Data-Optimized Contingency Screening

Ensuring the security of the power system is essential for stability and reliability, especially in the event of disruption. According to editorial research analyzed by Groundwork, Effective classification of contingency in power systems enables proactive decision-making and mitigates large-scale breakdowns and failures. In this article, we explore the use of machine learning algorithms to classify security levels of contingencies in power systems into safe, moderate, or severe classes.

Traditional contingency analysis relies on manual calculations and simulations, which can be time-consuming and prone to errors. Machine learning offers a scalable and powerful alternative, enabling real-time assessment of power system security. This study focuses on the use of machine learning algorithms to classify contingencies in power systems.

Methodology

The Newton-Raphson load flow method extracts system data from contingency scenarios, using the Overall Performance Index (OPI) as a safety measure. To address class imbalance and reduce dimensionality, Synthetic Minority Over-Sampling Technique (SMOTE) and Principal Component Analysis (PCA) are used for data pre-processing.

The study evaluates the performance of three machine learning algorithms: K-Nearest Neighbours (KNN), Random Forest (RF), and Support Vector Machines (SVM). These algorithms are trained and evaluated on datasets generated through N-k contingency scenarios for k equal 1, 2, and 3 on IEEE-14 and IEEE-30 bus systems.

Four hybrid pre-processing configurations are used:

1. Normalized: data is normalized to a common scale.
2. SMOTE-balanced: SMOTE is used to balance the classes.
3. PCA-transformed: PCA is used to reduce dimensionality.
4. Combined SMOTE PCA-transformed: both SMOTE and PCA are used.

Key Findings and Data Analysis

The study assesses the performance of the machine learning algorithms using precision, recall, and F1 score, with priority given to the severe contingency classes. The results show that:

• Random Forest (RF) achieved the highest F1 scores of 0.97 in IEEE-30 and 0.86 in IEEE-14.
• Support Vector Machines (SVM) benefits significantly from PCA and improves the accuracy of the classification.
• K-Nearest Neighbours (KNN) is best suited for SMOTE and PCA conversion.
• PCA contributes more than SMOTE to the overall performance of the model.
• SMOTE improves recall but can introduce false positives and is therefore a compromise of accuracy.

This study highlights machine learning as a scalable and powerful alternative to traditional contingency analysis, which improves the assessment of security in real time. The results demonstrate the effectiveness of machine learning algorithms in classifying security levels of contingencies in power systems.