A Causal Auditing Paradigm for Sub-health Diagnosis of Offshore Wind Assets via Multi-terminal Harmonic Fingerprinting

Project Identity: Shinar of Clark
Author: Yi Zeng
Framework: Causal Auditing for Offshore Wind Sub-health Diagnosis
📄 Manuscript PDF: Read the Full Paper Here
DOI: https://doi.org/10.5281/zenodo.20149720
github:https://github.com/Shinar-of-Clark/Clark-Paradigm-Initiative
Email:Clark@ShinarOfClark.com

📖 Introduction

This repository implements the sub-health diagnostic framework for offshore wind assets under the “Clark Paradigm”.

Unlike traditional black-box prediction models, this project establishes an “Electromagnetic Ledger” system. By conducting a real-time audit of the 1st-20th order harmonic fingerprints across the generator, converter, tower-base switchgear, and remote collector cables, it achieves a paradigm shift from “stochastic state prediction” to “deterministic causal auditing”.

Core Philosophy: Anomaly = Physical Reality – Causal Expectation (Δ=Physical Reality−Causal Expectation).

📁 Repository Structure

To ensure the transparency and reproducibility of our research, the supporting materials for this paper are organized as follows:

  • figures/: Contains all high-resolution figures, architectural diagrams, and result charts presented in the manuscript.
  • data/: Contains the corresponding datasets (raw and processed) used to generate each figure. Researchers can use this data to independently verify the causal auditing results.

🛠️ Technical Architecture

This platform adopts a distributed hardware topology, decoupling the perception layer from the audit layer:

  • Perception Layer (Field Layer): MCU nodes deployed inside the turbine and at the substation inlet execute high-frequency sampling at 10.24 kHz and local FFT extraction.
  • Audit Layer: Edge hosts perform feature decoupling and causal mapping via a Recursive Least Squares (RLS) engine and Multi-Head Attention mechanisms.

🚀 Key Features

  • Causal Auditing Engine: Utilizes the RLS operator to dynamically track the dielectric drift of assets, superseding simple threshold-based alarms.
  • Frequency-Selective Sovereignty:
    • 5th/7th Harmonics: Locks onto the health state of core power devices within full-scale converters. Precisely captures switching characteristic degradation induced by IGBT junction temperature elevation, dead-time drift, or gate drive delays, while also serving as an early diagnostic indicator for the capacitance decay of AC/DC-side filtering capacitors.
    • 11th/13th Harmonics: Probes the impedance evolution trajectory of long-distance subsea collection cables. Deeply analyzes distributed capacitance/inductance drift and high-frequency resonance point shifts triggered by “Water Treeing”, microscopic insulation dampness, and physical strain of cables caused by subsea currents.
    • 17th+ Harmonics: Captures feature shifts induced by transformer winding mechanical loosening or magnetic circuit saturation. Furthermore, prolonged spectral imbalance in this range imposes cumulative dielectric stress, serving as a latent indicator for potential partial discharge (PD) activities (while limited by frequency range to capture transient PD pulses directly, it serves as a lateral reflection of insulation aging trends).
  • Ultra-low Latency: The single-iteration computation time for the 21-dimensional causal vector is < 5ms, fully meeting the real-time requirements of 50Hz power infrastructure.
  • Game Theory Decision Optimization: Introduces a Nash Equilibrium model to balance generation revenue, failure risks, and O&M costs, maximizing residual asset value.

📊 Performance

Under extreme aerodynamic turbulence (e.g., sudden wind speed shifts), the error of traditional LSTM models may surge to 25%. In contrast, this framework, through rapid self-calibration, strictly restricts the audit residual Δ to within 5%.

📚 Citation

If you utilize the concepts or content of this project in your research, please cite our preprint paper:

APA Format:

Zeng, Y. (2026). A Causal Auditing Paradigm for Sub-health Diagnosis of Offshore Wind Assets via Multi-terminal Harmonic Fingerprinting (V1.0.0). Zenodo. https://doi.org/10.5281/zenodo.20149720

BibTeX:

@article{zeng2026clark,
  title={A Causal Auditing Paradigm for Sub-health Diagnosis of Offshore Wind Assets via Multi-terminal Harmonic Fingerprinting (V1.0.0)},
  author={Yi Zeng},
  year={2026},
  journal={Zenodo},
  doi={10.5281/zenodo.20149720}
}

🛡️ License

This project is licensed under the Creative Commons Attribution 4.0 International (CC BY 4.0) License.

License: CC BY 4.0

Rights Statement: You are free to share and adapt this work, provided that you give appropriate credit to the author Yi Zeng (Project Shinar of Clark) and indicate if changes were made.

滚动至顶部