📚 In-Situ Monitoring and Adaptive Control in Laser-based Additive Manufacturing

Welcome to this curated repository of research papers with open-source codes/dataset on in‑situ process monitoring and adaptive process control in laser-based additive manufacturing (AM). The goal is to highlight state-of-the-art approaches for ensuring zero-defect, autonomous AM through real-time sensing, data analysis, and control. We focus primarily on Laser Powder Bed Fusion (LPBF) and Laser Directed Energy Deposition (LDED), with selective inclusion of Wire Arc AM (WAAM) and other processes where relevant. Contributions are welcome – see the guidelines below! 😊

Details of our comprehensive review paper published in 2024: In-situ process monitoring and adaptive quality enhancement in laser additive manufacturing: a critical review (2024)

Table of Contents

• In-Situ Process Monitoring
  • Acoustic Signal-Based Monitoring
  • Vision-Based Monitoring
  • Thermal Infrared Monitoring
  • Surface Monitoring and Point Cloud Processing
  • Multisensor & Data Fusion
• Adaptive Process Control
• Simulation for Process Optimization
• Physics-Informed Machine Learning
• AM Robotic Path Planning
• AI-assisted Design of Advanced Materials
• Design and Optimization Tools
• Process-Structure-Property Datasets
• Contributing
• License

---

In-Situ Process Monitoring 🔎

Real-time monitoring is crucial for defect detection and quality assurance. Below, we categorize monitoring approaches by the primary sensing modality:

Acoustic Signal-Based Monitoring 📢

High-frequency acoustic emissions can reveal melt pool instabilities and defect formation. Researchers are using microphones or piezoelectric sensors to capture sound from the process and applying ML for anomaly detection.

Title (and Link)	Code 💻	Dataset 📂	Process	Tags	Year
Harmonizing Sound and Light: X-ray Imaging Unveils Acoustic Signatures of Stochastic Inter-regime Instabilities During Laser Melting M. Hamidi Nasab et al. (Nature Communications)	GitLab / GitHub	N/A	LPBF	X-rayImaging AcousticSignatures MeltPoolDynamics SignalSegmentation	2023
Optimizing In‑Situ Monitoring for LPBF: Deciphering Acoustic Emission and Sensor Sensitivity with Explainable ML V. Pandiyan et al. (J. Mat. Proc. Tech.)	GitHub	Dataset	LPBF	ExplainableML EMD FrequencyAnalysis	2023
In-Situ Alloying of Titanium-Fe: Acoustic Dynamics and Process Signatures V. Pandiyan et al.	GitHub	N/A	LPBF	In-situAlloying AcousticFingerprinting Ti-Fe	2023
Deep Transfer Learning of AM Mechanisms Across Materials in LPBF V. Pandiyan et al. (J. Mat. Proc. Tech.)	GitHub	N.A.	LPBF	TransferLearning PyTorch CrossMaterial	2022
Linking Acoustic Emission Signatures to Material Properties in AM V. Pandiyan et al. (Virtual Phys. Prototyp.)	GitHub	N/A	LPBF	MaterialFingerprinting FeatureSelection	2022
Self-Supervised Bayesian Representation Learning for Acoustic Monitoring in AM V. Pandiyan et al. (Materials & Design)	GitHub	N/A	LPBF	SelfSupervised BayesianLearning Representation	2022
Domain Adaptation for Bridging Dissimilar Process Maps Using Acoustic Signals V. Pandiyan et al. (Additive Manufacturing)	GitHub	N/A	LPBF	DomainAdaptation ProcessMaps CrossMaterial	2022
Feature Engineering for AM Acoustic Emission Monitoring V. Pandiyan et al. (Procedia CIRP)	GitHub	N/A	LPBF	FeatureEngineering WaveletTransform RealTime	2021
Semi-supervised Monitoring of LPBF Based on Acoustic Emissions V. Pandiyan et al. (Virtual Phys. Prototyp.)	GitHub	N/A	LPBF	SemiSupervised PyTorch AnomalyDetection	2021

Vision-Based Monitoring 📷

Optical monitoring uses cameras (visible or infrared) to observe the melt pool, spatter, and layer morphology in real-time. High-speed imaging and computer vision techniques can detect anomalies like pores or lack-of-fusion. Some works also release open datasets for benchmarking.

