基于平方根形式的无迹卡尔曼粒子滤波的电力系统动态估计

Dynamic Estimation of Power System Based on Power Unscented Kalman Particle Filter in Square Root Situation

针对无迹卡尔曼滤波(UKF)和粒子滤波(PF)状态估计精度低的缺点,把平方根形式的无迹卡尔曼粒子滤波(SRUPF)引入到电力系统状态估计中。在该方法中,无迹卡尔曼滤波作为概率密度函数进行更新,利用Markov链蒙特卡罗方法解决重采样后粒子的匮乏问题,利用平方根形式解决状态估计的收敛速度和稳定性问题。在保障精度的情况下,为了不牺牲大量的计算时间,适当的减少了粒子个数。最后在IEEE14进行了仿真验证,仿真结果表明SR UPF的引入可以有效提高电力系统的动态估计精度。

In terms of the shortcomings of unscented Kalman filter(UKF) and particle filter(PF) state estimation, this paper introduces unscented Kalman particle filter in square root form(SR-UPF) into power system state estimation.In this method, the unscented Kalman filter is updated as a probability density function.Markov chain Monte Carlo method is used to solve the problem of particle shortage after resampling, and the square root form solves the problem of convergence speed and stability of state estimation.In the case of ensuring accuracy, in order not to sacrifice a large amount of calculation time, the number of particles is appropriately reduced.Finally, the simulations are verified in IEEE 14.The simulation results show that the introduction of SR-UPF can effectively improve the dynamic estimation accuracy of the power system without sacrificing a large amount of computation time.