In recent years, due to the ever-increasing electrical demand and the limitations associated with energy resources, power systems are operated near the limits of their transmission capabilities. Therefore, real-time monitoring and control of the power system are necessary for secure operation of the system. Due to the increase of dynamic loads, especially asynchronous motors, the monitoring of short-term voltage stability (SVS) is necessary to be taken into consideration. With the introduction of wide area measurement systems (WAMS) and phasor measurement units (PMUs), real-time voltage and current phasors are provided for online monitoring and control of power systems. Therefore, data-based or model-free methods can be developed for online stability monitoring. In this thesis, two data-based methods are proposed for real-time monitoring of SVS. The first method is based on SVS indices and the other uses data mining techniques. In SVS indices-based methods, eleven appropriate indices are defined for online short-term voltage stability. The indices are calculated using the data measured by the PMUs and compared with pre-defined thresholds to assess the short-term voltage stability of the system. The performance of these indices is studied in terms of the time and accuracy required for determining the short-term voltage stability/instability. On the other hand, in methods which use data mining techniques, a suitable database is created by the PMUs measured data from the various disturbances over many years. This database is then clustered and To evaluate the performance of the proposed online SVS assessment methods, the IEEE 39-bus test system is simulated and the performance of the SVS indices are assessed. Furthermore a database consist of various scenarios related to short-term voltage phenomena is created and then the DT is designed in an offline manner, using the SVS indices as the decision features. Finally, the developed DT is used for online SVS assessment of the system. Furthermore, the performance of the proposed methods are compared. The constituted DT can detect the stable or unstable states at almost one second, which indicates the superiority of the proposed online SVS assessment over the previous methods. Keywords: Short-term voltage stability, real-time monitoring, WAMS, PMU, voltage stability indices, data mining, decision tree.
