Dynamic Voltage Collapse Prediction And Control In Power Systems

Oleh :
Muhammad Nizam - - Fak. Teknik

Voltage instability and voltage collapse problems are of great concern to many electric power industries. Such "roblems are often associated with contingencies like continuously increased loading, unexpected line and generator outages. In this situation, it is important to assess voltage stability of power systems by developing tools that can predict the distance to the point of collapse as well as identify the weak areas in a power system. To avoid power system from voltage collapse occurrence, ) tools for voltage control need to be considered as well. Much effort is currently been put into research on the phenomena of voltage collapse and many approaches have been explored. However, there is still a need for reducing the computational time in dynamic voltage stability assessment and control. Therefore, artificial intelligent techniques such as support vector machine (SVM) has been developed for fast and accurate prediction of voltage collapse and for identifying the voltage unstable areas in a power system. Initial work focused on the development of a new dynamic voltage collapse indicator named as the Power Transfer Stability Index (PTSI). The index is calculated by using information of total load apparent power, Thevenin voltage and impedance at a bus and phase angle between Thevenin and load buses. The value of PTSI will fall between 0 and I in which when PTSI value reaches 1, it indicates that a voltage collapse has occurred. The use of the proposed PTSI is then extended for identifying the voltage unstable buses in a power system as well as for determining the amount of shed load for the proposed adaptive under voltage load shedding scheme. The performance of the proposed techniques developed for dynamic voltage stability prediction and control are evaluated 'by implementing on three test systems, namely the IEEE 9 bus and 39 bus test systems and the 87 bus practical power system. Test results showed that the PTSI is more ac;curate and gives a better dynamic voltage collapse indication compared to the use of other indices such the power margin, voltage collapse proximity indicator and eigenvalues. The proposed adaptive load shedding method is found to give faster voltage recovery as compared to the other load shedding methods using controllers based on the fixed shed fixed delay and the variable shed variable delay schemes. Thus, the proposed adaptive load ~ shedding method can greatly reduce the computational time in control strategy and can be used as an effective counter measure against voltage collapse. For dynamic voltage instability prediction and weak area determination ln a power system using SVM, it is noted that the SVM takes much less training time as compared to artificial 1S1 neural network (ANN). For instance, for predicting dynamic voltage instability, the training times taken by SVM are 2.58 secs for the 39 bus system and 10.38 secs for the 87 bus system whereas the training times taken by ANN are 504 secs for the 39 bus system and 5551 secs for the 87 bus system. In terms ofaccuracy, the SVM gives accurate dynamic voltage ins tability prediction and determination of the voltage unstable buses in the test systems as comparable to ANN.