Additive Manufacturing Process Modeling and Evaluation

Additive manufacturing process modelling and evaluation encompasses a suite of computational and experimental strategies aimed at predicting, controlling and verifying the layer-by-layer fabrication of components. Physically based simulations—such as finite-element thermal and fluid dynamic models—capture heat transfer, melt-pool evolution and solidification kinetics, while data-driven frameworks leverage machine learning to extract key features from high-fidelity datasets and in situ measurements. Surrogate modelling and uncertainty quantification accelerate design iterations by balancing accuracy and computational cost, enabling real-time process monitoring and adaptive control. Evaluation metrics span from melt-pool geometry and porosity to microstructural characteristics and as-built mechanical properties, supporting qualification pathways in aerospace, biomedical implants and precision engineering. By integrating multiscale physics with advanced analytics, modern approaches deliver robust predictions of part quality, inform parameter optimisation and foster standardised validation protocols for broad industrial uptake.

Research from Nature Portfolio

No recent Nature Portfolio content available.

Research from all publishers

Recent work has introduced a factorial design analytics framework for laser powder-bed fusion, systematically exploring the interactions of material parameters in multiphysics simulations. By applying full factorial designs and LASSO-based variable selection, researchers have identified critical parameter combinations that govern thermal-fluid responses and validated findings against physics-based checkpoints, thereby improving predictive reliability under uncertainty.

A novel generative adversarial network architecture has been developed for melt-pool classification and image synthesis in laser powder-bed fusion. The model simultaneously classifies laser power, scan speed and direction across hundreds of classes with high accuracy and generates realistic melt-pool images for data augmentation. This approach enhances real-time monitoring capabilities and supports offline process optimisation by producing large, labelled datasets.

In directed energy deposition of Inconel 718, a hybrid finite‐element and convolutional neural network framework has been applied to predict spatially varying thermal histories and correlate cooling rates with microstructure and tensile properties. Automated feature extraction from simulated temperature sequences enables accurate mapping of process-structure-property relations, offering a pathway to data-driven design of alloy microstructures and mechanical performance.

Additive Manufacturing Process Modeling and Evaluation publication trend

The graph below shows the total number of articles in additive manufacturing process modeling and evaluation across all publications each year (not limited to Nature Index journals).

Technical terms

Additive Manufacturing (AM): Layer-wise fabrication of parts by depositing and consolidating materials under computer control.

Multiphysics modelling: Coupled simulation of interacting physical phenomena (e.g. heat transfer, fluid flow and phase change) in a single computational framework.

Surrogate model: Simplified computational model trained to approximate a high-fidelity simulation at reduced cost.

Convolutional Neural Network (CNN): Deep learning architecture that automatically extracts spatial features from sequential or grid-structured data.

Melt pool: Localised molten region created by an energy source during metal additive manufacturing, critical for defining microstructure and defects.

References

1. Factorial design analytics on effects of material parameter uncertainties in multiphysics modeling of additive manufacturing. npj Computational Materials (2023).
2. MeltPoolGAN: Auxiliary Classifier Generative Adversarial Network for melt pool classification and generation of laser power, scan speed and scan direction in Laser Powder Bed Fusion. Additive Manufacturing (2023).