Document Type

Thesis - Open Access

Award Date

2026

Degree Name

Master of Science (MS)

Department / School

Electrical Engineering and Computer Science

First Advisor

Xijin Ge

Abstract

Cascading failures are a major concern for modern power systems‚ where a small failure can cascade through the network to create a large blackout. Because such an event could have catastrophic economic and social costs it is important to understand cascading failures‚ how to model them‚ and the possibility of reducing them. In this thesis‚ we develop a data-driven framework based on actual utility outage data to model and reduce cascading failures. The proposed methodology is based on generation-dependent interaction models and in this study interaction matrices obtained from historical outage events are used to describe the interaction between generations. One of the most critical model parameters is the memory parameter that describes how future generations can be influenced by earlier generations. In previous studies‚ memory value is assumed to be fixed which is not practical. To address this‚ the optimal value of memory was determined from the NYISO (New York Independent System Operator) data using a genetic algorithm (GA) that minimizes the distance between the CCDs (Complementary Cumulative Distribution) of the observed and simulated cascade sizes. For the NYISO dataset‚ we found that the optimal memory value is 58.27% which fits better than the full-memory assumption. With the interaction model improved‚ the next step is to prevent cascading failures‚ which is the focus of a second optimization framework. This determines the critical components‚ and how much mitigation is required for each component. This is an application of a nested GA. So‚ effectively‚ the outer loop sets the minimum number of critical components‚ and the inner loop sets the minimum reinforcement level necessary to keep the cascade risk under the required threshold. Hence this method performs without enormously affecting the performance of the mitigation. This would eliminate about one third of the upgrades needed for the baseline method while offering almost similar mitigation performance. The data set from BPA (Bonneville Power Administration) was also tested with the proposed framework‚ and the optimal memory was 99.27%. This result is like the full-memory assumption used in previous studies and the optimal number of critical components remains consistent with existing results. This proves that this framework works well on other datasets and system conditions. In summary‚ in this work we provide practical and data-driven approaches to evaluate models of cascading failures.

Publisher

South Dakota State University

Download

Share

COinS
 

Rights Statement

In Copyright