Dual-inherent strain method for efficient residual stress prediction in additive manufacturing process

Abstract

Accurate and efficient estimation of residual stress and distortion at the design stage remains crucial and challenging, primarily due to the complex and subjective calibrations required by existing modeling methods, hindering their widespread adoption in the additive manufacturing (AM) process. The inherent strain method (ISM) has been widely used due to its proven effectiveness in predicting distortion and for its low computational cost as simulations rely on a single static analysis. Inherent strain is calculated from a meso-scale thermomechanical simulation and applied to a part-scale mechanical simulation, significantly reducing computational time. However, predicting residual stress with conventional ISM requires additional fictious non-physical property modeling, limiting its effectiveness. This study proposes a novel dual-inherent strain method (DISM) for efficient residual stress simulation in additive manufacturing, while retaining the distortion prediction accuracy. The method improves residual stress prediction with a new procedure to calculate and apply strain in part-scale simulations, capturing strain evolution during the AM process without needing calibration in property modeling. The proposed method is validated by experiments conducted on a Ti-6Al-4V double-cantilever beam processed by laser powder bed fusion. Additionally, the proposed approach eliminates the need for calibration of a threshold of fictious non-physical material property used in the conventional ISM, which significantly influences prediction results. On average, the new DISM achieves a 41 % reduction in computational time compared to the conventional ISM. This systematic method is adaptable to different materials and processes without the need for recalibration, making it broadly applicable to various scenarios in additive manufacturing.