Applying Digital Twin Methods for Process-Structure-Property Correlation Assessment in Metal Additive Manufacturing with Limited Experimental Data

Additive manufacturing (AM) is expanding for fabricating complex, high-value, and large-size and low-volume parts. While AM holds significant potential for producing a variety of components, it is essential to conduct quality assessment of AM components to ensure their safe and reliable operation. Low-volume fabrication of parts imposes limitations on the number of inspection samples for quality verification. Limited availability of inspection samples can be addressed by nondestructive testing (NDT) of fabricated samples, and the development of digital twin models. In this study, the Process-Structure-Property (P-S-P) of stainless-steel AM parts fabricated via laser powder bed fusion technique are evaluated using a limited number of inspection samples by various material characterization and NDT. The correlation between process parameters and part properties has been assessed. Results indicate the P-S-P relationship where apart from the ultimate strength values, all experimental results display bimodal distributions, featuring two distinct modes. The correlation analysis reveals that higher laser power is associated with increased RF Z-score and Modulus. Utilizing the experimental findings, a FEM serving as a digital twin of the parts is constructed. This model enables the extension of quality assessment for the parts across a broader spectrum and facilitates the examination of P-S-P relations.