Abstract
Between April and June of 2002, cracks were discovered in the fir-tree roots of several row 10 low-pressure 410 martensitic stainless steel turbine blades from an operating CANDU station. In total, 9 blades were eventually identified by MPI to have flaw indications near the inlet face between the first and second serrations. Among the population of blades examined fractography revealed cracks propagated by two different mechanisms: fatigue and stress corrosion cracking. In 7 of the 9 blades, the fracture surface morphology confirmed crack propagation by high-cycle fatigue, as evidenced by the beachmarks and ratchet marks produced by multiple initiation sites An analysis of the beachmarks suggested that cracks propagated independently and subsequently coalesced into a unified crack front. No significant pitting or other corrosion was found to accompany these defects, which might suggest a corrosion fatigue mechanism. Likewise, no consistent spatial relationship could be established between the crack path and either prior austenite grain boundaries, MnS stringer inclusions, or other metallurgical anomalies, which indicates their role in crack nucleation was minimal. Although hardness values measured were generally consistent with OEM's specifications, some evidence for over-tempering was observed (ripening of grain boundary precipitates/carbides, etc.). However, the specific role of these factors
More>>
Clark, M A;
[1]
Lehockey, E M;
Thompson, I;
[2]
Massey, R
[1]
- Kinectrics, Inc., Toronto, Ontario (Canada)
- Nuclear Safety Solutions, Toronto, Ontario (Canada)
Citation Formats
Clark, M A, Lehockey, E M, Thompson, I, and Massey, R.
Identification and management of cracking in 410 stainless turbine blade roots.
Canada: N. p.,
2003.
Web.
Clark, M A, Lehockey, E M, Thompson, I, & Massey, R.
Identification and management of cracking in 410 stainless turbine blade roots.
Canada.
Clark, M A, Lehockey, E M, Thompson, I, and Massey, R.
2003.
"Identification and management of cracking in 410 stainless turbine blade roots."
Canada.
@misc{etde_20617881,
title = {Identification and management of cracking in 410 stainless turbine blade roots}
author = {Clark, M A, Lehockey, E M, Thompson, I, and Massey, R}
abstractNote = {Between April and June of 2002, cracks were discovered in the fir-tree roots of several row 10 low-pressure 410 martensitic stainless steel turbine blades from an operating CANDU station. In total, 9 blades were eventually identified by MPI to have flaw indications near the inlet face between the first and second serrations. Among the population of blades examined fractography revealed cracks propagated by two different mechanisms: fatigue and stress corrosion cracking. In 7 of the 9 blades, the fracture surface morphology confirmed crack propagation by high-cycle fatigue, as evidenced by the beachmarks and ratchet marks produced by multiple initiation sites An analysis of the beachmarks suggested that cracks propagated independently and subsequently coalesced into a unified crack front. No significant pitting or other corrosion was found to accompany these defects, which might suggest a corrosion fatigue mechanism. Likewise, no consistent spatial relationship could be established between the crack path and either prior austenite grain boundaries, MnS stringer inclusions, or other metallurgical anomalies, which indicates their role in crack nucleation was minimal. Although hardness values measured were generally consistent with OEM's specifications, some evidence for over-tempering was observed (ripening of grain boundary precipitates/carbides, etc.). However, the specific role of these factors in promoting the fatigue failure could not be conclusively identified. Spacing between beachmarks within cracks among the (7) fatigued blades appeared similar suggesting that these cracks propagated under the influence of a common stress regime. Furthermore, the bulk of crack advance appeared to have occurred primarily at operating speeds given the number of beachmarks present far exceeded that expected to evolve solely from the stress transients generated during start/stop cycles. By correlating the array of major beachmarks with operating history, it was tentatively estimated that cracks had been propagating for between 6 and 10 years (of the total 20-year operating life). This is approximately consistent with general guidelines wherein cracks can be expected to initiate during the last 30% of fatigue life. On this basis, and in the absence of other compelling metallurgical evidence suggesting an alternate root-cause, these blades were believed to have reached end of fatigue life. Similitude in the beachmark patterns among the affected blades suggested stable and consistent growth characteristics. In addition, blades appeared to remain intact despite sustaining cracks that propagated through a large fraction of the root cross-section. The stable propagation characteristics combined with the apparently large 'safety' margins provided the opportunity for conducting regular inspections (phased array and MPI) to ensure reliable operation between turbine outages until replacement options could be developed and implemented. Subsequent inspections of row 10 blades in other identical turbines have not detected any other cracked blades to date. In contrast, two of the blades examined exhibited clear evidence of intergranular cracking along prior austenite grain boundaries consistent with stress corrosion cracking (Figure 3). Hardness on both of these blades approached or exceeded 400 HVN. Susceptibility to stress corrosion cracking increases significantly with hardness. Fracture surfaces contained elevated levels of Mo, 5, Cl, and F. Molybenum and sulfur were traced in part to the (Moly Paul) grease lubricant applied to mate the blade root to the turbine steeple during installation. Other contaminants established from Ion Chromatography included unusually high levels of carbon in the form of organics including formates, acetates, and glycolates as well