【作者】 李彬艳； 曹霄鹏； 陈宁； 高家元； 罗彪； 陈嘉瑶；

【Author】 Binyan Li;Xiaopeng Cao;Ning Chen;Jiayuan Gao;Biao Luo;Jiayao Chen;School of Electrical and Information Engineering,Changsha University of Science and Technology;School of Automation,Central South University;

【机构】 长沙理工大学电气与信息工程学院； 中南大学自动化学院；

【摘要】 大多数材料工业制备过程往往需要在高温环境下进行长时间的反应,烧结过程温度场的优化控制对于提升材料产品品质至关重要。然而,工业烧结环境的复杂导致外部干扰广泛存在,有效的烧结过程温度场优化控制方法较为缺乏。为此,针对温度场外部干扰多、控制过程数据传输与计算量大、复杂度高以及控制输入的频繁更新使得温度分布难以稳定的难点,研究一种基于自适应动态规划(Adaptive dynamic programming,ADP)的温度场事件触发H_∞控制方法。首先,对事件触发下的温度场最优控制问题进行描述,将所研究的温度场系统的H_∞控制问题,转化成求解低阶慢子系统的哈密顿-雅可比-伊萨克(Hamilton-Jacobi-Issac,HJI)方程的最优解的问题。为确定事件触发H_∞控制器的更新时刻,提出融合系统状态信息、控制信息等对温度场系统的事件触发条件进行设计的方法。以此为基础,采用ADP的评价-执行网络结构,研究温度场系统的事件触发H_∞优化控制方法。其中,评价网络连续更新以逼近最优值函数,执行网络根据事件触发条件进行非周期更新以获得控制策略。最后,将所提方法应用于辊道窑烧结过程,仿真结果表明了所提事件触发优化控制方法应用于实际工业过程的收敛性和有效性。更多还原

【Abstract】 The majority of material industrial preparation processes often require prolonged reactions under hightemperature conditions,so the optimal control of the sintering temperature field is crucial for enhancing the quality of material products.However,the complexity of the sintering environment leads to widespread external interferences,and the effective control methods for the temperature field are relatively scarce.To address these challenges,such as the multitude of external disturbances in the temperature field,the large volume of data transmission and computation in the control process,the high complexity,and the frequent updates of control inputs that make it difficult to stabilize the temperature distribution,the research has been conducted on an event-triggered H_∞ control method for the temperature field based on ADP(Adaptive dynamic programming).Firstly,the optimal control problem of the temperature field under the event trigger is described and transformed into the problem of finding the optimal solution of the Hamilton-JacobiIssac(HJI) equation for a dominant lower-order slow subsystem.To determine the update moments of the event-triggered H_∞ controller,a designing method for the event-triggered condition of the temperature field,which integrates the state and control information is proposed.On this basis,the ADP structure is employed to study the event-triggered H_∞ optimal control method for the temperature field system.In this context,the critic network continuously updates to approximate the optimal value function,while the actor network is updated nonperiodically to obtain the control strategy according to the event-triggering conditions.Finally,the proposed method is applied to a roller kiln,and the simulation results demonstrate the convergence and effectiveness of the proposed event-triggered control method when applied to actual industrial processes.更多还原