Data-Optimized Contingency Screening: A Machine Learning Approach to Power System Security

SummaryAI generated

A study published in September 2026 proposed using machine learning algorithms to classify the safety levels of power systems, aiming to improve system stability and prevent large-scale failures. The study used Newton-Raphson method to extract data, combined with SMOTE oversampling and PCA dimensionality reduction techniques to address sample imbalance issues. The performance of three models—KNN, random forest (RF), and support vector machine (SVM)—was evaluated in IEEE-14 and IEEE-30 test systems. The evaluation results showed that the random forest achieved the highest F1 score of 0.97 in the IEEE-30 system. The study found that PCA contributed more to the overall model performance than SMOTE, while SMOTE could improve recall rate but might introduce false positives. This study confirmed that machine learning is an effective alternative to traditional contingency analysis, enabling real-time optimization of safety assessments.

Related eventsRELATED EVENTS

2026-09-09 05:29OpenAI claims huge maths breakthrough on a famed ‘Millennium Problem’
2026-09-09 06:11OpenAI claims to have solved the Navier-Stokes equation problem, but its originality is questioned by the academic community
2026-09-08 22:00Google DeepMind Maps 9 Billion Possible DNA Variants

All reports (1)SOURCES

↗

A arXiv cs.AI en 2026-09-07 12:00

This study uses machine learning algorithms to classify the safety levels of power systems, aiming to improve system stability and prevent large-scale failures. The study employs Newton-Raphson method to extract data and combines SMOTE with PCA techniques to address sample imbalance and dimensionality reduction. Three models—KNN, Random Forest (RF), and Support Vector Machine (SVM)—were tested in IEEE-14 and IEEE-30 systems. Evaluation results show that Random Forest achieved the highest F1 score of 0.97 in IEEE-30 systems; PCA contributed more to overall model performance than SMOTE, while SMOTE could improve recall rate but might introduce false positives. This study confirms that machine learning is an effective alternative to traditional contingency analysis, enabling real-time optimization of safety assessments.