Self-organized Criticalstate Identification Method Of Power System With Centralized Large-scale Wind Power

Wind power has been the rapid development with doubling every year for five years,since the "Renewable Energy Law of the People's Republic of China'' was issued in 2005.The problems about wind power appeared one after another after 2010.The safe operation of the wind turbine easily destroyed due to the harsh operating environment and technical defect of wind power generation equipment away from the monitoring center,and thus it threaten the safe operation of the power system.A large number of wind turbines trip because of the frequency protection in the Western European grid "11 · 4“blackout,that destroyed the balance of power,resulting in frequency deviation for a period of time in 2006;In the wind power turbines gird-off failures of Zhangjiakou Hebei and Gansu Jiuquan area,a large number of wind turbines was grid-off due to voltage problems,leading to the loss of a large number of system load and significantly lower the grid frequency.Fortunately,these failures did not continue,did not lead to blackouts,but it cannot deny that the risks of large-scale wind power connected grid must pay close attention.Based on the self-organized criticality theory of the complex power system,the reason why cascading failure of power system occurs is that the power system is in high risk,which is called the self-organized critical state.If the power system state can be identified and monitored,and the control measures can be taken in time when the power system is in the self-organized critical state,the cascading failure can be prevented effectively.Against this background,the paper launches the research about the self-organized critical state identification methods of power system with large-scale wind power.The innovations are as follows:Based on the mechanism of the complex grid cascading failures revealed by the theory of self-organized criticality,the cascading failure(blackout)is the external manifestation of the grid self-organized criticality and the self-organized criticality is the inherent laws of the power grid cascading failures(large blackouts).On this basis,the evolution mechanism and mathematical characterization of power grid self-organized criticality is described systematically,including power law tail characteristics of time and the fractal dimension of space.By theoretical analysis of self-organized criticality,whether blackout data are consistent with a power-law tail feature can be used as a criterion for grid self-organized critical state identification.In terms of blackout data acquisition,improved OPA model is put forward.The volatility of wind power output is considered into the model,while removing the slow dynamic processes.The model use DC power flow algorithm to calculate flow distribution.Blackout data was got from multiple cascading failure simulation.The least squares method was used to fit power law tail curve generally.In this paper,maximum likelihood estimation method and K-S is proposed to fit power law tail curve due to existing defects of the least squares method,making the fitting results more accurate and reliable.It is a strong methods and tool for self-organized criticality analysis and identification.Based on analysis of the impact of large-scale wind power connected grid on self-organized critical state deeply,wind power fluctuation entropy is proposed to characterize the degree of ordering of wind power output volatility;The inhomogeneity of the power flow distribution is divided into two kinds caused by some transmission line with low load rate and by some transmission line with high load rate respectively,then it is put forward that the latter is the real factors of the grid to enter one of the self-organized critical state.The weighted power flow entropy is proposed to identify inhomogeneity of power flow distribution caused by some transmission line with high load rate.At last,using power law tail curve fitting method to analysis the impact various physical parameters proposed on the grid self-organized critical state respectively.The results show that the power flow entropy weighted,network topology entropy and wind power fluctuation entropy can be used as the key indicators of self-organized critical state identification.A method of self-organized critical state identification is put forward based on LVQ neural network,which services complex gird with centralized large-scale wind power.The entropy indicators extracted in the paper is considered as the main input object of the method.It is considered as training samples that power law tail curve fitting method generate.Using LVQ algorithm train LVQ neural network created,and then the network model trained is used to identify self-organization critical state of gird with centralized large-scale wind power.This method establishes a direct relationship between the physical indicators and self-organized critical state,to avoid the problem of subjective intervention and a number of simulation times using traditional identification methods.