We developed and applied a novel approach for shape agnostic detection of multiscale flaws in laser powder bed fusion (LPBF) additive manufacturing using heterogenous in-situ sensor data. Flaws in LPBF range from porosity at the micro-scale (< 100 µm), layer related inconsistencies at the meso-scale (100 µm to 1 mm) and geometry-related flaws at the macroscale (> 1 mm). Existing data-driven models are primarily focused on detecting a specific type of LPBF flaw using signals from one type of sensor. Such approaches, which are trained on data from simple cuboid and cylindrical-shaped coupons, have met limited success when used for detecting multiscale flaws in complex LPBF parts. The objective of this work is to develop a heterogenous sensor data fusion approach capable of detecting multiscale flaws across different LPBF part geometries and build conditions. Accordingly, data from an infrared camera, spatter imaging camera, and optical powder bed imaging camera were acquired across separate builds with differing part geometries and orientations (Inconel 718). Spectral graph-based process signatures were extracted from this heterogeneous thermo-optical sensor data and used as inputs to simple machine learning models. The approach detected porosity, layer-level distortion, and geometry-related flaws with statistical fidelity exceeding 93% (F-score).

Additive manufacturingAdditive manufacturing processes such as laser powder bed fusionLaser powder bed fusion produce material by localized melting of a powder feedstock layer by layer. The small melt pools and high energy density generate very different microstructuresMicrostructures in nickel superalloysSuperalloys when compared to more traditional cast or wrought processing, including features such as cellular structures and epitaxial grain growth. The features of these microstructuresMicrostructures vary depending on local thermal history, alloy chemistry, and processing parameters. There is a need to develop a systematic understanding of the influence the local thermal conditions during solidification have on the resulting microstructureMicrostructures. Such understanding will be useful in predicting and ultimately avoiding microstructural defects such as undesirable phases or non-optimal grain structures. In this work, in-situ Longwave Infrared imaging of a laser powder bed fusionLaser powder bed fusion process is used to characterize the local thermal conditions throughout additively manufactured builds for alloy IN718Alloy IN718 processed using systematically varied process parameters. This information is then correlated to observations of the microstructural features of these alloys in the as-built condition. This correlation analysis shows clear influence of the local thermal conditions during solidification on the dimensions of the dendritic microstructuresMicrostructures formed during the build process for IN718IN718. These dendritic structures arise due to segregation of elements such as niobium during solidification, an observation which can be predicted using a Scheil modeling approach.

Quality control and quality assurance are challenges in direct metal laser melting (DMLM). Intermittent machine diagnostics and downstream part inspections catch problems after undue cost has been incurred processing defective parts. In this paper we demonstrate two methodologies for in-process fault detection and part quality prediction that leverage existing commercial DMLM systems with minimal hardware modification. Novel features were derived from the time series of common photodiode sensors along with standard machine control signals. In one methodology, a Bayesian approach attributes measurements to one of multiple process states as a means of classifying process deviations. In a second approach, a least squares regression model predicts severity of certain material defects.

Additive manufacturing and specifically metal selective laser melting (SLM) processes are rapidly being industrialized. In order for this technology to see more widespread use as a production modality, especially in heavily regulated industries such as aerospace and medical device manufacturing, there is a need for robust process monitoring and control capabilities to be developed that reduce process variation and ensure quality. The current state of the art of such process monitoring technology is reviewed in this paper. The SLM process itself presents significant challenges as over 50 different process input variables impact the characteristics of the finished part. Understanding the impact of feed powder characteristics remains a challenge. Though many powder characterization techniques have been developed, there is a need for standardization of methods most relevant to additive manufacturing. In-process sensing technologies have primarily focused on monitoring melt pool signatures, either from a Lagrangian reference frame that follows the focal point of the laser or from a fixed Eulerian reference frame. Correlations between process measurements, process parameter settings, and quality metrics to date have been primarily qualitative. Some simple, first-generation process control strategies have also been demonstrated based on these measures. There remains a need for connecting process measurements to process models to enable robust model-based control.