刘守元,魏代兴,刘聪汉,宋晓轩,辛丽平,江才嘉,吴永玲,范锐,孙荣荣.间歇蒸煮过程的降阶模型与蒸煮终点卡伯值预测控制研究[J].中国造纸学报,2025,(2):164-172 本文二维码信息
二维码(扫一下试试看!)
间歇蒸煮过程的降阶模型与蒸煮终点卡伯值预测控制研究
Reduced-order Modeling and Predictive Control of Kappa Number at Cooking Endpoint in Batch Digestion Processes
投稿时间:2025-01-14  修订日期:2025-03-24
DOI:10.11981/j.issn.1000-6842.2025.02.164
中文关键词:  间歇蒸煮过程  卡伯值  扩展普渡模型  模型降阶  模型预测控制
Key Words:batch cooking process  Kappa number  extended Purdue model  model order reduction  model-predictive control
基金项目:山东省自然科学基金(ZR2021MF076、ZR2016FB04);山东省重点研发项目(2018GHY115025);中国博士后面上项目(2018M642611);国家自然科学基金(201606141、62303258)。
作者单位邮编
刘守元* 青岛理工大学信息与控制工程学院山东青岛266500 266500
魏代兴 青岛理工大学信息与控制工程学院山东青岛266500 266500
刘聪汉 青岛理工大学信息与控制工程学院山东青岛266500 266500
宋晓轩 青岛理工大学信息与控制工程学院山东青岛266500 266500
辛丽平* 青岛理工大学信息与控制工程学院山东青岛266500 266500
江才嘉 青岛理工大学信息与控制工程学院山东青岛266500 266500
吴永玲 青岛理工大学信息与控制工程学院山东青岛266500 266500
范锐 青岛理工大学信息与控制工程学院山东青岛266500 266500
孙荣荣 青岛理工大学信息与控制工程学院山东青岛266500 266500
摘要点击次数: 0
全文下载次数: 0
中文摘要:
      本研究提出一种基于子空间辨识的模型预测控制(MPC)算法,通过扩展普渡模型对间歇蒸煮制浆过程中固相、自由液相和截留液相的组分浓度变化进行模拟仿真,采用子空间系统辨识的数值算法(N4SID)对非线性间歇蒸煮过程的动力学模型进行降阶,建立低维降阶状态空间模型,并引入龙伯格观测器对其状态变量进行在线估计,结合MPC策略,实现间歇蒸煮过程的精确控制。Matlab仿真结果表明,在开发的MPC算法作用下,系统能够抑制蒸煮过程中的卡伯值波动,确保蒸煮终点的纸浆卡伯值达到预设值且误差≤2%;所建立的间歇蒸煮过程4阶状态空间模型与扩展普渡模型拟合度可达99.80%,能较好与真实系统匹配,有效减少预测控制算法的计算复杂度。
Abstract:
      This study proposed a model predictive control (MPC) algorithm based on subspace identification. By extending the Purdue model, it simulated and modeled the changes in component concentrations of the solid phase, free liquid phase, and retained liquid phase during the batch cooking process. The numerical algorithm (N4SID) for subspace system identification was used to reduce the order of the nonlinear batch cooking process’s kinetic model, establishing a low-dimensional reduced-order state space model. A Lombard observer was introduced to perform online estimation of the state variables. Combined with the MPC strategy, precise control of the intermittent cooking process was achieved. MATLAB simulation results showed that under the developed MPC algorithm, the system could suppress fluctuations in the cooking process’s Kappa number, ensuring that the pulp Kappa number at the end of cooking reached the preset value with an error of ≤2%. The established fourth-order state-space model of the intermittent cooking process and the extended Purdue model achieved a fitting degree of 99.80%, demonstrating good agreement with the actual system and effectively reducing the computational complexity of the predictive control algorithm.
查看全文  查看/发表评论  下载PDF阅读器  HTML

分享按钮