as nitrates, sulfates, and oxalates. Origins of this contamination include throw from cation resins in the water treatment plant, thermal degradation and oxidation by air ingress of the volatile amine (this case morpholine), and ingress of cooling water by way of condenser leaks. All these materials and by-products can ultimately accumulate in the turbine. The latter example demonstrates the importance of considering the implications of excursions in secondary side chemistry, condenser leaks, or other chronic mechanical and chemical factors that impact steam chemistry to ongoing turbine maintenance. This is particularly relevant in turbines containing older blades fabricated to specifications that permit wider than desirable variations in hardness which may render them susceptible to stress corrosion cracking in aggressive environments. A comprehensive inspection program based on in-situ hardness testing was instituted to identify and discard these high hardness blades. (author)}
place = {Canada}
year = {2003}
month = {Jul}
}
title = {Identification and management of cracking in 410 stainless turbine blade roots}
author = {Clark, M A, Lehockey, E M, Thompson, I, and Massey, R}
abstractNote = {Between April and June of 2002, cracks were discovered in the fir-tree roots of several row 10 low-pressure 410 martensitic stainless steel turbine blades from an operating CANDU station. In total, 9 blades were eventually identified by MPI to have flaw indications near the inlet face between the first and second serrations. Among the population of blades examined fractography revealed cracks propagated by two different mechanisms: fatigue and stress corrosion cracking. In 7 of the 9 blades, the fracture surface morphology confirmed crack propagation by high-cycle fatigue, as evidenced by the beachmarks and ratchet marks produced by multiple initiation sites An analysis of the beachmarks suggested that cracks propagated independently and subsequently coalesced into a unified crack front. No significant pitting or other corrosion was found to accompany these defects, which might suggest a corrosion fatigue mechanism. Likewise, no consistent spatial relationship could be established between the crack path and either prior austenite grain boundaries, MnS stringer inclusions, or other metallurgical anomalies, which indicates their role in crack nucleation was minimal. Although hardness values measured were generally consistent with OEM's specifications, some evidence for over-tempering was observed (ripening of grain boundary precipitates/carbides, etc.). However, the specific role of these factors in promoting the fatigue failure could not be conclusively identified. Spacing between beachmarks within cracks among the (7) fatigued blades appeared similar suggesting that these cracks propagated under the influence of a common stress regime. Furthermore, the bulk of crack advance appeared to have occurred primarily at operating speeds given the number of beachmarks present far exceeded that expected to evolve solely from the stress transients generated during start/stop cycles. By correlating the array of major beachmarks with operating history, it was tentatively estimated that cracks had been propagating for between 6 and 10 years (of the total 20-year operating life). This is approximately consistent with general guidelines wherein cracks can be expected to initiate during the last 30% of fatigue life. On this basis, and in the absence of other compelling metallurgical evidence suggesting an alternate root-cause, these blades were believed to have reached end of fatigue life. Similitude in the beachmark patterns among the affected blades suggested stable and consistent growth characteristics. In addition, blades appeared to remain intact despite sustaining cracks that propagated through a large fraction of the root cross-section. The stable propagation characteristics combined with the apparently large 'safety' margins provided the opportunity for conducting regular inspections (phased array and MPI) to ensure reliable operation between turbine outages until replacement options could be developed and implemented. Subsequent inspections of row 10 blades in other identical turbines have not detected any other cracked blades to date. In contrast, two of the blades examined exhibited clear evidence of intergranular cracking along prior austenite grain boundaries consistent with stress corrosion cracking (Figure 3). Hardness on both of these blades approached or exceeded 400 HVN. Susceptibility to stress corrosion cracking increases significantly with hardness. Fracture surfaces contained elevated levels of Mo, 5, Cl, and F. Molybenum and sulfur were traced in part to the (Moly Paul) grease lubricant applied to mate the blade root to the turbine steeple during installation. Other contaminants established from Ion Chromatography included unusually high levels of carbon in the form of organics including formates, acetates, and glycolates as well as nitrates, sulfates, and oxalates. Origins of this contamination include throw from cation resins in the water treatment plant, thermal degradation and oxidation by air ingress of the volatile amine (this case morpholine), and ingress of cooling water by way of condenser leaks. All these materials and by-products can ultimately accumulate in the turbine. The latter example demonstrates the importance of considering the implications of excursions in secondary side chemistry, condenser leaks, or other chronic mechanical and chemical factors that impact steam chemistry to ongoing turbine maintenance. This is particularly relevant in turbines containing older blades fabricated to specifications that permit wider than desirable variations in hardness which may render them susceptible to stress corrosion cracking in aggressive environments. A comprehensive inspection program based on in-situ hardness testing was instituted to identify and discard these high hardness blades. (author)}
place = {Canada}
year = {2003}
month = {Jul}
}