Exploitable results
The establishment of safe methods for computer aided transfer and/or extrapolation of material properties to metallic components under stress and dwell time at extended high temperature range has been achieved. 2 alloys were selected for evaluation: the austenitic steel Alloy 800H, tested at 850 C, and the ferritic steel 2.25Cr1Mo, tested at 550 C. Creep analysis was done using the theta projection method and a series of high accuracy constant stress tests were performed. For low cycle fatigue the strain range partitioning (SRP) method was selected. For 2.25Cr1Mo, the results indicated that SRP is sufficent for life prediction purposes, but for alloy 800H at 850 C the standard model could not properly account for the strong hold time observed. Metallographic investigations showed this to be related to surface oxidation effects. To overcome this, a modified SRP method was developed, which successfully correlates all project data and its accuracy has been further confirmed by checks using literature data.
2 advanced viscoplastic models, the Walker model and the Chaboche model have been implemented in the MARC and ABACUS FE codes respectively, which allow finite element analysis of high temperature components subject to typical service load conditions. A specially designed programme of test was performed to allow the constants in these models to be determined. More complex analysis has been performed with the Chaboche model for a notched pipe and with the Walker model by calculating the response of a turbine disc to a typical duty cycle. The latter example involved a 3-dimensional simulation of the disc itself, and complex nonisothermal and nonproportional loading conditions.
Stress strain histories for typical components were determined as a result of the numerical analysis. After some simplification, these were then simulated on uniaxial specimens. A series of such component stress strain history simulation tests were carried out for Alloy 800H at 850 C. The mo dified SRP method provided accurate life predictions for all these tests, whereas the standard SRP predictions were again strongly nonconservative.
Significant improvements in component behaviour assessment have been made by:
a refined analytical model to improve the correlation of long term relevent service behaviour with the short term creep test data;
modelling creep fatigue material deformation behaviour using a unified viscoplastic theory, and applying this to numerical analysis of components using optimised procedures;
the developemnt of a component life prediction method considering creep fatigue environment damage;
component life prediction via the use of the numerical and experimental methods for simulation of actual component stress strain histories.
Searching for OpenAIRE data...
There was an error trying to search data from OpenAIRE
No results available