The digital twin concept has emerged as means for predicting, observing, and controlling complex systems. These data-driven representations provide virtual interfaces for interacting with physical systems, allowing for the observations of changing states and behaviors. While the concept of the digital twin itself remains broad, when scoped by an application the opportunities become apparent. Digital twins in additive manufacturing (AM) are being developed as a means for observing and controlling both machines and processes. More recently, the digital twin is being explored as a means for observing the efficacy of the process to establish the provenance of the part being manufactured. This particular application is unique as the observations of the process can be associated with both the process and the part. In this context, the ability to validate expected process behaviors becomes an invaluable tool, as verification and validation at the process level can be leveraged to provide confidence into the behavior of the manufactured part. One of the major barriers in AM continues to process consistency. Slight changes in design, process, and material configurations can lead to substantial differences in final part quality. The general consensus is that variability in processes must be controlled to minimize potential discrepancies, that is to say repeatability of the process is necessary to develop qualified parts. Towards this goal, many have taken to various in process monitoring techniques to provide evidence of process consistency during the development of the part. Alternatively, if process variations can be better quantified against expected behaviors, the impacts of these variations on the part may be better understood and subsequently considered. This work investigates the nexus of process and part digital twins, and the use of predicted, physics-based process behaviors to validate observations made at the part meso-scale. By leveraging physics-based models we will look to leverage the digital twin concept to continuously validate the fabrication of a part until a macro-scale observation can be made. The multi-scale approach aims to provide newfound confidence in the fabricated part grounded in fundamental physics and in process measurements.