现代板带轧机数学模型的研究与应用

发布时间：2017-12-31 11:38

本文关键词：现代板带轧机数学模型的研究与应用　出处：《北京科技大学》2016年博士论文　论文类型：学位论文

【摘要】：板带生产综合自动化控制系统是复杂工业过程控制最具代表性的系统。板带轧制生产中控制对象具有多变量、强耦合、非线性、大滞后和时变性等复杂性,对控制系统而言尤其以高精度、高速度为特征,是冶金企业过程控制系统中复杂程度最强、控制难度最高的综合自动化系统,其控制水平与规模在相当程度上代表和反映一个国家工业化、自动化、信息化的发展水平。相对于传统的板带轧机而言, “现代”板带轧机具有液压驱动的机械设备、交流调速的主辅电气传动、完备的三级计算机控制系统和齐全的检测仪表等特征,轧制过程必依赖于数学模型。为此,数学模型是现代板带轧机控制中的核心和关键技术。建立和应用高可用性、高精度的数学模型,是冶金企业能够持续开发和生产出满足需要的板带材品种,不断提高产品质量,保持稳定生产,节能降耗,降低成本,从而提升综合竞争力的技术基础。论文以板带轧机生产过程为对象,从理论建模和数据驱动的角度出发,研究板带轧机复杂的机理模型;研究以优化产品质量,提高生产效率为目标,充分利用工业过程的历史和实时数据,运用人工智能、统计及有限元分析等基于知识和数据驱动的建模方法。从而分析并建立现代板带轧机的关键数学模型,把数据资源的优势转化为生产效益和产品质量的优势。本文主要研究工作及创新点有以下4个方面：1)提出将遗传算法与有限元分析合理结合,建立轧辊温度场及热辊形模型,并分析轧制参数对温度场的影响,使热辊形对板形的影响最小化,为板形控制提供有效的技术支持。2)结合小波多分辨分析和多RBF神经网络的轧制力建模方法,将轧制力分解重构为对应于不同影响因素的子信号,建立一个多RBF神经网络模型。仿真结果表明该模型可以真实反映轧制力变化的内在机理,而且网络训练时间短、预测精度高,是轧制力建模的一种有效方法。3)对弯辊力设定中的关键参数——工作辊相对挠度进行详细推导,并采用显式动力学弹塑性有限元法,建立四辊轧机弯辊力模型,研究并分析液压弯辊力对轧制过程中板形参数以及工作辊参数的影响规律,为进一步建立和优化板形控制模型奠定基础。4)推导以勒让德多项式为基模式的板形识别模型,并运用真实过程数据仿真验证其优越性。该模型计算量小,多项式系数直接对应于板形控制的三种基本手段,物理意义明确,工程实施灵活快捷。研究成果已成功应用在某铝箔厂1850mm冷轧生产线中,取得良好的控制效果。论文的研究进一步丰富了基于知识与数据驱动的复杂工业过程的优化建模理论与方法,以信息化带动工业化,实现现代板带轧机的优化运行与控制。
[Abstract]:Strip production of integrated automation control system is a complex industrial process control system. The most representative strip rolling production control object with multi variable, strong coupling, nonlinear, large delay and time-varying complexity, especially the control system with high precision, high speed characteristic, is the complexity of process control in metallurgical enterprises in the strongest, the difficulty to control the integrated automation system is the highest, the control level and scale to represent and reflect the industrialization of a country in a considerable degree of automation, the development level of informatization. Compared with the traditional rolling mill, the "modern" strip mill with mechanical equipment hydraulic drive, electric drive and auxiliary AC drive complete, three level computer control system and complete testing instruments and other characteristics, the rolling process will depend on the mathematical model. Therefore, the mathematical model of modern rolling mill control The core and key technology. The establishment and application of high availability, high accuracy mathematical model, is the metallurgical enterprises continue to develop and produce strip varieties need, and continuously improve product quality, stable production, energy saving, cost reduction, so as to enhance the comprehensive competitiveness of the basic technology. Based on the strip production the process of rolling mill as the research object, theoretical modeling and data driven from the perspective of research, mechanism model of strip mill complex; research to optimize product quality, improve the production efficiency as the goal, to make full use of the industrial process of historical and real-time data, using artificial intelligence, statistical analysis and finite element modeling method based on data-driven and knowledge to analysis and establish the mathematical model of the key of modern rolling mill, the data resources advantage into the production efficiency and product quality advantages. The main research work The following 4 aspects of innovation: 1) is put forward based on genetic algorithm and finite element analysis to establish a reasonable combination of roll temperature field and thermal contour model, and analyze the effect of rolling parameters on the temperature field, the heat roller shape to minimize the impact on the shape and shape control to provide effective technical support for the.2 board with the rolling force) modeling method of wavelet multiresolution analysis and RBF neural network, the decomposition and reconstruction for rolling force corresponding to the different influencing factors of sub signal, establish a multi RBF neural network model. The simulation results show that the model can reflect the inherent mechanism of the rolling force changes, and the network training time is short, the prediction accuracy high.3 is an effective method for rolling force modeling) detailed derivation of the key parameters of relative deflection of work roll bending force setting in, and by using the dynamic explicit elastic-plastic finite element method, the establishment of four roller mill roll bending Force model, research and analyze the influence of shape parameters in the rolling process of plate and the work roll parameters of hydraulic roll bending force, to further establish and optimize the shape control model is derived by Legendre.4) lay the foundation for flatness recognition model based on polynomial model, and use real process data to verify the superiority of the simulation. Model computation, polynomial coefficients directly corresponding to the three basic means to shape control, clear physical meaning, project implementation, flexible and efficient. The research results have been successfully applied in a 1850mm aluminum foil factory cold rolling production line, achieve good control effect. The paper further enriches the theory and method of optimization modeling of complex industrial process knowledge and based on data driven, drive industrialization with informatization, optimizing operation and control of modern rolling mill.

【学位授予单位】：北京科技大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：TG333;O174.2;TP183

【相似文献】