Finite Element Modelling of Creep Rupture on Grade 91 Steel using Monkman-Grant Ductility based Damage Model

Failure strain is a main parameter used in the ductility exhaustion based damage model in which the accuracy of the prediction is dependent on its input value. The experimental measured has indicated that the value of strain at fracture is extensively scattered, therefore may affect the prediction. This paper presents the result of creep rupture time using a modified creep damage model incorporating Monkman-Grant (MG) failure strain as an alternative to strain at fracture. Both strains at fracture and MG failure strain are separately employed in the damage model to predict the failure time of uniaxial smooth specimen and notched bar with different acuity ratios of 3.0 and 20. The FE model of the specimen is loaded under different stress values and the multiaxial failure strain at each stress level is estimated using Cock and Ashby void growth model. The predicted creep rupture time that is compared to the experimental data (in a range of 40-1000 hours) showing a good agreement within the scatter band of +/-factor of 2. Both approaches using strain at fracture and MG failure strain can be used in predicting the creep failure under uniaxial and multiaxial features. The advantage of using MG strain is that the laboratory creep testing can be interrupted prior to specimen fractured or once the secondary creep deformation occurs. Meanwhile, the determination of strain at fracture needs longer test duration where the test can bestopped only when the specimen broken.