Melt Pool Imaging for SLM Porosity — PatSnap Eureka
In-Situ Melt Pool Imaging: Reducing Porosity Variability in Selective Laser Melting
Real-time melt pool monitoring detects and corrects the process instabilities that cause sporadic porosity in SLM. Explore the sensing modalities, feedback architectures, and patent landscape — synthesised from 60+ sources across 2010–2024.
Two Distinct Porosity Regimes — Both Rooted in Melt Pool Instability
Porosity in SLM arises from two physically distinct regimes, each driven by melt pool instability. Lack-of-fusion porosity occurs when adjacent scan tracks or successive layers fail to overlap sufficiently, leaving unmelted powder trapped between solidified regions. Research from Politecnico di Milano (2022) shows that even when analytical models predict adequate melt pool overlap, the inherent statistical variability of melt pool dimensions — not merely their mean values — governs the probability of inter-track voids. Monte Carlo simulations in that study demonstrate that neglecting this variability leads to systematic underestimation of lack-of-fusion risk.
Keyhole porosity is generated when high energy density drives deep vapor depression collapse. As described by Lawrence Livermore National Laboratory (2016), recoil pressure and Marangoni convection together create a topological depression in the melt pool that, upon collapse, entraps gas at the pool bottom. Their powder-scale model incorporating laser ray-tracing distinguished pore formation at three spatially distinct locations: the edge of scan tracks, the pool bottom during depression collapse, and in denudation zones — all of which in-situ imaging must be sensitive enough to discriminate.
Compounding these mechanisms is the thermal history of the powder bed. Research from the University of Louisville (2022) showed that back-and-forth raster scanning progressively heats the substrate, altering melt pool size along each successive track. Scan length and hatch spacing were both identified as variables controlling the magnitude of this thermal drift — meaning a monitoring system must track melt pool evolution dynamically within each layer, not just at build initiation. Learn more about how PatSnap supports materials science R&D across advanced manufacturing disciplines.
In-Situ Melt Pool Imaging: Four Principal Sensing Modalities
From industrially deployed coaxial photodetectors to research-grade synchrotron X-ray systems, each modality offers distinct tradeoffs between resolution, throughput, and deployment practicality.
Coaxial Optical & Infrared Imaging
The most industrially deployed architecture integrates a photodetector or camera into the laser beam path, collecting emitted visible or near-infrared radiation. Concept Laser GmbH pioneered spatial mapping of photodetector signals to build coordinate space (2012), enabling layer-by-layer quality maps that detect a wide range of process anomalies during the build. Melt pool boundary extraction via spatial maximum of temperature gradient — rather than absolute temperature thresholds — improves robustness when emissivity variations confound measurements, as demonstrated for Ti-6Al-4V by Chongqing Institute (2019).
Layer-by-layer quality mappingTwo-Wavelength Pyrometry & Holography
Two-color pyrometry addresses emissivity uncertainty by computing temperature ratios between two wavelength bands. Imperial College London's coaxial two-wavelength camera operating at 100 kHz with 20 μm spatial resolution simultaneously resolved melt pool surface temperature fields and quantified thermal gradients (5–20 K/μm) and cooling rates (1–40 K/μs) in Ti-6Al-4V — sufficient resolution to discriminate between conduction-mode and keyhole-mode melting. ONERA extended this with multi-wavelength digital holography for full-field, single-shot acquisition of 316L melt pools.
5–20 K/μm thermal gradients resolvedSynchrotron X-ray Imaging
For research-grade characterization, synchrotron X-ray imaging provides unprecedented sub-surface visibility into melt pool and pore dynamics. Diamond Light Source (2018) used high-speed operando imaging to directly observe pore migration driven by Marangoni flow at 0.4 m/s, pore dissolution via laser re-melting, and the transition from continuous hemi-cylindrical tracks to disconnected beads as linear energy density decreases. Virginia Tech's multi-sensor study (2022) integrated synchrotron X-ray, high-speed IR, and high-spatial-resolution IR cameras to jointly characterize melt pool shape, keyhole geometry, vapor plume, and thermal evolution.
Direct pore migration observationHigh-Speed Camera + FEM Correlation
Xi'an University demonstrated combining finite element simulation with online high-speed imaging using laser supplementary lighting (2020). Simultaneous deployment of simulation and imaging allowed process parameter sensitivity to be mapped and discrepancies between predicted and observed melt pool geometry to be traced to specific physical causes — a methodology applicable to in-process anomaly attribution. This approach establishes the digital twin framework increasingly central to IP analytics in advanced manufacturing.
Anomaly attribution via FEM correlationQuantifying Melt Pool Sensitivity and Monitoring Capability
Key quantitative relationships from the patent and literature corpus, illustrating why passive SLM cannot maintain consistent porosity levels without active in-situ correction.
Parameter Deviation vs. Melt Pool Geometry Change
Nonlinear amplification: a ~5% parameter deviation produces ≥10% geometry change; an 11% beam size change causes >40% spatter volume change (UW-Madison, 2022).
