Case Study: Metallurgical Failure Analysis of Cracked Turbine Blades

Click here to view a .pdf of this Case Study including additional tables, images & charts.

Case Study: Metallurgical Failure Analysis of Cracked Turbine Blades

INTRODUCTIONBLADES

This failure analysis case study of cracked turbine blades is presented in a condensed version of MAI’s client report format. The report and analysis from which this case study was taken evaluated multiple blades and significantly more data than is presented here. Along with the analytical data, our reports include specific identification of how and why the failure occurred and practical options to prevent similar failures. These answers reduce our clients’ costs and liability exposure, and enhance the quality of their products. Images and data tables are located after the text.

BACKGROUND

Blades from a 660 megawatt steam turbine from an electric power utility were submitted to determine the mode and contributing factors to cracks that formed in service. The turbine operates at a main steam pressure of 2400 psig and a temperature of 1000°F. This unit was reportedly exposed to a contaminated water chemistry event in 1985, however, no cracks were found during an in-situ inspection in 1993. A recent inspection, however, revealed numerous cracks in the blade roots at the attachment to the hub. The operating temperature at the crack locations is approximately 250°F, which is slightly above the steam saturation temperature. The unit has reportedly been cycled through approximately 300 start/stops in its service history.

VISUAL EXAMINATION

Visual examination is often taken for granted. That’s a mistake. A thorough and thoughtful visual examination lays the foundation for an effective failure analysis, identifying critical features and aspects of the failed component.

Typical cracks exhibited by the blades were opened, which revealed dark prior cracks as shown in Figures 1. Examination of these opened prior cracks by optical stereomicroscopy revealed granular macroscopic features, but no evident defects that could have contributed to the failures. The laboratory generated fractures produced during opening of these prior cracks exhibit fibrous macroscopic features that are consistent with ductile single cycle overloads. Corrosion pits are present on the blade surfaces adjacent to these opened cracks.
MECHANICAL PROPERTIES

Tensile testing of bars machined parallel to the lengths of the blades indicates properties that are in conformance with Class 7 material per ASTM A470-65, “Standard Specification for Vacuum-Treated Carbon and Alloy Steel Forgings for Turbine Rotors and Shaft”. No abnormalities are present in the tensile properties of the blades that could have contributed to the cracking at the roots.

Charpy V-notch impact samples were also machined parallel to the length of the blades with the notches in the radial direction. The results of these impact tests are shown in Table 1. The impact properties are in conformance with the 40 ft.-lbs. minimum room temperature impact energy and the +50 °F maximum FATT specified in ASTM A470 for Class 7 material. No abnormalities are present in the impact properties of these two samples that could have contributed to the cracking at the dovetail hooks.

Chemical analysis performed prior to this investigation confirmed that the blades are made from the specified material.

SCANNING ELECTRON MICROSCOPY

The opened crack shown in Figure x was examined using a scanning electron microscope (SEM) equipped with an Energy Dispersive Spectrometer (EDS). EDS spectra are quantified, however, because EDS is a surface analysis technique, oxidation, contaminants and non-homogeneous surface composition may result in variations from bulk chemical analysis results.
EDS analysis of deposits in the dark discoloration on the opened crack revealed detectable amounts of sulfur, chlorine, chromium, manganese, iron, and nickel as shown in Figure 2. Light element EDS of these deposits revealed substantial oxygen and small amounts of carbon and potassium are also present as shown in Figure 3. The chromium, manganese, iron, and nickel are present in the rotor material. The remaining elements are present in the deposit. The substantial oxygen is consistent with an oxidation or corrosion product. The carbon indicates that some organic material is present, which could be a carbonate from the water. The chlorine and potassium are consistent with chloride salts. The source of the sulfur is not clearly evident from this investigation.

The dark deposits on the opened cracks were removed by ultrasonic cleaning in a laboratory grade detergent solution and the opened crack was examined by SEM. No material discontinuities are present on the opened crack. This crack occurred due to intergranular rupture as shown at higher magnifications in Figures 4. These features are consistent with stress corrosion cracking of an alloy steel.

Stress corrosion occurs when a susceptible material is exposed to a specific environment while subjected to sustained static tensile stress. Carbon and alloy steels are susceptible to stress corrosion cracking if exposed to a caustic environment, nitrates, dry ammonia containing carbon dioxide or oxygen, carbon dioxide and carbonate solutions, sulfuric acid, ferric chloride, and other environments. The sustained static tensile stress is due to the radial stresses imposed on the rotor by the mating rotating blades. The laboratory generated fracture produced during opening of this prior crack occurred due to microvoid coalescence as shown in Figure 5. This failure mode is indicative of a ductile single cycle overload, which is consistent with the method used to open the prior crack. This also indicates that the material is not inherently embrittled, which is consistent with the results of the Charpy V-notch impact testing, and that all four cracks occurred due to stress corrosion from exposure to a caustic environment.

METALLOGRAPHY

Transverse metallographic sections were prepared through unopened portions of the crack that was examined by scanning electron microscopy. A branching crack extends from the radius of the root as shown in Figure 6. The branching cracks are intergranular, which is consistent with the results of scanning electron microscopy. This indicates that the crack occurred due to intergranular stress corrosion cracking. Corrosive pitting is also evident on the surface and a small secondary crack extends from this pitting indicating that they are a contributory factor to the cracking. The microstructure adjacent to this crack consists of coarse grained tempered martensite, indicating that this rotor was through hardened by a quench and temper heat treatment as shown in Figure 7. No microstructural abnormalities are present that could have contributed to the cracking.
SUMMARY AND CONCLUSIONS

This investigation indicates that the tensile properties of the blade is consistent with Class 7 material per ASTM A470-65, “Standard Specification for Vacuum-Treated Carbon and Alloy Steel Forgings for Turbine Rotors and Shafts”. The room temperature Charpy V-notch impact energy and the fracture appearance transition temperature (FATT) are also in conformance with this material class. The microstructure of the blade indicates that it was hardened by quenching and tempering. No abnormalities are present in the tensile or impact properties, or in the microstructure, of this blade that could have contributed to the cracking in the root area.
The crack is intergranular, which is consistent with stress corrosion cracking of an alloy steel. Stress corrosion occurs when a susceptible material is exposed to a specific environment while subjected to sustained static tensile stress. Carbon and alloy steels are susceptible to stress corrosion cracking if exposed to a caustic environment, nitrates, dry ammonia containing carbon dioxide or oxygen, carbon dioxide and carbonate solutions, sulfuric acid, and ferric chloride, and other environments. The reported contaminated water chemistry event that occurred shortly after this unit was placed in service in 1985 may have had a contributory effect on the failures. However, because no cracking was detected for a significant period of time after this event, the concentration of caustic compounds due to possible alternate wetting and drying of the steam is the more probable source of corrosion that contributed to the failure. The substantial oxygen identified by Energy Dispersive Spectroscopy is consistent with a corrosion product, and carbon and some of the oxygen are consistent with carbonates from the steam. The chlorine and potassium are characteristic of chloride salts. The source of the sulfur is not clearly evident from this investigation, but could be associated with chemicals added to the water to control its chemistry.

It is recommended that the water chemistry used in this steam turbine be reviewed to identify any abnormalities that could have contributed to the stress corrosion cracking of the blade.