Title (and Link)	Code 💻	Dataset 📂	Process	Tags	Year
Co-Registered In-Situ and Ex-Situ Dataset from a Laser Powder Bed Fusion AM Process ORNL Peregrine Dataset	N/A	ORNL Peregrine v2023-10	LPBF	CoRegistered InSituExSitu Correlation	2023
Co-Registered In-Situ and Ex-Situ Tensile Properties Dataset from a Laser Powder Bed Fusion AM Process ORNL Peregrine Dataset	N/A	ORNL Peregrine v2023-11	LPBF	TensileProperties ProcessMonitoring PropertyPrediction	2023
Co-Registered In-Situ and Ex-Situ Dataset from an Electron Beam Powder Bed Fusion AM Process ORNL Peregrine Dataset	N/A	ORNL Peregrine v2023-09	EBPBF	ElectronBeam InSituMonitoring ProcessCorrelation	2023
Layer-wise Imaging Dataset from Powder Bed AM for Machine Learning Applications ORNL Peregrine Dataset	N/A	ORNL Peregrine v2022-10	LPBF	LayerWiseImaging MachineLearning Dataset	2022
Layer-wise Anomaly Detection and Classification for Powder Bed AM: Machine-Agnostic Algorithm for Real-Time Pixel-wise Semantic Segmentation L. Scime et al. (Additive Manufacturing)	N/A	ORNL Peregrine v2021-03	LPBF	SemanticSegmentation AnomalyDetection RealTime Peregrine	2020
RAISE-LPBF: Reference Dataset & Benchmark for Reconstructing Laser Parameters from On-Axis Video C. Blanc et al.	GitHub	Website	LPBF	HighSpeedImaging LaserParameters Benchmark	2023
Dataset of In‑Situ Coaxial Monitoring and Print's Cross-Section Images J. Akhavan et al. (Sci. Data)	N/A	Data	LDED	CoaxialVision CrossSection OpenAccess	2023
Real-Time Monitoring & Quality Assurance for L-DED via Coaxial Imaging + Self-Supervised Learning V. Pandiyan et al. (J. Intell. Manuf.)	GitHub	N/A	LDED	SelfSupervised CoaxialImaging QualityAssurance	2023
Manifold Learning for Process Zone Characterization in DED V. Pandiyan et al. (Manufacturing Letters)	GitHub	N/A	DED	ManifoldLearning Dimensionality ProcessZone	2022
Contrastive Learning for DED Process Monitoring V. Pandiyan et al. (Journal of Manufacturing Processes)	GitHub	N/A	DED	ContrastiveLearning SelfSupervised ProcessAnalysis	2022
Deep DIC: Deep Learning-Based Digital Image Correlation for End-to-End Displacement and Strain Measurement R. Yang et al. (J. Mat. Proc. Tech.)	GitHub	Drive	Various	DIC DeepLearning StrainMeasurement	2022

Thermal Infrared Monitoring 🌡️

Infrared cameras and pyrometers capture thermal signatures (melt pool temperature, cooling rates, etc.) for defect detection. Thermal monitoring can reveal anomalies like overheating, lack of fusion, or excessive cooling that correlate with defects.

Title (and Link)	Code 💻	Dataset 📂	Process	Tags	Year
In‑Situ Infrared Camera Monitoring for Defect & Anomaly Detection S. Hinnebusch et al. (arXiv preprint)	Python snippets	N/A	LPBF	FeatureExtraction AnomalyDetection ThermalSignatures	2024

Surface Monitoring and Point Cloud Processing 🛰️

Structured-light scanning, fringe projection, and 3D reconstruction techniques monitor the geometry of each layer. These produce point clouds or height maps to detect warping, recoater interference, or surface roughness issues.

