Powder-fed Laser Directed Energy Deposition of Near-β Titanium Alloy: A Comprehensive Analytical and Experimental Analysis
Loading...
Date
2021-10-12
Authors
Ansari, Mazyar
Journal Title
Journal ISSN
Volume Title
Publisher
University of Waterloo
Abstract
Laser directed energy deposition through powder feeding (LDED-PF), a class of metal additive manufacturing (AM), is a promising technique that enables the repair and refurbishment of metallic components with a variety of materials as well as the complete near-net-shape manufacturing of metallic parts with moderately complex geometries. The localized solidification and complex thermal cycles in LDED-PF have a significant effect on the deposition characteristics. The continuous change in the deposition rate occurs due to the thermal-based complexities and alters the deposition geometry and microstructure. There are still major knowledge gaps and challenges in achieving desirable dimensional and microstructural features of as-built components. A clear understanding of the process and its effects on the temperature field and the resulting dimensional and microstructural characteristics are of tremendous importance. Developing methodologies to predict the thermal-based complexities during the multi-track deposition and minimize the adverse effects is essential to enhancing LDED-PF. This thesis aims to develop a time-efficient predictive physics-based model that improves dimensional accuracy and process stability. The model also provides an effective tool for process optimization and microstructural engineering through process-microstructure linkages.
To this end, this research first tries to analytically couple the moving laser beam, powder stream, and semi-infinite substrate. A process map is developed for the single-track deposition, which draws the physical barriers to a stable process. To develop an effective heat source model that is more in harmony with the physics of the process, analytical solutions to three heat source models are introduced and compared to describe the transient temperature field in the single-track deposition. To improve the model fidelity, the enhanced thermal diffusivity and heat source radius are calibrated in terms of linear functions. To include the effect of scanning strategies in multi-track LDED-PF, the model is adapted for multi-track deposition of different scanning strategies. The temperature fields of cuboid-geometry depositions are simulated under four scanning strategies, namely, bidirectional, unidirectional, inward spiral, and S-pattern. The effect of heat accumulation is considered, and 2D thermal models are time-efficiently computed to obtain the melt pool shape. A new universal algorithm, based on parabolic functions, is developed to predict the geometrical profile of the overlapping beads, which applies to all scanning strategies. The model obtains the local heat flow direction at the longitudinal center-plane of each track. The solidification parameters of thermal gradient and solidification rate are then extracted from the local heat flow to reveal the linkages between the process and solidification microstructure. The solidification maps are established, and the microstructural transition of columnar-to-equiaxed is predicted. The developed model is validated for multi-track deposition of Ti-5Al-5V-5Mo-3Cr alloy at different laser powers, scanning speeds, and step-over distances under different scanning strategies. The developed knowledge of solidification characteristics and microstructural evolution can greatly contribute to the development of the Ti-5Al-5V-5Mo-3Cr alloy processed by LDED-PF.
Lastly, an adaptive prediction protocol is developed based on physics-based mathematical modeling to control the temperature and deposition dimensions. A simulation-based algorithm is designed for multi-track deposition of direction-parallel scanning strategies. The algorithm simulates the transient temperature field and geometry of subsequent tracks to adaptively predict the required laser power as the main processing parameter, such that each track arrives at the desired dimensions. The performance of the developed adaptive modeling is experimentally evaluated and validated. The adaptive protocol of modeling, as a cost-effective approach, only requires the thermophysical properties of the material and some basic information of the LDED-PF setup to successfully improve the dimensional accuracy and flatness of the deposited layers by neutralizing the adverse effect of heat accumulation and overlapping beads. This unique adaptive protocol can also ensure the uniformity of the deposition and the stability of the process by maintaining the standoff-distance constancy.