A characteristic time-based heat input model for simulating selective laser melting

The residual stress induced by rapid temperature cycling in a component during a selective laser melting (SLM) process has a significant detrimental effect on the component's quality and may cause component failure during or after the printing process. Predicting and therefore actively controlling the residual stress are important research topics in SLM. Numerical simulation is an effective tool to predict the residual stress and distortion throughout the printing process. Goldak heat input model has been commonly used to model the heat source in an SLM process. However, it incurs an unpractically high computational cost when simulating building a part with dimensions in centimetres due to the huge scale gap between the component size and the laser spot size. In the present work, a characteristic time-based heat input (CTI) model has been developed and implemented in a finite-element thermal-mechanical model to significantly reduce the process simulation time while satisfactorily predicting the temperature and residual stress state of SLMed parts. This characteristic time-based heat input model speeds up the computation by applying the integrated energy along the scan path over a characteristic heating time, which is defined as the ratio of the axis of the ellipsoidal heat source in the laser scanning direction to the scanning speed. The characteristic heating time ensures that the peak temperature and subsequent heat transfer of each deposited track can be satisfactorily captured in just one numerical step, respectively. The present model was calibrated and validated by identified isothermal curves on the cross section of a printed Ti6Al-4V track. Case studies demonstrate that the simulated temperature profile and stress field are in a good agreement with Goldak model. The maximum distortion of a 32 mm single-track sample predicted by CTI model is within 6% of that predicted by Goldak heat input model while CTI's computational time is just 0.4% of Goldak's.