ACCEPTABILITY OF DYNAMIC FINITE ELEMENT ANALYSES – MATERIAL FAILURE APPROACH - Presentation

IAEA [701] details how compliance with the regulatory tests may be achieved. The GB Competent Authority has always been open minded to new calculation procedures and has encouraged the use of finite element analysis where appropriate. In the last few years there has been a marked increase in the complexity of some analyses, which support applications, with attempts made to model material failure in a conservative manner. We have noticed a range of failure theories and modelling methods that may, or may not be acceptable. This paper attempts to clarify our expectations with regard to modelling failure. In elastic analyses, the main acceptance criterion is to check that the calculated stresses, e.g. Von-Mises or Tresca do not exceed the lower bound yield stress/es for the material/s. If the stresses exceed the yield limit/s, the determining criterion shifts from stress to strain; especially under multi-axial load conditions. Elastic-plastic analysis using finite element approach, solicits, true stress-strain curve for the material/s to be incorporated. The most relevant material property for comparison now is the available engineering uniaxial ductility. It is a global measure of failure, whereas, the local strains in the vicinity of the failure zone could be much higher. For a static analysis, the enumerated strains are available at the end of the analysis, whereas, in a dynamic analysis, the set period for analysis should be “long” enough to ensure that the response has been captured correctly at the end of the analysis. In most of the engineering applications, the state of stress is generally multi-axial, thus affecting the available ductility, which decreases non-linearly under predominant tensile multi-axial stress state. The elastic-plastic assessments need to show that to underwrite the integrity of a component, the final accumulated equivalent plastic strains are below the available multi-axial ductility for the respective material/s at the relevant temperature/s. Extensive FE analyses related to regulatory requirements will be done. Comments will be made on mesh quality, input parameters and post-processing diagnostics in an attempt to guide the analyst to a fit-for-purpose FE model. Finally, a quantifiable and recommendable margin when comparing the strains as stated above will be suggested.