Methodology to Determine Melt Pool Anomalies in Powder Bed Fusion of Metals Using a Laser Beam by Means of Process Monitoring and Sensor Data Fusion

Abstract

Additive manufacturing, in particular the powder bed fusion of metals using a laser beam, has a wide range of possible technical applications. Especially for safety-critical applications, a quality assurance of the components is indispensable. However, time-consuming and costly quality assurance measures, such as computer tomography, represent a barrier for further industrial spreading. For this reason, alternative methods for process anomaly detection using process monitoring systems have been developed. However, the defect detection quality of current methods is limited, as single monitoring systems only detect specific process anomalies. Therefore, a new methodology to evaluate the data of multiple monitoring systems is derived using sensor data fusion. Focus was placed on the causes and the appearance of defects in different monitoring systems (photodiodes, on- and off-axis high-speed cameras, and thermography). Based on this, indicators representing characteristics of the process were developed to reduce the data. Finally, deterministic models for the data fusion within a monitoring system and between the monitoring systems were developed. The result was a defect detection of up to 92% of the melt track defects. The methodology was thus able to determine process anomalies and to evaluate the suitability of a specific process monitoring system for the defect detection.

Die Forschungsinfrastruktur des SFB TRR 277 AMC Additive Fertigung im Bauwesen

“…Neben der Verwendung der EOS‐Anlagen im AMC werden diese in zahlreichen anderen Forschungsprojekten eingesetzt. Die Projekte befassen sich u. a. mit der experimentellen Validierung einer am iwb entwickelten Prozesssimulation, mit Untersuchungen zu den fertigungstechnischen Grenzen des Verfahrens oder der Fertigung von Mikrostrukturen mit wirksamem und kontrollierbarem Dämpfungsverhalten [42–50].…”

A deep learning solution for real-time quality assessment and control in additive manufacturing using point cloud data

“…Processing the raw data acquired from sensory modules is crucial in generating awareness and extracting knowledge from the undergoing processes. Researchers proposed various methods, from conventional thresholding and image processing techniques (Baumann & Roller, 2016;Borish et al, 2020;Harbig et al, 2022;Zhao et al, 2021) to the latest DL-based algorithms Pandiyan et al, 2022). Traditional methods require manual or semi-manual filters and feature generation and selection, which in small-scale problems would be feasible.…”

In-Situ Quality Assessment and Control in Additive Manufacturing Processes Using Laser Surface Profilometer and Deep Learning Techniques

“…Despite the traditional methods in which the data and features were known to the researcher for an unknown result, in ML-based methods, the data and the results are known, and the goal is to find the appropriate features and formula correlating the input data to the output. Jana et al (Harbig et al, 2022) proposed a manual filter selection and calibration approach for melt pool anomalies detection. The author utilized two photodiodes to gather 13800 frames per second with 160*140 pixels per frame.…”