Reproducibility data and code for auditable wind turbine early fault evidence promotion

It retains the original simulation and processed field evidence and adds three revision analyses: controlled availability selection at fixed within-role availability, a bridge to the original trailing-three-capture confirmation window, and threshold sensitivity with healthy verification burden and verifier utilization.

Fine-grained masks, counts, scores, frozen configurations, paired seed summaries, main and supplementary figure/table source data, and portable execution and independent numerical audit scripts are included. The included results are numerical simulation outputs and derived observational/mechanism evidence, not new physical measurements. They do not establish general field automatic-warning efficacy or acceptable maintenance cost.