Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Dynamic fault detection and diagnosis of industrial alkaline water electrolyzer process with variational Bayesian dictionary learning

Qi Zhang, Lei Xie, Weihua Xu, Hongye Su

TL;DR

This work tackles fault detection and diagnosis in industrial alkaline water electrolysis by addressing dynamic, noisy process data with a robust RDV DIB (RDVDL) framework. It combines sparse Bayesian dictionary learning to preserve dynamic structure with a low-rank vector autoregressive model to extract serial correlations under uncertainty, all inferred via variational Bayesian methods. Offline dictionary learning establishes a reconstruction-based monitoring baseline, while online stages compute dynamic and static fault statistics and diagnose faults via reconstruction-based contributions. The approach is validated on an industrial hydrogen production process, showing improved fault detection sensitivity and interpretable variable-level fault localization, with potential to enhance safety and reliability in AWE operations.

Abstract

Alkaline Water Electrolysis (AWE) is one of the simplest green hydrogen production method using renewable energy. AWE system typically yields process variables that are serially correlated and contaminated by measurement uncertainty. A novel robust dynamic variational Bayesian dictionary learning (RDVDL) monitoring approach is proposed to improve the reliability and safety of AWE operation. RDVDL employs a sparse Bayesian dictionary learning to preserve the dynamic mechanism information of AWE process which allows the easy interpretation of fault detection results. To improve the robustness to measurement uncertainty, a low-rank vector autoregressive (VAR) method is derived to reliably extract the serial correlation from process variables. The effectiveness of the proposed approach is demonstrated with an industrial hydrogen production process, and RDVDL can efficiently detect and diagnose critical AWE faults.

Dynamic fault detection and diagnosis of industrial alkaline water electrolyzer process with variational Bayesian dictionary learning

TL;DR

This work tackles fault detection and diagnosis in industrial alkaline water electrolysis by addressing dynamic, noisy process data with a robust RDV DIB (RDVDL) framework. It combines sparse Bayesian dictionary learning to preserve dynamic structure with a low-rank vector autoregressive model to extract serial correlations under uncertainty, all inferred via variational Bayesian methods. Offline dictionary learning establishes a reconstruction-based monitoring baseline, while online stages compute dynamic and static fault statistics and diagnose faults via reconstruction-based contributions. The approach is validated on an industrial hydrogen production process, showing improved fault detection sensitivity and interpretable variable-level fault localization, with potential to enhance safety and reliability in AWE operations.

Abstract

Alkaline Water Electrolysis (AWE) is one of the simplest green hydrogen production method using renewable energy. AWE system typically yields process variables that are serially correlated and contaminated by measurement uncertainty. A novel robust dynamic variational Bayesian dictionary learning (RDVDL) monitoring approach is proposed to improve the reliability and safety of AWE operation. RDVDL employs a sparse Bayesian dictionary learning to preserve the dynamic mechanism information of AWE process which allows the easy interpretation of fault detection results. To improve the robustness to measurement uncertainty, a low-rank vector autoregressive (VAR) method is derived to reliably extract the serial correlation from process variables. The effectiveness of the proposed approach is demonstrated with an industrial hydrogen production process, and RDVDL can efficiently detect and diagnose critical AWE faults.
Paper Structure (25 sections, 40 equations, 8 figures, 3 tables, 1 algorithm)

This paper contains 25 sections, 40 equations, 8 figures, 3 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. AWE hydrogen production
  3. AWE hydrogen production system
  4. Electrolytic module
  5. Gas-liquid separation module
  6. Circulating module
  7. Summary of different type of faults in AWE electrolyzer.
  8. Data-driven process monitoring methods
  9. Review of PCA
  10. Review of Dynamic PCA
  11. Review of Dynamic-inner PCA
  12. Review of Dynamic-inner CCA
  13. Dynamic Variational Inference Dictionary Learning
  14. Bayesian Dictionary Learning
  15. Variational Bayesian Inference
  16. ...and 10 more sections

Figures (8)

  • Figure 1: Flow chart of alkaline electrolysis system.
  • Figure 2: The chemical reactions in the alkaline electrolyzer.
  • Figure 3: Fault detection result of fault 1.
  • Figure 4: Fault diagnosis result of fault 1.
  • Figure 5: Fault detection result of fault 2.
  • ...and 3 more figures