Consequence-driven optimization for detection and localization of stealth false data injection attacks against state estimation in power distribution systems
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Monitoring modern power distribution systems through state estimation (SE) is crucial for optimizing grid operation and ensuring reliability. However, SE is vulnerable to stealth false data injection attacks (FDIAs). Stealth FDIAs can evade conventional bad data detection algorithms, leading to operator misjudgments and erroneous decisions. FDIA misidentifications, i.e., false alarms and undetected FDIAs, have distinct monetary and operational consequences. Even within each type of misidentification, these consequences can vary based on factors like meter location, customer type, and sustained energy not supplied (ENS). This paper, therefore, proposes consequence-driven cost functions to quantify the monetary impact of FDIA misidentifications in the SE. The proposed method explicitly accounts for system topology, customer type, and ENS. The proposed approach is model-agnostic and can operate with any anomaly detection method. We use an autoencoder (AE) as a sample anomaly detection method to illustrate the proposed consequence-driven framework. The AE is trained on FDIA-free data to reconstruct normal meter behavior. Deviations are then passed to the largest normalized residual (LNR) test for detection and localization, enabling a detailed evaluation of FDIA misidentification costs. Additionally, an optimization formulation is introduced to adjust the LNR thresholds for each meter, minimizing the total misidentification cost. Simulations use IEEE 13-bus and 123-bus test feeders. Results show that optimal thresholds can reduce FDIA misidentification costs by up to 66 %. This offers a consequence-driven alternative to the accuracy-based metrics commonly used in the literature. It also provides a better fit for the complex, cyber-physical nature of power systems.
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.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 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 it