Economic production of small lots of components is a long-standing problem in industry. A potential solution is additive manufacturing (AM) which has low preparation cost and high capability for producing complex geometry. It is a manufacturing process that adds material bottom-up. Non-critical components fabricated via AM are now being used in motor vehicles, consumer products, medical products and aerospace devices. 

For engineering applications, high value-added components require consistency in fatigue properties. However, components fabricated by AM have large variation in the fatigue properties compared to those by conventional manufacturing processes. To alleviate unpredictable catastrophic failures, it is essential to study fatigue behavior. Previous study reported that fatigue crack initiation mechanism accounts for a large portion of fatigue life, especially for low amplitude and high cycle fatigue under real world loading conditions. However, this major portion of fatigue life prediction is mostly ignored by main-stream researchers working on fatigue modeling. This research has made a major advance in this new fatigue research and modeling direction. 

The objective of this research is to develop a fatigue crack initiation model for metal components produced by AM. To improve life prediction accuracy, the model must incorporate the effect of different microstructures produced in the components, which is complex due to numerous cycles of re-heating and re-cooling processes. The project is thus separated into three tasks: (1) developing a temperature model to account for the thermal cycling history, (2) modeling the component’s microstructures produced by the cycling temperatures for the potential crack initiation zone, and (3) developing a fatigue crack initiation model for life estimation. A summary of each task is provided in the following.

First, the role of temperature model is to predict the temperature history at each material point for the material volume of interest, as an input to the microstructure-temperature model for predicting the distribution of microstructures resulted from the AM process. The existing temperature models are computationally expensive to for accurate prediction of the temperature resulted from repetitive heating and cooling. The main reason is that these models considered entire boundary conditions of all the material points. In this project, we proposed and employed the new concept of effective computation zone, which can save the computational time significantly for the AM process. 

Second, The microstructures considered consists the grain size and the phase. The grain size variation is modeled by using representative volume element, which is defined as a volume of heterogeneous material that is sufficiently large to be statistically representative of the real component’s microstructure. The material phase of each point is obtained by modeling the process of phase transformation. A continuous cooling transformation (CCT) diagram is used with a thermal model for predicting the resultant microstructure phases. Since traditional CCT diagrams are developed based on slow and monotonic cooling processes, such as furnace cooling and air cooling, which are very different from the repetitive heating and cooling processes in AM. In this study, a new general methodology is presented to create CCT diagrams for materials fabricated by AM. The result showed that the effect of the segmented duration within the critical temperature range, which induced precipitate formation, could be cumulative.

Third, the existing fatigue crack initiation life model has poor accuracy. One reason for the poor accuracy is the coefficients change due to the variation in the microstructure is not accounted for. In this research task, a semi-empirical fatigue crack initiation model was developed with microstructure-dependent coefficients of maximum persistent slipband width, energy efficiency coefficient, resolved shear stress and plastic slip rate per cycle. Methods of determining these coefficients as functions of microstructures, which can improve the accuracy of fatigue life estimation. 

The outcome of this project is to provide a new engineering tool for understanding and designing the melting AM process for metals based on scientific models. With this tool, the fundamental mechanism contributing to a large variation of the fatigue life of the metal components made by AM process can be understood, attributed, and predicted. The seemly ‘stochastic’ nature of fatigue life of the AM components can be changed to be more deterministic and pedictable. This approach represents a major advance in new and effective fatigue modeling and microstructure prediction methodology for AM made components.  The models and methodology developed can be used as tools for research, design, and control for laser-based AM process applications.

Last Modified: 06/25/2020
Modified by: C. Richard Liu