半终粉磨系统建模及数据驱动控制研究
发布时间：2018-11-06 07:27
【摘要】：水泥粉磨是水泥生产的最后一个环节,直接决定了水泥生产线最终的产量及质量。近年来,半终粉磨工艺的应用在一定程度上增加了水泥产量,降低了能源消耗;在线激光粒度分析仪在水泥粉磨环节上的应用,显现了其在提质、节能、降耗和增产方面的巨大潜力。本文以在线激光粒度分析仪为检测手段,结合半终粉磨工艺机理,围绕水泥粉磨环节的磨机负荷和水泥粒度,开展半终粉磨系统建模及数据驱动控制研究。本文研究内容为山东省重大专项“智能化工厂关键技术研究与应用示范(2015ZDXX010F01)”和国际合作项目“面向节能减排的水泥生产过程集成控制系统研究(SQ2013ZOC600)”两个项目的核心内容之一,主要研究工作如下:(1)针对水泥半终粉磨系统的磨机负荷和水泥粒度两个关键参数,在分析其影响因素的基础上,建立了其各自数学模型。选取循环风机转速和磨机电流分别作为神经网络的输入和输出参量,采用极限学习机神经网络(ELM)算法建立了磨机负荷的数学模型,其输入层与隐含层的权值和隐含层神经元的阈值在辨识过程中随机产生且保持不变,隐含层神经元的数量被确定之后便可得到唯一最优解,并进行了仿真验证;选取后选粉机转速为模型输入,小于45?m颗粒含量为模型输出,采用最小二乘法建立了水泥粒度的数学模型,仿真验证显现了该模型与水泥粒度的动态变化具有良好的一致性,为后续水泥粒度控制算法的研究奠定了基础。(2)为提高半终粉磨粒度控制的稳定性和鲁棒性,给出了一种基于数据驱动技术的水泥粒度自适应PID控制方法,解决了控制方法对模型的依赖问题。在建立水泥粒度系统紧格式动态线性化数据模型的基础上,应用粒度控制系统的I/O数据(后选粉机转速和小于45?m颗粒含量)估计其伪偏导数(PPD),调整PID控制器的参数;仿真结果验证了该控制算法的有效性。(3)提出了包括系统软硬件架构、数据库设计、Bang-Bang与数据驱动自适应PID控制相结合的半终粉磨粒度优化控制方案,研发了半终粉磨粒度优化控制系统,进行了工程应用,取得良好运行效果。
[Abstract]:Cement grinding is the last link in cement production, which directly determines the final output and quality of cement production line. In recent years, the application of semi-finished grinding technology has increased cement production to a certain extent and reduced energy consumption. The application of on-line laser particle size analyzer in cement grinding shows its great potential in improving quality, saving energy, reducing consumption and increasing production. In this paper, based on on-line laser particle size analyzer and combined with the mechanism of semi-finish grinding, the modeling and data-driven control of semi-finished grinding system are studied around the mill load and cement particle size of cement grinding link. The research contents of this paper are "key Technology Research and Application demonstration (2015ZDXX010F01) of Intelligent Factory" and "Research on Integrated Control system of cement production process oriented to Energy Saving and Emission reduction (SQ2013ZOC600)" and International Cooperation Project "Research on cement production process Integrated Control system for Energy Saving and Emission reduction (SQ2013ZOC600)" One of the core elements of the project, The main research work is as follows: (1) aiming at the two key parameters of grinding machine load and cement particle size of the cement semi-finished grinding system, based on the analysis of its influencing factors, the respective mathematical models are established. The rotational speed of circulating fan and the current of mill are selected as the input and output parameters of neural network, and the mathematical model of mill load is established by using the (ELM) algorithm of extreme learning machine neural network. The weights of the input layer and the hidden layer and the threshold value of the neuron in the hidden layer are generated randomly and remain unchanged during the identification process. The number of neurons in the hidden layer is determined and the unique optimal solution is obtained. Selecting the rotational speed of the separator as the model input and the particle content less than 45 m as the model output, the mathematical model of cement particle size is established by using the least square method. The simulation results show that the model is in good agreement with the dynamic change of cement particle size, which lays a foundation for the subsequent research of cement granularity control algorithm. (2) to improve the stability and robustness of semi-finished grinding granularity control. An adaptive PID control method for cement granularity based on data-driven technology is presented, and the dependence of the control method on the model is solved. Based on the dynamic linearization data model of cement granularity system compact format, the pseudo-partial derivative (PPD), is estimated by using the I / O data of the granularity control system (the speed of the later separator and the particle content less than 45m). Adjust the parameters of PID controller; Simulation results verify the effectiveness of the control algorithm. (3) A semi-finish-grinding granularity optimization control scheme, which includes system hardware and software architecture, database design, Bang-Bang and data-driven adaptive PID control, is proposed. The particle size optimization control system of semi-finish grinding has been developed and applied in engineering, and good operation effect has been obtained.
