Detecting the extent of co-existing anomalies in additively manufactured metal matrix composites through explainable selection and fusion of multi-camera deep learning features
Bibliographic record
Abstract
Process development for customised additively manufactured materials is challenging and labour-intensive. Advanced in-situ monitoring coupled with modern machine learning (ML) methods can expedite defect detection and qualification of additive manufacturing (AM) parts. Directed energy deposition (DED) processes offer flexibility to deposit material on existing complex parts for hybrid manufacturing and repairs. DED enables custom metal matrix composites (MMCs) like nickel tungsten carbide (Ni-WC) overlays on ferrous mining tools for enhanced wear resistance. However, co-existing anomalies specific to defects in the matrix, reinforcement and their interaction present development challenges. The challenge is compounded since the co-existing anomalies can exist in varying extents (e.g. absent, low, high). This study investigates dual mid-wave infrared (MWIR) cameras (FLIR and CLAMIR) for defect extent detection in Ni-WC MMCs. Deep learning features extracted with a fine-tuned vision transformer outperformed conventional methods by improving anomaly separability and revealing process-regime-aware feature distributions. Explainable artificial intelligence identified key MWIR features detecting six defect categories. Data ablation revealed FLIR’s superior accuracy and generalisability under noise, while CLAMIR demonstrated robustness under instability. Explainable fusion enabled effective selection of camera features. Our work provides a foundation for ML-assisted development of AM-based Ni-WC and similar MMCs by facilitating in-situ detection of co-existing anomalies.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".