Voltage stability and critical bus detection using a decision tree approach based on PMU measurements

Monitoring voltage stability is important to maintain the power system operation safe during normal and disturbance conditions. Continuation power flow (CPF) is usually used to perform voltage stability assessment by creating a nose curve showing the distance between the collapse point and the current operating condition. However, the CPF requires high computation time, especially for large and complex power systems. On the other hand, the development of synchronous measurement unit from phasor measurement unit (PMU) in the wide-area monitoring systems (WAMSs) can give the power system operator real-time monitoring by providing voltage magnitude and angle measurement. High-resolution data from WAMS are valuable to create the data-driven model to assist the power system operator. Using this approach, voltage stability and critical bus detection are proposed. The proposed model is based on a decision tree and implemented in the simulated power systems using modified IEEE 14 and 118 buses test system. The proposed method will be tested with different operation conditions, variables as a feature, different numbers of PMUs, and PMU locations. The effectiveness of the proposed method is compared to another machine learning methods, such as support vector machine (SVM), random forest (RF), multilayer perceptron (MLP), extra tree (ET), and radius neighbors (RN). The proposed method has better performance than SVM, RF, ET, and RN and faster computation time than MLP. All the best scenarios across different types of comparisons conducted in the simulations can predict stability detection and critical bus > 90 within 2-s computation time.