Power System Fault Screening And Ranking Based On Dynamic Security Region

With the development of the power system, the domestic grid is showing a trend whose features are large-scale?pressure and connectivity and so on. Grid expanding brings huge economic benefits and the potential problems of the safe operation of power systems. The theory of power system dynamic security domain(DSR) provides a new problem-solving ideas considering the transient safety assessment. So far, the two mature methods which is used to solve dynamic security domain hyperplane coefficients is fitting and analytical method. Fitting has the characteristics of exact calculation result and slow calculation due to the large amount of the critical power injection point. Analytical method is based on a critical point of injection, a method of quickly obtaining security domain boundaries. The method has the characteristics of quick calculation speed and poor accuracy.There are two methods of power system transient stability analysis: probabilistic analysis and uncertainty analysis. Although the uncertainty analysis has clear physical concept, the results are only limited stability and instability, difficultly giving accurate information to establish an effective decision system. Probabilistic analysis takes into account the random factors such as the failure location, giving the probabilistic insecurity index.The theoretical basis and its solving method of probabilistic insecurity indexes?the BP neural network algorithm and fault filtering and sorting method are analyzed in this paper. The fault filtering model based on the degree of the instability of the system and the fault sorting model based on the system insecurity probability index are presented. The key step of computing system instability degree and probabilistic insecurity index is calculating a given failure hyper plane coefficient of dynamic security domain. Based on a large amount of data verification, the paper proposed that the BP neural network algorithm could be used to calculate a given fault hyper plane coefficient of the dynamic security domain. and the calculation accuracy satisfies the requirement of engineering practice. In order to ensure the speed and accuracy, when calculating the power system fault filtering, we use the analytical method to calculate the DSR hyper plane coefficient to get rapid filtering results; When calculating the fault sorting problem, we use the BP neural network to calculate DSR hyper plane coefficient to get the accurate calculation results. If any special cases appeared in the final results, fitting process calculation will be used in comparison.This paper proposes a new transient evaluation model: power system fault filtering and sorting assessment model. The model complete the screening based on the system instability, then complete failure ranking based on probabilistic insecurity index. The prerequisite of calculating instability degree and probability insecurity index is solving DSR hyper plane coefficients. This model considers random factors such as fault type, truly reflecting the level of the given fault system risk, providing effective reference information for the operating personnel.