基于EKF与UKF的乙苯化工过程状态估计性能评估

Performance evaluation of ethylbenzene chemical process state estimation based on EKF and UKF

当动态系统内部状态无法直接测量时,状态估计是获取内部状态数据有效的解决方法。该文基于扩展卡尔曼滤波器和无迹卡尔曼滤波器对乙苯化工过程中的状态估计进行了分析比较研究,依靠少量测量数据估计出了乙苯化工过程系统中的25个状态值。经过MATLAB一系列仿真实验验证了两种估计器的有效性,分析了在高斯噪声下,两种滤波器的性能与时间复杂度。结果表明:无迹卡尔曼滤波器能够提供更好的状态估计性能,两者都能够处理较大的输入扰动,无迹卡尔曼滤波器在状态估计计算时长上更具优势。

When the internal state of dynamic system cannot be measured directly,state estimation is an effective method to obtain internal state data.In this paper,state estimation in ethylbenzene chemical process is analyzed and compared on the basis of Extended Kalman Filter and Unscented Kalman Filter.Twenty-five state values in ethylbenzene chemical process system are estimated based on a few measurement data.Through a series of MATLAB simulation experiments,the effectiveness of the two estimators is verified,and the performance and time complexity of the two filters under Gaussian noise are analyzed.The results show that Unscented Kalman Filter can provide better performance of state estimation,both can deal with larger input perturbations,and Unscented Kalman Filter has more advantages in the computation time of state estimation.