基于ISR-UKF的PET固相缩聚过程参数和状态估计

连续式PET固相缩聚移动床反应器具有显著的分布参数特性,由于建模简化或过程时变等原因使得所建模型参数不精确,导致反应器状态的估计失真。首先采用正交配置方法离散PET固相缩聚过程的偏微分方程模型,然后基于改进的平方根不敏Kalman滤波算法(ISR-UKF),设计自适应联合估计器,同时获得参数和状态估计值。实验结果表明,参数估计结果合理,状态估计精度较高且稳定性好;并获得了频率因子、活化能近似值和有效系数随反应温度动态变化的规律,表明所提出的联合估计器能获得较好的实际应用效果。

State and parameter estimation of solid-state polymerization process for PET based on ISR-UKF

The moving bed reactor of the continuous solid-state polymerization (SSP)process for PET is a kind of distributed parameter system.The model parameters are easy to become inaccurate and uncertain due to the modeling simplification and time-varying property, which leads to the state estimation error.An adaptive joint estimator for the SSP process is designed based on the ISR-UKF for the real-time parameter and state estimation, following the orthogonal collocation method to discretize the process model described by partial differential equations.Experiment and simulation results show the estimator can produce reasonable parameter estimate values and have excellent state track capability and good stability.The approximate frequency factor and active energy and the efficiency coefficient dynamics with the reaction temperature are obtained by the joint estimator, which shows the joint estimator can be well used for practical applications.