چكيده به لاتين
Accurate prediction of creep life is critical to ensure the health and long-term performance of mechanical parts and equipment that are exposed to high stress and temperature, such as power plant components, aerospace industries, and so on. Continuous damage mechanics models, as one of the common methods, through fundamental equations, are an important step in the direction of creep phenomenon modeling and creep life estimation. The significant advantage of these methods is cost-effectiveness and providing more accurate results compared to other life prediction methods. However, the main challenge in applying these models is to determine the material parameters associated with each model, which are often dependent on applied stress and temperature .
In this research, the performance of four continuum damage mechanics models, including Kachanov-Robotnov (KR), Liu-Murakami, Othman-Hayhurst-Dyson (OHD) and Kowalewski-Hayhurst-Dyson (KHD) in predicting the failure time under uniaxial tension, for short-term tests Three materials were evaluated: 91P steel, stainless steel and Vaspaloy superalloy. Estimation of material parameters for each model was done using single-objective genetic algorithm and model error optimization based on short-term test data. Then, the long-term creep strain-time curve was extracted in each model and compared with the data. The evaluation criteria of the models included failure time prediction error and creep strain curve fitting error .
The results showed that the KR model provides the best performance in predicting the failure time for all three materials. OHD model performed better in curve fitting on laboratory data for Vaspaloy superalloy and Liu-Murakami model for two other materials (91P steel and stainless steel). Finally, it can be concluded from this research that in order to predict the life of materials with a simpler microstructure, accurate estimation of the material parameters of a model takes precedence over the complexity of that model. However, sometimes for materials with a more complex microstructure, we have to choose a more complex model while accurately estimating the material parameters.
Keywords: Creep, Continuum damage mechanics, Genetic algorithm, Short- and long-term data, Model lifetime estimation.