Two-Wavelength Melt Pool Imaging: Measurement Capabilities
Imperial College London (2018) coaxial setup achieved 100 kHz frame rate, 20 μm spatial resolution, resolving 5–20 K/μm thermal gradients and 1–40 K/μs cooling rates in Ti-6Al-4V.
Closed-Loop Melt Pool Monitoring Architecture: From Signal to Porosity Reduction
The complete signal chain from raw melt pool emission through feature extraction, anomaly detection, and feedback actuation — synthesised from Concept Laser (2010/2012), GE (2021/2024), and Nottingham Trent University (2021).
Linking Melt Pool Signals to Internal Flaws
Translating raw imaging data into actionable porosity predictions requires statistically validated signal-to-flaw correlations — the frontier of process qualification research.
Statistical Spread, Not Mean, Governs Porosity Risk
Chalmers University of Technology's landmark study (2021) on Hastelloy X was among the first to directly link the statistical spread of melt pool signals — not just their mean level — to the incidence of porosity. This reframes monitoring from anomaly detection to population-level quality prediction, requiring systems to capture variability across the full process space. The PatSnap analytics platform supports this kind of process mapping at scale.
Melt Pool Radiation Intensity Predicts Microhardness
Air Force Engineering University (2022) demonstrated that a data-driven model fusing process parameters with power spectrum features of melt pool intensity signals can predict microhardness of LPBF specimens. This shows that melt pool optical signals carry not only geometric information but also microstructure-correlated thermal history signatures, which relate indirectly to porosity content through densification behavior.
From Melt Pool Signal to Closed-Loop Process Control
Three complementary control strategies — reactive feedback, model-predictive feed-forward, and shape-descriptor regulation — together address the full spectrum of melt pool instability.
Identify Freedom-to-Operate Risks in Melt Pool Control Patents
GE holds active US and EP patents on boundary geometric length control. Understand the claim landscape before you build with PatSnap.
Key Assignees and Their Distinctive Contributions
The melt pool monitoring patent and literature space is shaped by a small group of highly active organisations, each contributing a distinct methodological advance.
General Electric Company
Holds the most commercially significant active patents in the dataset, with claims covering geometric-length-based melt pool boundary mapping and process control in both US (2021) and EP (2024) jurisdictions. GE's approach advances the field by moving from single-scalar signals (area, width) to a shape-descriptor — boundary length — that captures morphological irregularities associated with process instability. This represents the current frontier of industrial melt pool control IP. The PatSnap customer success team supports IP professionals navigating landscapes like this one.
Boundary geometric length control (US + EP)Chalmers University of Technology
Contributes the most rigorous statistical framework linking in-situ melt pool monitoring signal distributions to internal flaw statistics in Hastelloy X. Their 2021 study is among the first to directly link the statistical spread of melt pool signals — not just their mean level — to the incidence of porosity, establishing that monitoring must capture variability, not merely central tendency, to serve as a reliable quality proxy. This methodological advance reframes monitoring from anomaly detection to population-level quality prediction.
Statistical signal-to-flaw framework (Hastelloy X)University of Wisconsin-Madison
Contributes two high-impact X-ray imaging studies quantifying how parameter deviations and powder particle size distributions propagate into melt pool geometry uncertainty. Their 2022 study demonstrated that a deviation of only ~5% from optimized laser parameters produces a ~10% or greater change in depression zone and melt pool geometry, while an 11% change in beam size can alter spatter volume by over 40% — providing the physical basis for why monitoring is necessary even under nominally optimized conditions.
~5% deviation → ≥10% geometry change (Ti-6Al-4V)Nottingham Trent University
Provides the most complete feedback control system design in the dataset, including thermal modeling of inter-track heat coupling and a controller architecture targeting regulated melt pool cross-sectional area. Their 2021 study modeled the disturbing heat from neighboring tracks as an initial temperature offset and designed a controller to reject this disturbance by dynamically adjusting laser power or speed — directly addressing the residual heat mechanism and demonstrating the practical control pathway from mechanism understanding to closed-loop implementation. Explore related process control research across industries.
Constant melt pool area feedback controllerMelt Pool Imaging for SLM Porosity — key questions answered
Porosity in SLM arises from two physically distinct regimes. The first is lack-of-fusion porosity, produced when adjacent scan tracks or successive layers fail to overlap sufficiently, leaving unmelted powder trapped between solidified regions. The second is keyhole porosity, generated when high energy density drives deep vapor depression collapse — recoil pressure and Marangoni convection create a topological depression in the melt pool that, upon collapse, entraps gas at the pool bottom.
Research from the University of Wisconsin-Madison demonstrated that a deviation of only ~5% from optimized laser parameters produces a ~10% or greater change in depression zone and melt pool geometry, while an 11% change in beam size can alter spatter volume by over 40%. This nonlinear amplification of input uncertainty into geometric variability is the primary motivation for active in-situ correction.
Monte Carlo simulations from Politecnico di Milano (2022) show that neglecting melt pool dimension variability leads to systematic underestimation of lack-of-fusion risk. Chalmers University of Technology's landmark study (2021) directly linked the statistical spread of melt pool signals — not just their mean level — to the incidence of porosity, establishing that melt pool monitoring must capture variability, not merely central tendency, to serve as a reliable quality proxy.