Title (and Link)	Code 💻	Dataset 📂	Process	Tags	Year
Etna: On-line 3D Monitoring for Robotized LMD AIMEN OpenLMD Toolkit (Open-Source Project)	GitHub	N/A	LDED	3DScan RealTime Robotized	2017

Multisensor & Data Fusion 🧩

Combining multiple sensors (e.g., optical, acoustic, thermal) can provide a richer picture of the process. Multisensor approaches often leverage machine learning to fuse data and detect defects more reliably, sometimes using ground truth from X-ray or CT for validation.

Title (and Link)	Code 💻	Dataset 📂	Process	Tags	Year
Audio-Visual Cross-Modality Knowledge Transfer for ML-Based In-Situ Monitoring in Laser AM Xie et al. (Additive Manufacturing)	N/A	Zenodo	LDED	CrossModality AudioVisual KnowledgeTransfer	2025
Deep Learning Monitoring of LPBF with Heterogeneous Sensing and X-Ray Guidance V. Pandiyan et al. (Addit. Manuf.)	GitHub	N.A.	LPBF	Multisensor XRay VariableTimeScales	2022
Multi-Material Composition Monitoring Using Sensor Fusion in LPBF V. Pandiyan et al.	GitHub	N/A	LPBF	SensorFusion MultiMaterial CompositionMonitoring	2021

Adaptive Process Control 🕹️

Adaptive control closes the loop by adjusting process parameters in real-time based on sensor feedback. The aim is to correct defects on-the-fly – for example, modulating laser power or scan speed to maintain a stable melt pool and avoid flaws, driving AM toward autonomous operation.

Title (and Link)	Code 💻	Dataset 📂	Process	Tags	Year
CladPlus: Closed-Loop Control for Laser Cladding AIMEN OpenLMD Toolkit (Open-Source Project), OpenLMD, an open source middleware and toolkit (García-Díaz et al. RCIM)	GitHub	N/A	LDED	ClosedLoop LaserCladding InfraredControl	2017

Physics-based Simulations 🖥️

Physics-based simulations (thermal, mechanical, microstructural) help predict outcomes of process parameters and optimize them without costly trial-and-error. Open-source tools and multi-physics models are enabling "digital twins" of AM processes.

Title (and Link)	Code 💻	Dataset 📂	Process	Tags	Year
JAX-AM: Differentiable Simulation Toolkit for Additive Manufacturing T. Xue et al. (Open-Source Project)	GitHub	N/A	Multi (LPBF/DED)	Differentiable GPU-Accelerated JAX	2023
Efficient GPU-accelerated Thermomechanical Solver for Residual Stress Prediction in AM S. Liao et al. (Computational Mechanics)	GitHub	N/A	LDED/LPBF	GPU-Accelerated FEM ResidualStress CuPy	2023
Graph-Based Modeling for Additive Manufacturing AMPL Research Group	GitHub	N/A	LPBF, LDED	GraphModels GNN ProcessModeling	2023
An iterative machine learning framework for predicting temperature profiles Arindam Paul et al.	GitHub	N/A	LPBF	IterativePrediction BuildQuality XCT	2022
Mechanistic Data-Driven Prediction of As-Built Mechanical Properties in Metal Additive Manufacturing X. Xie et al. (npj Computational Materials)	GitHub	GitHub	DED	ThermalHistory CNN MechanicalProperties	2021

Physics-Informed Machine Learning 🧠

Physics-informed ML integrates fundamental physics (conservation laws, constitutive models, etc.) into data-driven models. In AM, this can mean using simulation data to train ML models, constraining neural nets with physical laws, or creating surrogate models that are faster than physics simulators but more accurate than black-box ML alone.

