Multistage Fatigue Modeling of AZ91 and AM60 Magnesium Alloys

Jump to: navigation, search


In order to construct physically-motivated and microstructurally-sensitive fatigue estimations of the AZ91 and AM60 cast magnesium (Mg) alloys presented in this study, the Multistage Fatigue (MSF) model was employed.


Model Correlations and Fatigue Estimations

Figure 18a-c shows the MSF model correlations for the AZ91 as-cast, AZ91-T6, and AM60 as-cast alloys, respectively. The mean MSF model predictions for each data set were based on the stabilized cyclic stress-strain behavior and the average measured microstructural features, including the dendrite cell size (DCS) and the average size of the pore responsible for crack initiation. Regarding the percentage of incubation life in the mean life predictions, an average incubation life of 34%, 56%, and 47% of total life was found suitable for the AM60 as-cast, AZ91 as-cast, and AZ91-T6 alloys, respectively. Upper and lower bounds of the strain-life data sets were determined by using the extreme pores sizes and DCS’s for each data set. The model showed good correlation to both the upper and lower bounds as compared to the experimental results. As such, the MSF model was generally able to bind the scatter of each fatigue data set considering only the microstructure variation observed between locations L1 and L2. The model showed more sensitivity to pore size than to DCS, which is not unexpected, considering there was not much variation of the DCS from locations L1 and L2. As such, the defect size was the primary factor determining number of cycles to failure for both the AM60 and the AZ91 alloys.

Figure 18. Comparison of the strain-life experimental results for locations L1 and L2 to the mean, upper, and lower bound estimations of the Multistage fatigue model: a) AZ91 as-cast; b) AZ91-T6; c) AM60 as-cast
Personal tools

Material Models