Two-wavelength (or two-color pyrometry) setups address emissivity uncertainty directly. Imperial College London's work used a coaxial two-wavelength high-speed camera operating at 100 kHz with 20 μm spatial resolution to simultaneously resolve the melt pool surface temperature field and quantify thermal gradients (5–20 K/μm) and cooling rates (1–40 K/μs) in Ti-6Al-4V. This level of thermal resolution is sufficient to distinguish between conduction-mode and keyhole-mode melting, which respectively produce gas and collapse-induced porosity.
General Electric's active patents claim a method of generating a real-time image of the melt pool, mapping the melt pool boundary from pixel-level physical property measurements, computing the geometric length of that boundary, and using this length metric to control at least one aspect of the additive manufacturing process. The geometric length of the boundary — a scalar derived from the full boundary shape rather than a single width or area dimension — provides a more complete characterization of melt pool morphology deviations associated with instability and incipient porosity.
As shown by the University of Louisville (2022), back-and-forth raster scanning progressively heats the substrate, altering melt pool size along each successive track. Scan length and hatch spacing were both identified as variables controlling the magnitude of this thermal drift — meaning a monitoring system must track melt pool evolution dynamically within each layer, not just at build initiation.
Still have questions? Let PatSnap Eureka search the full patent and literature corpus for you.
Ask PatSnap Eureka About Melt Pool MonitoringAccelerate Your SLM Process Qualification with AI-Powered Patent Intelligence
Join 18,000+ innovators already using PatSnap Eureka to accelerate their R&D — search 60+ melt pool monitoring assignees, map FTO risk, and identify the next breakthrough in closed-loop SLM control.
References
- Linking In Situ Melt Pool Monitoring to Melt Pool Size Distributions and Internal Flaws in Laser Powder Bed Fusion — Chalmers University of Technology, 2021
- Uncertainties Induced by Processing Parameter Variation in Selective Laser Melting of Ti6Al4V Revealed by In-Situ X-ray Imaging — University of Wisconsin-Madison, 2022
- On the Lack of fusion porosity in L-PBF processes — Politecnico di Milano, 2022
- Laser powder-bed fusion additive manufacturing: Physics of complex melt flow and formation mechanisms of pores, spatter, and denudation zones — Lawrence Livermore National Laboratory, 2016
- Residual Heat Effect on the Melt Pool Geometry during the Laser Powder Bed Fusion Process — University of Louisville, 2022
- Detection of Process Failures in Layerwise Laser Melting with Optical Process Monitoring — Concept Laser GmbH, 2012
- Melt pool boundary extraction and its width prediction from infrared images in selective laser melting — Chongqing Institute of Green and Intelligent Technology, Chinese Academy of Sciences, 2019
- Melt pool temperature and cooling rates in laser powder bed fusion — Imperial College London, 2018
- Melt pool monitoring in laser beam melting with two-wavelength holographic imaging — ONERA, Université Paris Saclay, 2022
- Determining process stability of Laser Powder Bed Fusion using pyrometry — inspire AG, 2020
- In situ X-ray imaging of defect and molten pool dynamics in laser additive manufacturing — Diamond Light Source Ltd, 2018
- In situ melt pool measurements for laser powder bed fusion using multi sensing and correlation analysis — Virginia Tech, 2022
- Effects of Particle Size Distribution with Efficient Packing on Powder Flowability and Selective Laser Melting Process — University of Wisconsin-Madison, 2022
- Simulation of the Evolution of Thermal Dynamics during Selective Laser Melting and Experimental Verification Using Online Monitoring — Xi'an University, 2020
- Melt pool monitoring for laser beam melting of metals: assistance for material qualification for the stainless steel 1.4057 — Friedrich-Alexander-Universität Erlangen-Nürnberg, 2018
- Hardness Prediction of Laser Powder Bed Fusion Product Based on Melt Pool Radiation Intensity — Air Force Engineering University, 2022
- Feedback Control of Melt Pool Area in Selective Laser Melting Additive Manufacturing Process — Nottingham Trent University, 2021
- Method for melt pool monitoring using geometric length (US) — General Electric Company, 2021
- Method for melt pool monitoring using geometric length (EP) — General Electric Company, 2024
- Simulating melt pool characteristics for selective laser melting additive manufacturing — Robert Bosch GmbH, 2020
- Feedback control of Layerwise Laser Melting using optical sensors — CONCEPT Laser GmbH, 2010
- An improved methodology of melt pool monitoring of direct energy deposition processes — TWI, 2020
- Predictions of in-situ melt pool geometric signatures via machine learning techniques for laser metal deposition — CSIRO Manufacturing, 2022
- NIST — Additive Manufacturing Standards and Metrology
- Lawrence Livermore National Laboratory — Advanced Manufacturing Program
- Diamond Light Source — Synchrotron Science Facility
All data and statistics on this page are sourced from the references above and from PatSnap's proprietary innovation intelligence platform.
PatSnap Eureka searches patents and research to answer instantly.