Additive Manufacturing Process Root Selection Using Bayesian Network

Abstract

Additive manufacturing emerged as a highly promising technology within the realm of manufacturing, offering advantages such as rapid customization, on-demand production, the ability to create intricate shapes, and the absence of economies of scale dependency. Its widespread adoption is evident across various industries, including aviation, oil and gas, and the automotive sector. But finding the most favorable process root to 3D print a part that makes sense both technically and economically is a challenging and tricky task because there are many different methods, each with its own unique constraints. This article focuses on the complex decision-making process of choosing which process root to be selected for 3D printing something, considering various attributes. It introduces a novel framework combining the Delphi methodology, the Neutrosophic Best-Worst Method, and Bayesian Network analysis to discern the most viable additive manufacturing process root for a given part. To validate the efficacy of this framework, a practical case study involving Ferro Oil Tech India Pvt. Ltd. is employed. The study evaluates three discrete additive manufacturing process routes: direct fabrication via direct metal laser sintering, selective laser sintering combined with investment casting, and stereolithography coupled with investment casting. The findings indicate that the direct metal laser sintering process proves to be both economically viable and technically feasible.