Enhanced Power System State Estimation Using Machine Learning Algorithms

Abstract

The widespread implementation of renewable energy sources is posing new and distinct challenges for power systems. Consequently, power system state estimation has become increasingly essential for monitoring, operating, and safeguarding modern power systems. Conventionally, physics-based models such as weighted least square or weighted least absolute value were utilized, which classically analyze a single snapshot of the systems and fail to capture the temporal connections of system states. Thus, this study exploits the potential of machine learning approaches to forecast the state values of power systems. The performance and stability of innovative machine learning methodologies are validated using the IEEE systems. The results of the simulations are encouraging, which shows the effectiveness and feasibility of the proposed machine learning methods for power system state estimation.