Evaluation of machine learning tools for inspection of steam generator tube structures using pulsed eddy current
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Degradation of nuclear steam generator (SG) tubes and support structures can result in a loss of reactor efficiency. Regular in-service inspection, by conventional eddy current testing (ECT), permits detection of cracks, measurement of wall loss, and identification of other SG tube degradation modes. However, ECT is challenged by overlapping degradation modes such as might occur for SG tube fretting accompanied by tube off-set within a corroding ferromagnetic support structure. Pulsed eddy current (PEC) is an emerging technology examined here for inspection of Alloy-800 SG tubes and associated carbon steel drilled support structures. Support structure hole size was varied to simulate uniform corrosion, while SG tube was off-set relative to hole axis. PEC measurements were performed using a single driver with an 8 pick-up coil configuration in the presence of flat-bottom rectangular frets as an overlapping degradation mode. A modified principal component analysis (MPCA) was performed on the time-voltage data in order to reduce data dimensionality. The MPCA scores were then used to train a support vector machine (SVM) that simultaneously targeted four independent parameters associated with; support structure hole size, tube off-centering in two dimensions and fret depth. The support vector machine was trained, tested, and validated on experimental data. Results were compared with a previously developed artificial neural network (ANN) trained on the same data. Estimates of tube position showed comparable results between the two machine learning tools. However, the ANN produced better estimates of hole inner diameter and fret depth. The better results from ANN analysis was attributed to challenges associated with the SVM when non-constant variance is present in the data.
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.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.002 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 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 it