【作者】 李冬；

【导师】 刘建昌；

【作者基本信息】 东北大学 ， 控制理论与控制工程， 2015， 博士

【摘要】 厚度精度是热轧板带钢的重要质量指标之一,它直接影响下游制造业的生产效率及产品质量。随着科学技术的日趋发展,用户对板带钢的厚度精度提出了越来越高的要求。厚度自动控制(Automatic gauge control,AGC)系统的主要作用就是将板带厚度控制在设定的目标偏差范围内,其性能的好坏直接影响板带厚度精度的高低。因此,本文对板带热连轧机的厚度自动控制系统进行了全面、深入的研究,主要研究内容如下:(1)新型监控AGC及其改进研究。为了缩短监控AGC的调节周期以提高其厚差修正速度,深入研究了一种新型监控AGC系统,并对其存在的问题进行了分析。为了保持各机架轧制力分配比例不变以利于轧制过程稳定和板形质量控制,提出了新型监控AGC厚差分配系数的改进算法,较好地解决了厚差分配系数合理计算的问题。为了获得最佳的综合厚度控制效果,对新型监控AGC的控制结构进行了改进,有效地解决了新型监控AGC与厚度计AGC产生相互干扰的问题。(2)轧辊偏心补偿方法研究。为了避免轧辊偏心引起厚度计AGC误动作,研究了一种轧辊偏心动态死区滤波方法,较好地解决了死区宽度固定无法适应轧辊偏心状态变化的问题。动态死区滤波方法通过抑制厚度计AGC误动作能够防止轧辊偏心造成厚度精度恶化,但无法消除轧辊偏心对板带厚度造成的不良影响。为了进一步提高板带厚度精度,提出了基于群智能辨识的轧辊偏心补偿方法,丰富了主动轧辊偏心补偿方法。通过融合蚁群算法和粒子群算法设计出的群智能算法,较好地解决了线性递减惯性权重的粒子群算法由于种群多样性缺失过快造成的易陷入局部最优的问题。(3)热连轧负荷分配多目标优化方法研究。针对热连轧负荷分配这一多目标优化问题,设计了基于改进蚁群系统算法的负荷分配多目标优化方法,提高了负荷分配优化的求解速度和精度。为了避免解空间网格数量难以确定的问题影响改进蚁群系统算法的优化性能,提出了用于求解负荷分配多目标优化问题的多目标平衡迭代算法。此算法的优化求解速度和精度均优于改进蚁群系统算法,并且还具有对初始值要求低、原理及参数设置简单等特点,体现出良好的在线应用前景。(4)厚度自动控制系统设计。设计了某钢厂650mm板带热连轧机的厚度自动控制系统,给出了该系统实际使用的控制算法。为了提高末机架厚度计AGC的控制精度,设计了基于轧制数据的轧机刚度测量方法,提高了轧机刚度测量的准确性。基于现场实验数据,对确定轧机辊缝参考零点的效果、轧机刚度的测量效果、负荷分配的优化设定效果、实际板带厚度的控制效果和基于群智能辨识的轧辊偏心补偿效果进行了分析和评价。本文的研究结果对我国板带热连轧厚度控制水平的提高具有积极的促进作用。更多还原

【Abstract】 The thickness accuracy is one of the important quality indexes in hot rolled strips.It directly affects the production efficiency and product quality of the downstream manufacturing industry.With the development of science and technology,the requirement of the strip thickness accuracy from users is getting higher and higher.The main function of the automatic gauge control(AGC)system is to control the strip thickness within the range of setting target deviation.Its performance directly affects the precision of strip thickness.Therefore,a comprehensive and in-depth research was carried out on AGC system for hot strip mill in this dissertation.The main research contents are as follows:(1)The new monitor AGC and its improvements were researched.To shorten adjustment cycle of monitor AGC for improving its correction speed of thickness deviation,the intensive study on a new type of monitor AGC system was conducted and its existing problems were analyzed.In order to keep constant distribution ratio of the rolling force on each stand in favor of the stability of rolling process and the control of shape quality,an improved algorithm of thickness-deviation distribution coefficients for the new monitor AGC was proposed,and then the problem of reasonable calculation of thickness-deviation distribution coefficients was well solved.For the sake of obtaining the best comprehensive thickness control effect,the control structure of the new monitor AGC was improved,and the problem of mutual interference between the new monitor AGC and gauge meter AGC(GM-AGC)was solved efficiently.(2)Compensation methods of roll eccentricity were investigated.In order to avoid the misoperation of GM-AGC caused by roll eccentricity,the research on dynamic-deadband filter method of roll eccentricity was done.Then,the problem that the fixed deadband width could not accommodate the variation of roll eccentricity was solved preferably.The dynamic-deadband filter method could prevent the deterioration of thickness accuracy causing by roll eccentricity through suppressing the malfunction of GM-AGC,but it could not eliminate the harmful effect of roll eccentricity on strip thickness.To further improve the accuracy of the strip thickness,a method of roll eccentricity compensation based on swarm intelligence identification was proposed and the active roll eccentricity compensation method was enriched.The swarm intelligence algorithm based on the fusion of ant colony optimization algorithm and particle swarm optimization(PSO)algorithm was designed to preferably solve the problem that PSO algorithm with linear decreasing inertia weight was easy to fall into local optimum due to the rapid loss of population diversity.(3)The multi-objective optimization methods of load distribution for hot continuous rolling were studied.To solve the multi-objective optimization problem in load distribution of hot continuous rolling,multi-objective optimization method for load distribution based on improved ant-colony system(IACS)algorithm was designed.This method improved the solving speed and accuracy of load distribution optimization.To avoid the problem of the difficulty in determining grid number of solution space affecting the optimization performance of IACS algorithm,a multi-objective balance iteration algorithm for solving the problem of the multi-objective optimization of load distribution was proposed.Its optimization solution speed and accuracy were both better than that of IACS algorithm.Moreover,the multi-objective balance iteration algorithm had the features,such as low requirement for initial value,uncomplicated principle and simple parameter setting and so on,and it also had the good prospect for online application.(4)The AGC system was designed.The AGC system for the 650mm hot strip mill in a steel plant was designed and its control algorithms for practical use were given.To improve the control precision of GM-AGC in the last stand,the measurement method of mill stiffness based on rolling data was designed,and the accuracy of mill stiffness measurement was improved.According to the field experimental data,the effects of experimental on the determination of the reference zero of mill roll gap,the measurement of the mill stiffness,multi-objective optimization of load distribution,the control of actual strip thickness and the roll eccentricity compensation based on swarm intelligence identification were analyzed and evaluated.The research results of this dissertation have a positive effect on the improvement of thickness control level for hot strip mill in China.更多还原