Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A SISA-based Machine Unlearning Framework for Power Transformer Inter-Turn Short-Circuit Fault Localization

Nanhong Liu, Jingyi Yan, Mucun Sun, Jie Zhang

TL;DR

A SISA (Sharded, Isolated, Sliced, and Aggregated)-based machine unlearning (MU) framework for power transformer inter-turn short-circuit fault (ITSCF) localization that achieves almost identical diagnostic accuracy to full retraining, while reducing retraining time significantly.

Abstract

In practical data-driven applications on electrical equipment fault diagnosis, training data can be poisoned by sensor failures, which can severely degrade the performance of machine learning (ML) models. However, once the ML model has been trained, removing the influence of such harmful data is challenging, as full retraining is both computationally intensive and time-consuming. To address this challenge, this paper proposes a SISA (Sharded, Isolated, Sliced, and Aggregated)-based machine unlearning (MU) framework for power transformer inter-turn short-circuit fault (ITSCF) localization. The SISA method partitions the training data into shards and slices, ensuring that the influence of each data point is isolated within specific constituent models through independent training. When poisoned data are detected, only the affected shards are retrained, avoiding retraining the entire model from scratch. Experiments on simulated ITSCF conditions demonstrate that the proposed framework achieves almost identical diagnostic accuracy to full retraining, while reducing retraining time significantly.

A SISA-based Machine Unlearning Framework for Power Transformer Inter-Turn Short-Circuit Fault Localization

TL;DR

A SISA (Sharded, Isolated, Sliced, and Aggregated)-based machine unlearning (MU) framework for power transformer inter-turn short-circuit fault (ITSCF) localization that achieves almost identical diagnostic accuracy to full retraining, while reducing retraining time significantly.

Abstract

In practical data-driven applications on electrical equipment fault diagnosis, training data can be poisoned by sensor failures, which can severely degrade the performance of machine learning (ML) models. However, once the ML model has been trained, removing the influence of such harmful data is challenging, as full retraining is both computationally intensive and time-consuming. To address this challenge, this paper proposes a SISA (Sharded, Isolated, Sliced, and Aggregated)-based machine unlearning (MU) framework for power transformer inter-turn short-circuit fault (ITSCF) localization. The SISA method partitions the training data into shards and slices, ensuring that the influence of each data point is isolated within specific constituent models through independent training. When poisoned data are detected, only the affected shards are retrained, avoiding retraining the entire model from scratch. Experiments on simulated ITSCF conditions demonstrate that the proposed framework achieves almost identical diagnostic accuracy to full retraining, while reducing retraining time significantly.
Paper Structure (12 sections, 3 equations, 4 figures)

This paper contains 12 sections, 3 equations, 4 figures.

Table of Contents

  1. Introduction
  2. ITSCF Data Acquisition
  3. ITSCF Simulation
  4. Sensor Failures
  5. SISA Framework
  6. ML and MU Algorithms
  7. SISA Framework for ITSCF Localization
  8. Results Analysis
  9. ITSCF Data Processing
  10. ML Model Structure and Hyperparameters
  11. Experimental Results Analysis
  12. Conclusion

Figures (4)

  • Figure 1: The SISA framework for power transformer ITSCF localization.
  • Figure 2: Training accuracy and time on different numbers of shards, i.e., $S$. If $S=1$, it is a non-SISA model, and the training process removing poisoned data is full retraining.
  • Figure 3: Comparison of confusion matrices for four cases under 1 ITSCF condition with poisoned data caused by sensor failures.
  • Figure 4: Comparison of confusion matrices for four cases under 6 ITSCF conditions with poisoned data caused by sensor failures.