A Method of Predicting Powder Flowability for Selective Laser Sintering

TL;DR

This paper presents a novel approach combining Revolution Powder Analysis and machine learning to reliably classify powder flowability for Selective Laser Sintering, aiding pre-screening of material systems.

Contribution

It introduces a new method integrating RPA and machine learning to predict powder flowability, which is more reliable than traditional surface roughness and layer density measures.

Findings

RPA can reliably classify powder flowability.

Layer density and surface roughness are less effective for classification.

The method improves pre-screening efficiency for SLS materials.

Abstract

This work investigates a method for pre-screening material systems for Selective Laser Sintering (SLS) using a combination of Revolution Powder Analysis (RPA) and machine learning. To develop this method, nylon was mixed with alumina or carbon fibers in different wt.% to form material systems with varying flowability. The materials were measured in a custom RPA device and the results compared with as-spread layer density and surface roughness. Machine learning was used to attempt classification of all powders for each method. Ultimately, it was found that the RPA method is able to reliably classify powders based on their flowability, but as-spread layer density and surface roughness were not able to be classified.