A Novel Assessment of Delayed Neutron Detector Data in CANDU Reactors
Bibliographic record
Abstract
Abstract A common opportunity for nuclear power plant operators is ensuring that routinely collected data are fully leveraged. Exploiting data analytics can enable improvements in anomaly detection and condition monitoring by identifying previously unseen data trends and correlations without major financial investment. One such opportunity is in facilitating the detection of fuel defects by augmenting the delayed neutron (DN) monitoring system deployed in the majority of Canada deuterium uranium (CANDU) reactors. In this paper, we demonstrate using archive data that the detection of fuel defects can be accelerated using this system in combination with the use of a deeper historical dataset and the introduction of a smoothing algorithm. The current defect identification process relies on the analysis of data of high variance and is subject to the judgment of a domain expert, resulting in variable defect identification periods. The proposed approaches seek to mitigate this and alleviate the variable identification time. Initial results presented here show that for an initial batch of 30 defects, identification periods can be meaningfully reduced compared to the current process, with defects potentially visible on an average of 11.4 days earlier. By shortening this identification period, fuel containing defects can be scheduled for earlier removal, reducing the risk of statutory shutdown obligations, protecting personnel, and promoting industry best practice. Exploring a historical dataset identifies previously undocumented trends and we discuss the potential to produce correlations with other reactor parameters. The application of this knowledge can lead to opportunities in the use of machine learning algorithms and, ultimately, more accurate predictions.
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 | 0.002 | 0.004 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 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".