What does this research mean for the field?

An uncertainty-aware deep ensemble framework integrating multimodal in-situ sensing improves calibration, defect prediction reliability, and reduces intervention costs in metal additive manufacturing compared to deterministic approaches. Novelty: ClaimNovelty.METHODOLOGICAL. Consensus alignment: ConsensusAlignment.NEUTRAL.

What question did this study set out to answer?

To develop an uncertainty-aware deep ensemble framework that integrates multimodal in-situ sensing to predict layer-wise porosity with calibrated confidence for real-time operator intervention.

September 14, 2026Nondestructive Testing And Evaluation

Uncertainty-aware deep ensembles for real-time porosity prediction in metal additive manufacturing with calibrated confidence for operator decision support

View Full Paper

Ask AI

Bookmark

Authors

TST. Suresh SMS. A. Sahaaya Arul MaryChrist University SKSumanth Ratna KandavalliNew York University

Discussion

Member takes

Implication

Validation study demonstrates calibrated deep ensembles accurately predict porosity in metal additive manufacturing, highlighting reliable real-time operator decision support.

Key Points

To develop an uncertainty-aware deep ensemble framework that integrates multimodal in-situ sensing to predict layer-wise porosity with calibrated confidence for real-time operator intervention.
Integrated in-situ thermal imaging, acoustic emission, and spectral monitoring via a shared CNN–LSTM architecture for simultaneous defect classification and porosity regression.
Decomposed predictive uncertainty into aleatoric and epistemic components using the law of total variance, calibrated confidence via temperature and variance scaling, and implemented an uncertainty-gated decision policy.
Validated performance using micro-CT ground truth on NIST AMMT, AM-Bench 2022, and multi-material builds of Ti-6Al-4V and Inconel 718.
Demonstrated improved confidence calibration and defect prediction reliability over conventional deterministic monitoring approaches.
Reduced overall intervention costs by establishing an automated mapping from calibrated predictive distributions to operational actions, including continue, inspect, re-scan, and abort.

Cite This Study

Suresh et al. (2026) studied this question.

synapsesocial.com/papers/6aa7b2c70926e14a848b142f https://doi.org/10.1080/10589759.2026.2721360

View Full Paper

Ask AI

Bookmark

Implication

Validation study demonstrates calibrated deep ensembles accurately predict porosity in metal additive manufacturing, highlighting reliable real-time operator decision support.

Uncertainty-aware deep ensembles for real-time porosity prediction in metal additive manufacturing with calibrated confidence for operator decision support

Discussion

Member takes

Implication

Key Points

Cite This Study

Uncertainty-aware deep ensembles for real-time porosity prediction in metal additive manufacturing with calibrated confidence for operator decision support

Discussion

Member takes

Implication

Key Points

Cite This Study

Also Consider

Also Consider