Cascading Failures Time Deduction And Preventive Control Based On Nodal Load Changes

Modern interconnected power system is large-scale, which has a wide range of components, closely inter-systeem, complex operating conditions and is suffered by many various and complicated interference.When system failure happen, if not handled properly, the faults is easy to expend and develop into cascading failure, then the system is blackout. Therefore, analysis the state changes after accident, deduce the evolution from initial failure to cascading failures, thus propose safety warning abd make prevention and control measure, which has important theoretical significance and practical value to prevent the occurrence of blackout.This paper propose a method and deduction model of cascading failures based on the time span, when summarized the existing analysis model. Time-oriented process method is uesd to deduct the development of cascading failures by this model, which is based on nodal load data and extract contingency set by reactance betweenness of network. In the process of cascading failure deduction, calculate the operating state by power flow calculation, take the system transient stability into account under the circumstance of the initial failure, line trip outage and load shedding, simulate the control measures of stability and monitor the low-voltage of power system. Interplay of many factors are considered in the process of cascading failures, the simulation results are more in line with development process of blackout in recent years.For some stage of cascading faliure lead to blakout, three security trends analysis methods are presented to confirm system operations after initial fault, and screen the moment of deteriorated operations, then give waring and control measures. There are three method: slope act which can reflect the change of total laod; hierarchical clustering method which divide the nodal load data and extract the characteristic section that represent the change trend of load; trend analysis based on electric network. Slope act can reflect the trend of total load directly, but ignores the inconsistencies of node load changes. Hierarchical clustering method can effectively reflect the node load changes. Both methods have fast calculation, but only consider the disposal of load data, the analysis of physical features of electric grid is insufficient. Trend analysis based on grid features can reflect the changes in node load,but the calculation a little slower.Finally, a optimal load shedding of OPF is adopted to make prevention and control measures. The result of example, show that the control measures can prevent the subsequent failure, prevent failure scale from swelling further.