Title (and Link)	Code 💻	Dataset 📂	Process	Tags	Year
A Physics-Informed CNN with Custom Loss Functions for Porosity Prediction in LMD Erin McGowan. et al (Sensors)	N/A	Harvard Dataverse	LMD	CustomLoss CNN Porosity	2022
Differentiable Physics Simulation of Dynamics-aware Metal Powder-bed Fusion Additive Manufacturing Components M. Mozaffar et al.	GitHub	N/A	LPBF	DifferentiableSimulation AutoDiff Optimization	2022
Physics-Informed Neural Networks for Multiphysics Data Assimilation with Application to AM S. Liao et al. (Computational Mechanics)	GitHub	Drive	LDED/LPBF	PINN ThermalMechanical DataAssimilation	2022

AM Robotic Path Planning 🤖

Path planning is crucial for directed-energy AM processes (LDED, WAAM, etc.), especially with multi-axis robots. Efficient algorithms are needed to determine the deposition toolpath that ensures uniform material deposition, avoids collisions, and can handle complex geometries or repairs.

Title (and Link)	Code 💻	Dataset 📂	Process	Tags	Year
RobPath: Off‑Line Path Planning Tool for LMD Automation AIMEN OpenLMD Toolkit (Open-Source Project)	GitHub	N/A	LMD	ToolpathGeneration Robotics CAD-Based	2020
Toolpath Design for AM Using Reinforcement Learning M. Mozaffar et al.	GitHub	N/A	FDM	ReinforcementLearning ToolpathOptimization	2022

Design and Optimization Tools 🛠️

Title (and Link)	Code 💻	Dataset 📂	Process	Tags	Year
A Flexible and Easy-to-Use Open-Source Tool for Designing Functionally Graded 3D Porous Structures N. Korshunova et al. (Virtual and Physical Prototyping)	GitHub	N/A	Various	TPMS LatticeStructures FunctionallyGraded	2022

Process-Structure-Property Datasets 📊

Title (and Link)	Code 💻	Dataset 📂	Process	Tags	Year
Machine Learning for Knowledge Transfer Across Multiple Metals Additive Manufacturing Printers J. Gaikwad et al. (Additive Manufacturing)	N/A	Citrination	LPBF	KnowledgeTransfer PSPRelationship MultiPrinter	2021
New Insight into the Multivariate Relationships Among Process, Structure, and Properties in Laser Powder Bed Fusion AlSi10Mg Q. Luo et al. (Additive Manufacturing)	N/A	Available in Paper	LPBF	PSPRelationship AlSi10Mg MultivariateAnalysis	2023
Effect of Processing Parameters on Pore Structures, Grain Features, and Mechanical Properties in Ti-6Al-4V by LPBF Q. Luo et al. (Additive Manufacturing)	N/A	Zenodo	LPBF	Ti6Al4V ProcessParameters MechanicalProperties	2022
Dataset of Process-Structure-Property Feature Relationship for LPBF Ti-6Al-4V Material Q. Luo et al. (Data in Brief)	N/A	Zenodo	LPBF	Ti6Al4V PSPRelationship FeatureData	2023

---

🤝 Contributing

We welcome contributions from the community! Please feel free to suggest a paper or add a new entry via Pull Request. To maintain consistency, please follow these guidelines:

• Scope: Ensure the paper is relevant to in-situ monitoring or adaptive control in AM (or closely related, like simulation or hybrid processes). Laser-based metal AM is the focus, but notable work in other AM or sensing modalities can be included if it fits a category.
• Table Format: Use the same table structure: Title (with link to paper), Code (link if available), Dataset (link if available), Process type, Tags, and Year. Indicate N/A if code or data are not available.
• Tags: Use descriptive tags in backticks like `ExplainableML`. Aim for 2-4 tags that capture the work's key aspects, focusing on methods, algorithms, or specific techniques rather than general categories already covered by the section.
• Style: Keep descriptions concise. You can add a short note below a table if needed to explain context or significance (as done above). Use present tense and clear language.

---

Happy Printing & Processing! 🚀 Feel free to open an issue for any questions or discussions.