Manufacturer specifications for the resolution of an additive manufacturing (AM) machine can be ten times smaller (more optimistic) than the actual size of manufacturable features. Existing methods used to establish a manufacturable design rule-set are conservative piecewise-constant approximations. This paper aims to evaluate the effectiveness of a first-order model for producing improved design rule-sets for feature manufacturability, accounting for process variation.
A framework is presented which uses an interpolation method and a statistical model to estimate the minimum size for a wide range of features from a set of iterative experiments.
For an SLS process, using this approach improves the accuracy and reliability of minimum feature size estimates for a wider variety of features than assessed by most existing test artifacts.
More research is needed to provide better interpolation models, broaden applicability and account for additional geometric and process parameters which significantly impact the results. This research focuses on manufacturability and does not address dimensional accuracy of the features produced.
An application to the design of thin channels in a prosthetic hand shows the utility of the results in a real-world scenario.
This study is among the first to investigate statistical variation of “pass/fail” features in AM process characterization, propose a means of estimating minimum feature sizes for shapes not directly tested and incorporate a more efficient iterative experimental protocol.
