Research On Reactive Power Optimization And Fault Diagnosis Of Power System With Large-scale Wind Farms

Due to the clean and renewable characteristics of wind energy and the corresponding call for energy conservation and emission reduction,wind power generation has become a new trend of global power development.However,wind energy is also random and intermittent,so large-scale wind power grid connection is bound to bring adverse effects on the stable operation of power system,such as voltage stability,frequency fluctuation and so on.In addition,when a serious fault occurs in the power system,temporary overvoltage will lead to the disconnection of relevant wind farms,affect the local voltage level and frequency of the power grid,and is not conducive to the stable operation of the power system.In order to ensure the voltage stability of power system with large-scale wind farm,starting with the research of wind speed sequence prediction and wind farm output correlation,the corresponding reactive power compensation measures are determined based on the wind farm output correlation,to ensure the safe and stable operation of power system with large-scale wind farm.Then,in the case of wind turbine trip-offs in the extreme faults,the electrical quantity characteristics of relevant off grid wind turbines are analyzed,and the artificial intelligence method is used to identify the fault types.There are main research contents of this paper as follows:(1)In this paper,particle swarm optimization algorithm is used to optimize the number of sub modes and quadratic penalty factor of variational modal decomposition.After determining the optimal number of sub modes and the quadratic penalty factor,the original wind speed sequence is decomposed based on the improved variational modal decomposition method,to form multiple sub sequences with different characteristics.The ant colony algorithm is used to optimize the BP neural network.Based on the improved BP neural network,the features of each wind speed subsequence are extracted to form the corresponding prediction results.The prediction results of each wind speed subsequence are superimposed to obtain the final prediction results of wind speed series.Taking wind speed prediction in Northwest China as an example,the results show that this method can effectively predict wind speed and has high accuracy.(2)The type of wind turbine is determined,wind farm models with DFIG wind turbines are built in PSCAD,and then the wind speed sequence prediction results are taken as the input of wind turbine,to achieve the effect of simulating the actual operation of northwest electric field.The common reactive power compensation devices are introduced,and the advantages and disadvantages of various reactive power compensation devices are analyzed.SVC is selected as reactive power compensation equipment in this paper.In addition,the installation node of SVC is determined,and the parameters of corresponding SVC is determined by GA-ACO algorithm,which reduces the active and reactive power loss of power grid.Taking sum variance as the basis for dividing the number of typical scenes,the variation of weak node voltage is analyzed after adding shunt capacitor.Additionally,compared with the case of voltage level regulation with shunt capacitor,it is verified that the effect of voltage level regulation with SVC is better.(3)Faults such as short circuit,disconnection and abnormal wind speed are set in the wind farm model built above,to obtain the waveform of wind turbine disconnected due to protection action.Based on the transient voltage stability evaluation criteria and the low/high voltage ride through requirements of wind turbines,when the wind farm is disconnected,the data of voltage and output active power of corresponding wind turbines are summarized.Similarly,the ant colony algorithm is used to optimize the parameters of BP neural network,and the data such as voltage and output active power are used as the inputs of BP neural network for model training.The corresponding example analysis verifies the effectiveness of the fault identification method proposed in this paper.