论文部分内容阅读
针对间歇过程中关键参数,在线检测精度低、离线分析时滞大的问题,提出一种融合时滞测量值的状态估计方法。鉴于在线和离线检测值的采样周期不同,分仅有在线检测值和两种检测值并存等两种情况进行分析。考虑间歇过程的非线性、非高斯分布等特点,引入粒子滤波算法并基于贝叶斯方法对其进行扩展,实现两种检测值的信息融合。将提出的算法应用在啤酒发酵过程中,并与不考虑时滞测量值的估计效果对比。实验结果表明,该方法能够较好地处理考虑时滞值的状态估计问题,且效果优于不考虑时滞测量值的情况。
Aiming at the key parameters of intermittent process, the problem of low online detection accuracy and large offline analysis delay, a state estimation method based on time lag measurement is proposed. In view of the difference between the sampling period of on-line and off-line detection values, there are only two cases where online detection value and two detection values coexist. Considering the non-linear and non-Gaussian distribution of batch process, the particle filter algorithm is introduced and expanded based on Bayesian method to realize the information fusion of the two measured values. The proposed algorithm is applied to the beer fermentation process and compared with the estimated effect without considering the time lag measurement. The experimental results show that the proposed method can deal with the state estimation problem with delay value well, and its effect is better than that without delay measurement.