microexamination revealed that crack propagation was transgranular further supporting the possibility of stress-corrosion cracking. Simulation tests carried out in the vessel confirmed this mode of cracking. It was recommended that weld seams be furnace heat treated at a temperature of 600 to 640 deg C (1110 to 1180 deg F) for a minimum of 1 h per inch of section thickness.

A deformed steel tube was received for failure analysis after buckling during a heat-treat operation. The tube was subjected to various metallurgical tests as well as nondestructive testing to confirm the presence of residual stresses. The microstructure of the tube was found to be homogenous and had no banded structure. However, x-ray diffraction analysis confirmed the presence of up to 6% retained austenite which likely caused the tube to buckle during the 910 °C heat treating procedure.

Two complete aircraft undercarriage-leg 2014 aluminum alloy forgings and a number of sectional ends that exhibited cracks during nondestructive testing were examined to determine the extent of damage and the type of cracking. Cracks were primarily confined to the diaphragm and adjoining wall between the steel sleeve and the steel diaphragm washer. Metallographic analysis and accelerated corrosion tests showed that the cracks had originated as stress-corrosion failures.

A newly installed pipeline leaked during cleaning prior to hydrotest at a pressure of approximately 400 psig. The intended hydrotest pressure was 750 psig. The pipeline was constructed from spiral-welded API 5L-X65 HSLA steel and was intended for seawater injection. Analysis included nondestructive testing, metallography, and scanning electron microscopy. Based on the results, the failure was attributed to transit fatigue, caused during highway transportation. Cracks along the toes of the weld from both the outside and inside surfaces, the transgranular nature of cracking, and the presence of fatigue striations all supported transit fatigue as the damage mechanism.

A back wall riser tube in a high pressure boiler failed, interrupting operations in a cogeneration plant. The failure occurred in a tube facing the furnace, causing eight ruptured openings over a 1.8 m section. The investigation consisted of an on-site visual inspection, nondestructive testing, energy dispersive x-ray analysis, and inductively coupled plasma mass spectrometry. The tube was made from SA 210A1 carbon steel that had been compromised by wall thinning and the accumulation of fire and water-side scale deposits. Investigators determined that the tube failed due to prolonged caustic attack that led to ruptures in areas of high stress. The escaping steam eroded the outer surface of the tube causing heavy loss of metal around the rupture points.

Two 25 x 40 mm (1 x 1.5 in.) AISI 4150 hot-rolled steel bars that cracked during heat treatment were examined to determine whether the heat treating procedure had contributed to the failure. Metallographic examination of a cross section taken through the fracture revealed an oxide coating on both sides of the fracture surface. The oxide was also found on the top and bottom sides of the sample. Sawcut sides of the bar did not exhibit the oxide layer The presence of the oxide in the fracture, combined with its absence on all exterior surfaces, indicated that the fracture occurred as a result of an oxide seam in the original material rather than from oxide from heat treating. Nondestructive testing prior to machining and heat treatment was recommended.

An A-470 steel rotor disk was removed from the high-pressure portion of a steam turbine-powered compressor after nondestructive testing revealed cracks in the shoulder of the disk during a scheduled outage. Samples containing cracks were examined using various methods. Multiple cracks, primarily intergranular were found on the inlet and outlet faces along prior-austenite grain boundaries. The cracks initiated at the surface and propagated inward. Multiple crack branching was observed. Many of the cracks were filled with iron oxide. X-ray photoelectron spectroscopy indicated the presence of sodium on crack surfaces, which is indicative of NaOH-induced stress-corrosion cracking. Failure was attributed to superheater problems that resulted in caustic carryover from the boiler. Two options for disk repair, installing a shrink-fit disk or applying weld buildup, were recommended. Weld repair was chosen, and the rotor was returned to service; it has performed for more than 1 year without further incident.

This article provides information on life assessment strategies and conceptually illustrates the interplay of nondestructive evaluation (NDE) and fracture mechanics in the damage tolerant approach. It presents information on probability of detection (POD) and probability of false alarm (PFA). The article describes the damage tolerance approach to life management of cyclic-limited engine components and lists the commonly used nondestructive evaluation methods. It concludes with an illustration on the role of NDE, as quantified by POD, in fully probabilistic life management.

Of the many different nondestructive evaluation (NDE) techniques, ultrasonic inspection continues to be the leading nondestructive method for inspecting composite materials, because measurements can be quantitative and the typical defect geometries and orientations lend themselves to detection and characterization. This article focuses on the three common methods for ultrasonic nondestructive inspection of plastics, namely pitch-catch, through-transmission, and pulse-echo, as well as the three basic types of ultrasonic NDE scans: the A-scan, B-scan, and C-scan. The discussion includes the linear and phased array systems that are sometimes used for large-scale inspection tasks to reduce scan times, the various gating and image processing techniques, and how ultrasonic data are interpreted and presented. A brief section on future trends in ultrasonic inspection is presented at the end of the article.

