Abstract
Investigations were performed on a cracked turbine disc from the Cooper Nuclear Power Station, and on two failed turbine discs (governor and generator ends) from the Yankee-Rowe Nuclear Power Station. Cooper is a boiling water reactor (BWR) which went into commercial operation in July 1974, and Yankee-Rowe is a pressurized water reactor (PWR) which went into commercial operation in June 1961. Cracks were identified in the bore of the Cooper disc after 41,913 hours of operation, and the disc removed for repair. At Yankee-Rowe two discs failed after 100,000 hours of operation. Samples of the Cooper disc and both Yankee-Rowe discs (one from the governor and one from the generator end of the LP turbine) were sent to Brookhaven National Laboratory (BNL) for failure analysis.
The fracture face on the Cooper disc was observed to have both intergranular and transgranular cracks. The governor end disc at Yankee-Rowe showed no indication of intergranular attack, and failed as a result of fast fracture, probably following impact by one or more of the fragments produced by failure of the generator disc. The generator end disc had numerous intergranular axial cracks in the bore area. MoS2 was observed on the bore of all three discs. The Yankee-Rowe discs were subjected to SEM/EDAX, uniaxial tension tests, hardness testing, Charpy "V" notch testing, and environmental notched tensile tests. The notched tensile tests of the generator discs performed with MoS2 had the effect of reducing the notch tensile strength of the disc by a factor of 3.5. This result was consistent with a similar test conducted with H2S at ambient temperature and pressure. On the governor disc, the H2S reduced the notched tensile strength by only a factor of 2.1, suggesting a reason for the absence of intergranular cracks on this disc. Laboratory data indicate that MoS2 can hydrolyze in a steam environment to form H2S, causing stress corrosion cracks to initiate. The results of this investigation support a model that the cracks initiated at startup of the turbine, probably from H2S produced by hydrolysis of MoS2, and grew at a rate consistent with the published British data for propagation of cracks in pure steam.