Detecting voids in 3D printing using melt pool time series data
- PDF / 1,439,360 Bytes
- 8 Pages / 595.276 x 790.866 pts Page_size
- 53 Downloads / 175 Views
Detecting voids in 3D printing using melt pool time series data Vivek Mahato1 · Muhannad Ahmed Obeidi2 · Dermot Brabazon2 · Pádraig Cunningham1 Received: 27 April 2020 / Accepted: 11 October 2020 © Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract Powder Bed Fusion (PBF) has emerged as an important process in the additive manufacture of metals. However, PBF is sensitive to process parameters and careful management is required to ensure the high quality of parts produced. In PBF, a laser or electron beam is used to fuse powder to the part. It is recognised that the temperature of the melt pool is an important signal representing the health of the process. In this paper, Machine Learning (ML) methods on time-series data are used to monitor melt pool temperature to detect anomalies. In line with other ML research on time-series classification, Dynamic Time Warping and k-Nearest Neighbour classifiers are used. The presented process is effective in detecting voids in PBF. A strategy is then proposed to speed up classification time, an important consideration given the volume of data involved. Keywords Process monitoring · Classification · Time-series
Introduction This study is focused on metal powder printing, an additive manufacturing process that is having a high impact in industries such as automotive, aerospace and biomedical. In additive manufacturing, functioning parts can be produced directly from a computer aided design model without the need for moulding or prototyping. The PBF process is shown in Fig. 1. The part is built as the laser traces out the part in layers in the powder bed. The overall process is complex and build quality is very sensitive to process parameters (Aminzadeh and Kurfess 2019; Kwon et al. 2020). For this reason, there has been considerable research on analytics methods for in-situ monitoring to detect the onset of defects. Grasso and Colosimo (2017) provide a comprehensive review of this research. It is clear that the temperature of the melt pool is an important signal for monitoring build quality. In this paper we explore the hypothesis that anomalies such as pores can be detected by identifying characteristic signatures in melt pool time-series data.
B
Pádraig Cunningham [email protected] Dermot Brabazon [email protected]
1
University College Dublin, Dublin, Ireland
2
Dublin City University, Dublin, Ireland
While the ultimate objective is to detect pores arising from incorrect process parameter, the data analysed here comes from blocks containing voids by design. The size of these voids is set to correspond to the size of pore defects arising during normal PFB machine part production. The fact that the void locations are known by design provides a ground-truth to assess the performance of the ML methods. Our analysis uses data from the Aconity MINI 3D printer.1 The temperature data comes from pyrometers that monitor melt pool temperature. Two pyrometers supplied by KLEIBER Infrared GmbH detect the heat emission light in the range of 1500 t
Data Loading...
