Transient Voltage Stability Evaluation Of Wind Power Grid Connected System Based On CPSO-BP Neural Network

With the continuous expansion of the installed scale and capacity of wind farms,the impact of wind power access on the security and stability of power system can’t be ignored.Among many problems related to the safe and stable operation of the whole power system after large-scale wind power grid connection,system transient voltage stability is a problem that needs special attention.In order to evaluate the transient voltage stability of the system after the wind farm is connected to the power system as soon as possible and accurately,a transient voltage stability evaluation method based on CPSO-BP neural network is proposed in this thesis.Firstly,starting with the development status of wind power at home and abroad,this thesis analyzes the operation characteristics of wind power,and establishes a multi machine dynamic equivalent model of wind farm based on CPSO-FCM algorithm.The model combines Chaotic Particle Swarm Optimization(CPSO)with Fuzzy C-Means(FCM)algorithm,and takes the operation data such as wind speed,generator angular speed,rotor current,real-time active and reactive power as clustering indexes to carry out multi machine dynamic equivalence of wind farm.In the case of three-phase short circuit,through the evaluation of voltage error,active power error and reactive power error,the effectiveness of the proposed method is verified,which saves time for input feature acquisition in subsequent chapters.Then,according to the feature selection principle and combined with the existing research,this thesis selects 32 original feature quantities,including traditional feature quantities and wind power feature quantities.Then,the Kernel Principal Component Analysis(KPCA)is used to analyze and calculate the 32 feature quantities,and the dimension of the feature quantity is reduced from 32 to 13,which shortens the calculation time for the subsequent chapters to evaluate the transient voltage stability of wind power grid connected system.Finally,because the current traditional methods are difficult to quickly and accurately judge whether the system transient voltage is stable after the wind farm is connected to the system,this thesis proposes a transient voltage stability evaluation method of wind power grid connected system based on CPSO-BP combination.Chaos theory is used to improve the shortcomings of particle swarm optimization algorithm.The improved algorithm is applied to the optimization of the initial weight and threshold of back propagation(BP)neural network.Then,according to the input feature acquisition principle,the traditional physical quantities before and after system failure and the physical quantities related to wind farm are collected as the input feature quantities of CPSO-BP neural network for supervised learning.The obtained model is first classified according to the margin value of Critical Clearing Time(CTT),and the classified samples are used for transient voltage stability evaluation and CCT margin prediction of wind power grid connected system.The simulation analysis of wind power grid connection is carried out by using the standard example of England 10 machine 39 bus system,and the results show the effectiveness of the proposed method.