改进粒子群算法辨识噪声模型的卡尔曼直接转矩控制

Direct Torque Control Using EKF Based on Noise Model Identification by Improved Particle Swarm Algorithms

针对感应电机扩展卡尔曼滤波器状态估计中难以获得最优噪声矩阵问题,提出了一种基于改进粒子群算法辨识卡尔曼滤波器噪声矩阵的方法.通过调整粒子觅食策略对粒子群算法进行改进,运用改进算法优化滤波器噪声矩阵,再将优化的卡尔曼滤波器应用于感应电机无传感器直接转矩控制系统中.仿真结果表明,与试探法、标准粒子群算法和自适应粒子群算法相比,改进粒子群算法能够改善卡尔曼滤波器滤波性能,从而提高感应电机无传感器直接转矩控制系统的控制精度.

Aiming at the problem of getting optimal noise matrix in state estimation of the extended Kalman filter in induction motor,an optimized noise matrix method,which is based on the improved particle swarm optimization algorithms,is proposed. Updated strategies are adopted in standard particle swarm optimization. Based on the improved particle swarm optimization,a global optimized noise covariance estimation method is applied to extended Kalman filter in direct torque control system. Simulation results show that the proposed method can effectively improve the performance,much better than trial and error,particle swarm optimization and adaptive particle swarm optimizatio.