Increasingly, Laser Powder Bed Fusion (L-PBF) Additive Manufacturing (AM) machines have been instrumented with nondestructive sensors that collect both in-situ and ex-situ measurement data. In-situ sensors include co-axial imagers, off-axis imagers, acoustic sensors, accelerometers, and chamber environment sensors. Typically, each sensor is designed to collect only one type of data. Consequently, when monitoring L-PBF processes, multi-sensor data fusion is necessary to provide a broad range of data on measured powder, melt-pool geometry, layer defects, and fabricated part. The presentation provides data fusion scenarios that describe possible measurement approaches to extract process signatures from the multiple types of sensor data. Signatures are used to monitor and control the L-PBF AM process. However, aligning those different data types to the AM-part-build coordinate system, which is necessary for data fusion, is still a major problem. In addition, the quantity and quality of the different data types are causing curation, organization, and administration problems that can impact the benefits of collecting those different data types. The selected scenario describes how data registration, data fusion, and data analytics can be used to detect anomalies in the AM process. Anomalies can occur while spreading a powder layer, melting the powder using the laser, scanning a completed layer, and the inspecting the fabricated part. Usually, detecting anomalies can only be done accurately by using and fusing data from a combination of sensors for analyzing that fused data using predictive modeling tools, such as machine learning techniques.
- Take away how non-destructive sensors can detect anomalies in powder bed fusion processes for quality enhancement
- Have an overview of a spectrum of non-destructive sensors used in in-situ and ex-situ monitoring of additive manufacturing process and fabricated products