A crack was found in an aircraft main wing spar flange fabricated from 7079-T6 aluminum alloy during a routine nondestructive x-ray inspection after the craft had logged 300 h. Scanning electron microscopy (SEM) revealed an intergranular fracture pattern indicative of stress-corrosion cracking (SCC) and fatigue striations near the crack origin. Visual examination of the crack edge revealed that the installation of the fasteners produced a fit up stress. Further inspection of the opened fracture showed that the crack had been present for some time because a heavy buildup of corrosion products was seen on the fractured surface. Metallographic examination of the flange in the area of fracture initiation showed the presence of end grain exposure, which would promote SCC. Electron optical examination of the fracture clearly showed the flange was cracking by a mixed mode of stress corrosion and fatigue. The cracking was accelerated because of an inadvertent fit up stress during installation. The age of the crack could not be established. However, a reevaluation of prior x-ray inspections in this area would result in some close estimate of the age of the crack. End grain exposure further promoted SCC.

Three wing flap hinge bearings were received by the laboratory for analysis. The bearings were fabricated from chromium-plated type 440C martensitic stainless steel. The intergranular fracture pattern seen in the electron fractographs, coupled with the corrosion pits observed on the inner diam of the bearings, strongly suggested that failure initiated by pitting and progressed by SCC or hydrogen embrittlement from the plating operation. It was recommended that the extent of the flap hinge bearing cracking problem be determined by using nondestructive inspection because it is possible to crack hardened type 440C during the chromium plating process. An inspection for pitting on the bearing inner diam was also recommended. It was suggested that electroless nickel be used as a coating for the entire bearing. A review of the chromium plating and baking sequence was recommended also to ensure that a source of hydrogen is not introduced during the plating operation.

A lever (machined from a casting made of AISI type 410 stainless steel, then surface hardened by nitriding) that was a component of the main fuel-control linkage of an aircraft engine fractured in flight after a service life of less than 50 h. Investigation (radiographic inspection) supported the conclusions that the lever broke at a cold shut extending through approximately 95% of the cross section. The normally applied load constituted an overload of the remainder of the lever. Recommendations included adding magnetic-particle inspection to the inspection procedures for this cast lever.

Cadmium-plated high-strength steel bolts were used to facilitate quick disassembly of a vehicle. One bolt was found fractured across the root of a thread after being torqued in place for one week. The bolts were made of 8735 steel heat treated to a tensile strength of 1241 to 1379 MPa (180 to 200 ksi) with a hardness of 39 to 43 HRC, followed by cadmium plating. The bolt that failed and several that did not were examined. It was found that failure of the bolts was the result of time-dependent hydrogen embrittlement. Had the remaining bolts been torqued to the normal stress levels, all would have failed within two weeks. The bolts were baked, as specified by ASTM B 242, at 205 deg C (400 deg F) for 30 min. No further failures occurred. Baking for 30 min is the minimum baking time; however, baking times up to 24 h are recommended for greater safety.

Two types of chromium-plated hydraulic cylinders failed by cracking on their outer surfaces. In one case, the parts had a history of cracking in the nominally unstressed, as-fabricated condition. In another, cracks were detected after the cylinders were subjected to a pressure impulse test. Both part types were made of 15-5 PH (UNS S15500) precipitation hardening stainless steel. Hydrogen embrittlement cracking was the likely cause of failure for both part types. Cracking of the as-fabricated parts was ultimately prevented by changing the manufacturing procedure to allow for a reheat treatment. For parts that cracked after pressure testing, excessive dimensional changes precluded the inclusion of a reheat treatment as a manufacturing step, and further failure was averted by carefully employing proper machining practices, avoiding abusive machining.

As the result of a leak detected in a plate-formed header at PENELEC'S Shawville Unit No. 3, an extensive failure investigation was initiated to determine the origin of cracking visible along the longitudinal weld seam. Fabricated from SA387-D material and designed for a superheater outlet temperature of 566 deg C, the 11.4 cm thick header had operated for approximately 187,000 h at the time of the failure. Discussion focuses on the results of a metallographic examination of boat samples removed from the longitudinal seam weldment in the vicinity of the failure and at other areas of the header where peak temperatures were believed to have been reached. The long-term mechanical properties of the service-exposed base metal and creep-damaged weld metal were determined by creep testing. Based on the utility's decision to replace the header within one to three years, an isostress overtemperature lead specimen approach was taken, whereby failure of a test specimen in the laboratory would precede failures in the plant. These tests revealed approximately a 2:1 difference in life for the base metal as compared to weld metal.