智能预测控制在磨机自动控制系统中的应用

发布时间：2018-08-23 08:32

【摘要】：磨机是水泥粉磨系统中重要的设备之一,磨机负荷的稳定直接影响磨机的粉磨效率和水泥的成品质量,因此,对于磨机负荷的研究具有重大的意义。随着自动化水平的不断提高,对于磨机自动控制系统的模型研究和优化控制的要求将越来越高。由于磨机是一个具有非线性、纯滞后、强耦合等特点复杂工业被控对象,采用常规的控制算法难以取得良好的控制效果。本文以某水泥厂磨机自动控制系统作为研究对象,在了解水泥粉磨系统国内外研究现状的基础上,提出了采用智能预测控制对磨机负荷进行优化控制。根据联合粉磨工艺流程,选取了选粉机转速作为系统输入,出磨提升机电流作为系统输出,并采取了基于神经网路的预测控制策略对磨机负荷进行优化控制。在此基础上,开发了磨机自动控制系统的上位机软件。本文主要工作如下:(1)利用三层前向BP神经网络辨识磨机负荷模型并进行预测控制器设计。首先采集现场的输入输出数据,辨识系统模型的输入输出延时阶数;然后用辨识出来的神经网络模型作为预测模型,运用N-R滚动优化方法计算系统未来控制序列,反馈校正用来克服扰动带来的模型预测误差;最后,通过Matlab仿真验证了该算法在水泥磨机自动控制系统的有效性。(2)运用了VC++中MFC和Matlab混合编程的方式开发预测控制软件。通过Matlab Compiler方法把用m文件编写的算法程序编译成VC++的DLL动态链接库文件,在VC++中调用DLL文件实现算法模块的功能;运用OPC技术实现数据的读取和写入的功能;最后,为了实现动态效果的显示,在MFC程序中导入TeeChart5控件,用来实时的显示出磨提升机电流的变化趋势以实现用户界面的设计。(3)为了工程的实际应用,验证了优化控制软件的有效性。还原现场运行工况,运行软件,通过观察实时曲线来验证软件的优化控制效果。效果显示本文开发的优化控制软件能够很好的调节出磨提升机电流到达设定值,使得磨机负荷处于稳定状态。
[Abstract]:Grinding machine is one of the important equipment in cement grinding system. The stability of mill load directly affects the grinding efficiency of grinding machine and the quality of cement finished product. Therefore, it is of great significance to study the mill load. With the improvement of automation level, the requirements of model research and optimization control of mill automatic control system will be higher and higher. Because the mill is a complex industrial controlled object with nonlinear, pure lag and strong coupling, it is difficult to obtain good control effect by using conventional control algorithm. In this paper, the automatic control system of a cement mill is taken as the research object. On the basis of understanding the present research situation of cement grinding system at home and abroad, the intelligent predictive control is proposed to optimize the mill load control. According to the combined grinding process, the speed of the separator is selected as the input of the system, the current of the output hoist is taken as the output of the system, and the predictive control strategy based on neural network is adopted to optimize the load of the mill. On this basis, the upper computer software of the mill automatic control system is developed. The main work of this paper is as follows: (1) Identification of mill load model and design of predictive controller using three-layer forward BP neural network. First, the input and output data are collected to identify the order of the input and output delay of the system model, and then the neural network model is used as the prediction model, and the N-R rolling optimization method is used to calculate the future control sequence of the system. Feedback correction is used to overcome the model prediction error caused by disturbance. Finally, the effectiveness of the algorithm in the cement mill automatic control system is verified by Matlab simulation. (2) the predictive control software is developed by using the mixed programming of MFC and Matlab in VC. The algorithm program written in m file is compiled into VC's DLL dynamic link library file by Matlab Compiler method, and the function of algorithm module is realized by calling DLL file in VC, and the function of data reading and writing is realized by OPC technology. In order to display the dynamic effect, the TeeChart5 control is imported into the MFC program to display the changing trend of the grinding hoist current in real time in order to realize the design of the user interface. (3) for the practical application of the project, the effectiveness of the optimized control software is verified. The optimal control effect of the software is verified by observing the real-time curve. The results show that the optimal control software developed in this paper can adjust the current of the mill hoist to the set value and make the mill load stable.
【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】