“Potential Development of Improved Fatigue Design Methods” G. H. Weidenhamer and W. J. O’Donnell, presented at ASME Winter Annual Meeting, New Orleans, Louisiana, December 10-14, 1984, ASME Paper 84-WA/PVP-6.
Keywords: ASME Code; fatigue design criteria; safety margins; crack initiation; crack propogation; manufacturing defects; fatigue of weldments
The technology base for the current ASME Code fatigue design criteria is described. Verified crack initiation and failure data on full size vessels is compared with both the laboratory specimen failure data and the design curves. This comparison indicates that the existing design curves do provide the intended safety margins. However, questions have been raised about the propagation of acceptable manufacturing defects. Published results based on crack propagation technology indicate that such manufacturing defects can propagate completely thru-the-wall within the fatigue design limits of Section III of the Code. The potential for developing improved fatigue design criteria by considering the crack initiation, propagation, and instability phases of fatigue failure is established. The importance, complexity and potential for extending these methods to include the fatigue of weldments are explored and the direction of technology developments needed to improve Codes and Standards is suggested.
CURRENT ASME CODE FATIGUE DESIGN BASIS
Current ASME Code fatigue design criteria for nuclear components (Ref. 1) are based on 30-year old technology. Extensive test results exist showing that cracks propagate much faster in reactor water environments than in air. The S-N fatigue design curves in the Code, however, are based entirely on fatigue tests conducted in air. In the low-cycle fatigue regime, 90% of the fatigue life involves crack propagation. Test results have also shown that cracks initiated in the low-cycle regime will propagate below the endurance limit of the unnotched specimens which are used as the sole basis for the fatigue design curves in Section III of the Code.
These considerations raise the question whether ornot the current fatigue design criteria and curves are adequate. Present fatigue design evaluation methods in Section III of the ASME Code were developed by applying a factor of 2 on the “stress” or 20 on the cycles-to-failure from the “best fit” curve through failure data obtained on polished unnotched specimens tested in air_ The “stress” amplitude is actually half the total strain range multiplied by the modulus of elasticity to obtain a fictitious stress useful with elastic analyses. The Tresca failure criterion is used for general stress conditions.
FULL SIZE VESSEL CRACKING VS. FATIGUE DESIGN CURVES
An experimental study (Refs. 2, 3, 4) sponsored by the Pressure Vessel Research Committee and the Atomic Energy Commission was conducted to determine the low-cycle fatigue characteristics of full size pressure vessels incorporating a variety of nozzle configurations of interest to the reactor designer and pressure vessel industry at large. These investigations were carried out over 20 years ago at Southwest Research Institute. Thirty-six inch I.D. vessels with 2-inch nominal nozzle wall thickness constructed of ASTM A-201 Grade A and other low carbon steels and ASTM A-302 Grade B and other low alloy high strength steels were cyclically tested.
A 2-inch thick plate used was the thinnest requiring fine-grain practice. Consequently, these vessels may be regarded as “full-size”, and the data obtained should be representative of full-size vessels. The local stress conditions were carefully measured using brittle coatings and nearly 900 strain gages. Peak stresses occurred at the inside corners of the nozzles and fatigue cracks developed at these locations. Strain ranges continued to be measured as the vessels were pressure-cycled and essentially stabilized after 10 cycles. Crack initiation and crack growth were observed.