Predicting voltage stability margin using a machine learning approach

نوع: Type: Thesis

مقطع: Segment: masters

عنوان: Title: Predicting voltage stability margin using a machine learning approach

ارائه دهنده: Provider: Nima Safarkhani

اساتید راهنما: Supervisors: Dr. Mohammad Amin Ghasemi - Dr. Saleh Razini

اساتید مشاور: Advisory Professors:

اساتید ممتحن یا داور: Examining professors or referees: Dr. Alireza Hatami - Dr. Mohsen Nozadian

زمان و تاریخ ارائه: Time and date of presentation: 2025

مکان ارائه: Place of presentation: 33

چکیده: Abstract: In this study, multiple machine learning methods are used to estimate the loading margin based on regression techniques and voltage stability indices. Various scenarios for load changes and generator voltage magnitude variations are considered. Different operating conditions are obtained from power flow analysis, and stability indices along with loading margins are calculated for each scenario. This research introduces a machine learning approach for long-term voltage stability margin prediction. The key feature of the proposed technique is the utilization of various voltage stability indices as inputs to a set of machine learning models. Additionally, a method for generating training data under different operational conditions and N-1 contingency scenarios is proposed for training these models. The inputs and outputs of the model correspond to voltage phasor values and voltage stability margin indices, respectively. The best machine learning algorithm and the most suitable input categories for voltage stability indices (VSIs) are selected through a comparative study. The experiments are conducted on the IEEE 39-bus system, and results indicate that the artificial neural network method provides the best performance.