【学位授予单位】：济南大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TQ172.632;TP273
本文编号：2313555
[Abstract]:Cement grinding is the last link in cement production, which directly determines the final output and quality of cement production line. In recent years, the application of semi-finished grinding technology has increased cement production to a certain extent and reduced energy consumption. The application of on-line laser particle size analyzer in cement grinding shows its great potential in improving quality, saving energy, reducing consumption and increasing production. In this paper, based on on-line laser particle size analyzer and combined with the mechanism of semi-finish grinding, the modeling and data-driven control of semi-finished grinding system are studied around the mill load and cement particle size of cement grinding link. The research contents of this paper are "key Technology Research and Application demonstration (2015ZDXX010F01) of Intelligent Factory" and "Research on Integrated Control system of cement production process oriented to Energy Saving and Emission reduction (SQ2013ZOC600)" and International Cooperation Project "Research on cement production process Integrated Control system for Energy Saving and Emission reduction (SQ2013ZOC600)" One of the core elements of the project, The main research work is as follows: (1) aiming at the two key parameters of grinding machine load and cement particle size of the cement semi-finished grinding system, based on the analysis of its influencing factors, the respective mathematical models are established. The rotational speed of circulating fan and the current of mill are selected as the input and output parameters of neural network, and the mathematical model of mill load is established by using the (ELM) algorithm of extreme learning machine neural network. The weights of the input layer and the hidden layer and the threshold value of the neuron in the hidden layer are generated randomly and remain unchanged during the identification process. The number of neurons in the hidden layer is determined and the unique optimal solution is obtained. Selecting the rotational speed of the separator as the model input and the particle content less than 45 m as the model output, the mathematical model of cement particle size is established by using the least square method. The simulation results show that the model is in good agreement with the dynamic change of cement particle size, which lays a foundation for the subsequent research of cement granularity control algorithm. (2) to improve the stability and robustness of semi-finished grinding granularity control. An adaptive PID control method for cement granularity based on data-driven technology is presented, and the dependence of the control method on the model is solved. Based on the dynamic linearization data model of cement granularity system compact format, the pseudo-partial derivative (PPD), is estimated by using the I / O data of the granularity control system (the speed of the later separator and the particle content less than 45m). Adjust the parameters of PID controller; Simulation results verify the effectiveness of the control algorithm. (3) A semi-finish-grinding granularity optimization control scheme, which includes system hardware and software architecture, database design, Bang-Bang and data-driven adaptive PID control, is proposed. The particle size optimization control system of semi-finish grinding has been developed and applied in engineering, and good operation effect has been obtained.
【学位授予单位】：济南大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TQ172.632;TP273
【参考文献】
相关期刊论文 前10条
1 吴茂胜;袁铸钢;张强;;基于水泥粒度工况模板的联合粉磨系统建模[J];控制工程;2016年09期
2 石国平;徐向升;;双圈流水泥粉磨系统与半终水泥粉磨系统之比较[J];中国水泥;2016年06期
3 代桃桃;张强;申涛;;水泥磨机负荷的LPV预测控制[J];济南大学学报(自然科学版);2016年02期
4 田力;袁东;杨国春;;水泥工业粉磨系统的节电方法[J];新世纪水泥导报;2015年01期
5 张先垒;袁铸钢;张强;;基于Bang-Bang的水泥立式辊压磨模糊PID控制[J];济南大学学报(自然科学版);2015年02期
6 孙毅刚;刘静雅;赵珍;;基于极限学习机的航空发动机传感器故障诊断[J];传感器与微系统;2014年08期
7 刘仁德;;浅析影响闭路水泥粉磨产能的主要因素[J];四川水泥;2014年02期
8 刘平成;;水泥半终粉磨关键技术综述[J];四川水泥;2014年01期
9 汤健;柴天佑;余文;赵立杰;;在线KPLS建模方法及在磨机负荷参数集成建模中的应用[J];自动化学报;2013年05期
10 颜文俊;秦伟;;水泥立磨流程的建模和控制优化[J];控制工程;2012年06期
相关硕士学位论文 前10条
1 吴茂胜;水泥联合粉磨系统的建模与预测控制研究[D];济南大学;2016年
2 刘亚东;水泥联合粉磨粒度控制系统研发[D];济南大学;2016年
3 张先垒;水泥生产联合粉磨标准工况模板及建模研究[D];济南大学;2015年
4 郑鉴君;联合粉磨系统磨机负荷辨识方法研究[D];济南大学;2015年
5 任万杰;水泥粉磨系统的建模与控制方法研究[D];济南大学;2013年
6 王陆;数据驱动的预测控制器设计与性能监控[D];上海交通大学;2013年
7 曲艳召;水泥细度与碱硫含量对混凝土强度发展的影响[D];重庆大学;2012年
8 李克雷;水泥生产过程分解炉智能控制系统的设计与开发[D];东北大学;2010年
9 周扬铭;水泥生产关键设备及其节能降耗技术研究[D];武汉理工大学;2010年
10 范文礼;水泥粉磨技术的研究与应用[D];山东大学;2008年
，本文编号：2313555
本文链接：https://www.wllwen.com/kejilunwen/huaxuehuagong/2313555.html
