【作者】 仲兆准；

【导师】 王景成；

【作者基本信息】 上海交通大学 ， 控制理论与控制工程， 2011， 博士

【摘要】 带钢热连轧精轧是钢铁工业的重要生产过程,其产品在表面质量、几何尺寸及机械性能等方面都有严格的要求。由于其生产过程工艺复杂,工况恶劣,现代热连轧生产线往往装备昂贵的高精设备,且由多种自动控制系统联合工作,包括自动厚度控制系统,自动宽度控制系统,自动板形控制系统等。另一方面,由于各类扰动的存在,前一轧机带钢出口速度和后一轧机带钢入口速度往往并不相等,这样的速差必然导致前后轧机间带钢长度与张力的波动。为保证在恒定张力情况下进行轧制,目前绝大多数热连轧生产线在前后两轧机间都安装了活套装置。由于传统的控制方法将厚度、活套及张力,作为独立的子系统进行控制,不能协调处理系统间的耦合,控制性能很难满足市场对热轧带钢产品日益苛刻的要求。本文以厚度、活套及张力综合系统为研究对象,在滑模控制,H∞控制,几乎干扰解耦控制及预测控制方面做了相关探索性研究。主要内容包括以下几个方面:(1)提出了基于线性模型及滑模方法的活套张力控制方案,采用耦合项对消的方法,实现了系统在滑模面上的解耦,增加了系统鲁棒性,提高了控制性能。另外,针对传统滑模控制无法处理非匹配扰动的缺点,提出了积分滑模和H∞控制相结合的方法,积分滑模方法处理匹配扰动,H∞控制处理非匹配扰动,进一步提高了活套张力控制系统性能,获得了比单独使用一种方法更好的效果。(2)针对活套张力系统固有的非线性特性,提出了基于非线性模型及滑模方法的解耦控制方案和基于反馈线性化的几乎干扰解耦控制,在活套张力系统非线性模型的基础上,对系统的非线性控制方法做了探索性研究。滑模控制能实现非线性耦合项的对消,并能较好地抑制匹配扰动,对执行器的未建模动态具有较好的鲁棒性。基于反馈线性化的几乎干扰解耦控制,则能实现参数化控制器的设计,系统对扰动的抑制能力可调,且参数调节直观方便。非线性控制器具有全局稳定,可运行范围大的优点,能有效解决线性控制方案由于运行区域有限、需要随工作点变化切换控制器的问题。(3)针对现代热连轧生产线大都通过数字可编程逻辑控制器进行系统控制,离散控制器更适合实际应用的特点,并针对厚度系统与活套张力系统间的强耦合,建立了厚度-活套综合系统状态空间模型,并在其基础上,设计约束离散预测控制器,通过加权矩阵的调节,协调厚度、活套与张力的响应动态,提高了综合控制系统性能。(4)以实际生产过程数据为基础进行仿真分析,采用实际扰动作为仿真输入,验证了基于厚度-活套综合系统状态空间模型的仿真程序的可信性。并进一步,对本文提出的约束预测控制算法进行仿真,证明了新算法的有效性,可充分提高综合控制系统的性能。更多还原

【Abstract】 Hot strip finishing mills (HSFMs) are important processes of the iron and steel industry. The specifications which should be satisfied in HSFMs are surface and dimensional quality, mechanical properties and so on. The process is rather complicated and challenging, and modern HSFMs are equipped with expensive equipments with high accuracy. The dimension quality is guaranteed by dedicated control systems such as AGC (Automatic Gauge Control), AWC (Automatic Width Control), ASC (Automatic Shape Control) and so on. On the other hand, the strip speed leaving the upstream mill usually doesn’t equal the the strip speed entering the downstream mill due to various disturbances, which will cause the deviations of the strip length and tension between the mills. Looper is equipped between each pair of mills to guarantee stable operation of the process with constant strip tension. Gauge, looper and tension are controlled independly in the conventional control strategies, which are not able to deal with the interactions between subsystems and are not suitable for the demanding market. In this paper, gauge, looper and tension control systems are studied from the SMC (Sliding Mode Control), nonlinear control and MPC (Model Predictive Control) aspect, including:(1) Looper-tension control strategy based on linearized model and SMC method is proposed to improve the robustness and performances of the system, and the looper-tension system is decoupled on the sliding manifold by interaction item canceling. And H∞control strategy based on ISMC (Integral SMC) is proposed to deal with both matching and mismatching disturbances, which achieves better performances than adopting H∞control strategy alone. (2) Nonlinear control strategies are proposed to deal with the nonlinear nature of the looper-tension system. SMC strategy is considered for its well-known robustness withstand matching disturbances and uncertainties, and with the help of nonlinear interaction canceling, the looper and tension interaction could be greatly reduced. A more systematic method: looper-tension almost disturbance decoupling control based on feedback linearization, is proposed and results in parameterized controllers which are easy to tune. Nonlinear controllers are global stable and can work within a wide range, whereas the linear controllers only can work within a limited range and usually needs to change according to the work point.(3) Most of the modern HSFMs are controlled based on PLC (Programmbale Logic Controller) which means discrete control strategies will be more adequate for application. For the strong interaction between the gauge system and looper-tension system, state space model of gauge-looper integrated system is propsed. Based on the integrated model, CDMPC (Constraint Discrete-time MPC) strategy is suggest to control the integrated system coordinately and systematicly, which is easy to deal with constraints on inputs and outputs of the intergral system, and convenient to adjust weight matrices to improve the performances of the overall system.(4) The reliability of the simulation program, which is built based on the state space model of the gauge-looper integrated system, is verified through simulations of disturbances of the integrated system with real plant data. And the effectiveness of the proposed CDMPC is thereby proved through numerical simulations based on the verified program.更多还原