Dynamic State Estimation Of Power System Containing Renewable Energy

Identifying bad data and ensuring data accuracy is one of the basic tasks of state estimation.With the random and fluctuating renewable energy,such as wind and solar energy,connected to the power system,the stochastic characteristics of power system are more obvious.The difficulty of power grid parameter identification increases,and the data identification method fails.A higher performance level is required and more uncertain factors should be considered in the state estimation.In power system,the fluctuation measurement data caused by the stochastic characteristic of renewable energy is not the bad data.However,the traditional robust algorithm of dynamic state estimation cannot identify the above data and the fluctuation data will be treated as bad data,due to which the true characteristics of measurement data are lost.The provided wrong information will affect the operation safety of power system.Therefore,with renewable energy accessing to the power grid,how to ensure the robustness and accuracy of power system dynamic state estimation is particularly necessary.Firstly,the dynamic state estimation model of the power system is established.To solve the problem that Holt's two-parameter exponential smoothing method cannot guarantee the accuracy of state forecasting when faced with renewable energy,a dynamic state equation based on continuation forecasting method is proposed.By introducing renewable energy and load forecasting data,the accuracy of state forecasting is improved.Several common Kalman filtering algorithms for power system dynamic state estimation are studied.Focusing on the extended Kalman filtering and cubature Kalman filtering,the state estimation accuracy of linear filtering and nonlinear filtering algorithms are compared.Secondly,according to the data characteristics of renewable energy power system,the measurement data are classified.The bad data and the fluctuation data are collectively classified as the sudden change data,and an improved robust cubature Kalman filtering based on twolevel data identification is proposed.The innovation covariance is calculated according to the innovation sequence.The sudden change data is obtained by comparing the innovation covariance to the measurement variance.After the first-level identification,the data correlation is calculated according to the measurement sequence,and the bad data are distinguished from fluctuation data in the second-level identification.The bad data is corrected by the measurement noise scale factor.The simulation results show that the impact of bad data is reduced,and the fluctuation data is not mishandled by using the improved robust cubature Kalman filtering.The state estimation performance maintains a high accuracy.Finally,based on the above research,the uncertainties of renewable energy and load forecasting are considered and the interval dynamic state estimation is proposed.The two-point estimation method is used to solve the model,and the corresponding process is given.The interval estimation results based on Monte Carlo simulation are used as reference to verify the effectiveness of the two-point estimation method.By fitting the distribution characteristics of wind power and load forecasting errors in an actual power grid,the interval estimation results of the improved robust cubature Kalman filtering are analyzed.Compared with the original algorithm,the robustness and accuracy of interval estimation are verified.Compared with the point estimation,the interval estimation can provide operators with more state information.