Correlation of localised thermal shock to isothermal biaxial crack growth rates in thin plates of AISI 316 stainless steel.

The work presented covers numerical analyses of edge-cracked thin plates of austenitic stainless steel, AISI 316, loaded with transient thermal downshock. Further to this isothermal centre-cracked plates loaded under isothermal biaxial conditions are modelled. Crack growth rates are then applied to the models to determine a correlation between the two geometries. The work presented demonstrates that a correlation exists between thermal downshock and biaxial isothermal crack growth rates. For such a correlation to be determined, the thermal shock and isothermal biaxial loading must be studied with an elastic-plastic material response. It is also determined that biaxiality in thermal shock is a function of the shock localisation due to localised contraction of the specimen in the direction of heat flow, making it necessary to evaluate thermal shock stresses in a two-dimensional sense. The relationship between biaxiality and localisation is found to be non-linear. At shocked regions greater than 25% of the available area maximum stresses remain largely unchanged. Below 25% of the available area maximum stresses drop significantly, by up to 50%. Finally a method is proposed whereby thermal shock crack growth rates can be estimated by the bounding conditions of isothermal biaxial loading. This method uses an estimation of the heat transfer coefficient determined through a thorough analysis of a broad range of thermal shock cycles for AISI 316. Using the defining temperatures and time period of the cycle the h value can be estimated to within 3%. Once known the h value is fed into a simple log function describing maximum thermal stress, which can then be converted for any localisation present. It is also found that the correlating isothermal peak load is approximately equal to that of the thermal stress calculated. From this a modified Paris law is proposed to predict two sets of crack growth rates; one at equibiaxial conditions and the second at a biaxiality determined from the thermal shock load. This is shown to be a minimum of 0.35 at maximum localisation. This will calculate the upper, and lower, bounding limits of the thermal shock crack growth rates, providing a good estimation of the thermal shock crack growth rates.