High fidelity multi-physics modeling of laser metal interaction and keyhole dynamics in powder bed fusion

Porosity defects in laser powder bed fusion processes are critical, particularly when initiated in the high laser energy density regime, known as the keyhole regime. In-situ understanding of melt-pool dynamics is essential, which can be achieved through high energy radiography techniques. However, these techniques are costly and time-consuming, prompting the need for high fidelity numerical models. This paper presents a numerical model aimed at studying the correlation between powder bed fusion process inputs and the formation of deep and narrow vapor depressions (keyholes) thereby revealing interesting aspects that are difficult if not impossible to identify using experimental setups. The model consists of two coupled components: a multi-phase, multi-physics model to solve melt pool dynamics and a high-performance ray tracing-based multi-reflection laser radiation model. The metal free surface predicted by the first model is input into the radiation model, which calculates the effective energy absorbed, resulting in a fully coupled simulation Results show good agreement with experimental benchmark data from dynamic x-ray radiography techniques, with maximum deviations of 11% in average keyhole depths and 2 to 8% in laser absorption time profile and final pores were successfully predicted. The investigation of the impact of maximal temperature spot on keyhole shape and dynamics, revealed distinct keyhole shapes categorized by average maximal temperature spot and uncovering a clear pattern behind pore-keyhole separation, leading to a better understanding of chaotic keyhole mode where pores are generated and pushed into the melt pool. The developed numerical model offers a cost-effective alternative to experimental techniques and can aid in optimizing process parameters to mitigate porosity defects in additive manufacturing.