Field-driven data processing paradigm for multi-information additive manufacturing

Abstract

Multifunctional components with various information such as material, structure, process, and performance have become viable and accessible to the industry due to the rapid growth of additive manufacturing (AM) technology from contour modeling to functional modeling. However, multi-information additive manufacturing requirements far outpaced the processing capabilities of current data processing systems to model and interpret multi-information digital models. This paper proposes a field-driven processing paradigm with the ability to describe and parse complex, multi-information distributions. The robustness and flexibility of field-driven design are fully exploited by uniformly converting data including common AM models such as meshes, function formulas, and point clouds into control field representations and setting reasonable control-property mapping rules. The complexity of multi-information modeling is significantly reduced via a combined strategy of control fields and mapping relationships. A complete multi-information model only consists of initial data related to desired properties and corresponding mapping rules. Compared with the voxel method, the amount of data is reduced by more than 80%. The multi-information model is discretely parsed in different dimensions according to manufacturing requirements to efficiently and accurately generate property data for industrial manufacturing. Two representative multi-information demonstrators incorporating material-structure dual gradients and structure-process integration are designed and printed while 3 AM techniques are employed to validate the utility of this paradigm. The paradigm is anticipated to open up an efficient route for the realization of material-structure-process-property integrated AM.