PMU Data Driven Machine Learning Regression Method for Real Time Voltage Stability Assessment

With the expansion of power grids, the probability of encountering a variety of faults and disturbances has also increased. Voltage stability is now a big issue for how power systems work. This paper suggests using a machine learning method to analyse voltage stability real-time. This approach uses different line and node voltage stability indices to check the system's voltage condition. The approach relies on Feed Forward Neural Networks (FFNN) and Random Forest (RF) regression models. Inputs for the model are derived from data collected by optimally positioned Phasor Measurement Units (PMUs). Optimal placements for the PMUs are found using Integer Linear Programming (ILP). This approach has been validated utilizing the IEEE-30 bus system. The results shows that this approach works well for real time monitoring and online applications.