Optimizing the Expected Utility of Shape Distortion Compensation Strategies for Additive Manufacturing

Abstract

In the past two decades, the field of additive manufacturing (AM) has seen tremendous growth, especially in the production of functional parts. Unfortunately, improving the dimensional accuracy of these printed parts to the point where they can be used for a broad range of applications has proven challenging. Several methodologies to improve the dimensional accuracy of 3D printed parts have been proposed in the literature. One approach that has seen a considerable amount of work in recent years is product design adjustment based on predictive modeling. Under this approach, predictions of geometric deviations across the surface of a part are used to modify the shape of a part before printing so as to counteract or compensate for the predicted deviations. However, a majority of compensation methods aim at minimizing expected geometric and dimensional error, with a lack of consideration of cost and uncertainty. This study presents a new strategy based on multi-attribute utility theory to account for cost and inherent uncertainty associated with a compensation decision. By establishing manufacturer preferences and prior beliefs about the efficacy of a predictive model, the proposed decision-making strategy for compensation significantly increases the value of a given print to a manufacturer under simulated preferences.