Research On Online Load Modeling Method Based On Adaptive Kalman Filter

With the advancement of China's main grid interconnection and hybrid connection receiving end grid process,new energy and power electronic components is increasing,the load demand and characteristics of different regions also show great differences,randomness and dispersion.The load is relatively behind the modeling work of generators and transmission networks,which affects the accuracy of the entire power system model calculation.Therefore,it is certainly urgent to perform online parameters based on realtime dynamic data Recognition faced with time-varying structural parameters and complex power system loads to model.Aiming at the new features of the power system,this article applies the real-time small disturbance data provided by the Wide Area Measurement System.Taking the comprehensive load model as the research object,the load parameters are identified online based on the adaptive Kalman filter algorithm,and the genetic algorithm is used to obtain more accurate covariance matrix for system noise estimation,which provides effective parameter identification results for power system operation analysis,and improves the ability to monitor grid load.In terms of online load parameter identification,in view of the limitations of traditional load modeling methods in data acquisition,identification accuracy and online identification,an adaptive Kalman filter online load identification algorithm based on Wide Area Measurement System measurement technology is proposed.First consider the composition of the load model,establish and linearize the comprehensive load model.Based on the small disturbance data measured in real time by the Phasor Measurement Unit,a data preprocessing method is proposed.Compared with the traditional Kalman filter algorithm and the improved Sage-Husa adaptive Kalman filter algorithm,the identification problem is solved by using the prediction error method.In terms of system noise optimization,an adaptive Kalman filter algorithm based on genetic algorithm optimization is proposed.First,adaptive adjustments are made to genetic algorithm crossover,genetics and selection to avoid being prone to premature convergence and converging to the local optimum.Then use the improved genetic algorithm to determine the initial estimation error covariance matrix offline to improve the calculation efficiency and accuracy of the adaptive Kalman filter.Finally,the traditional KF,AKF and AKF optimized by this algorithm identify parameters and compare the fit of active power and reactive power to verify the effectiveness of genetic algorithm optimization.The load parameter online identification method based on the improved adaptive Kalman filter and the improved genetic algorithm for optimizing model parameters proposed in this paper are verified by real-time operating data of substation of Zhejiang Power Grid.The results of calculation examples show that the online identification algorithm proposed in this paper has Better applicability and higher accuracy.