基于神经网络辨识的煤泥滚筒干燥控制研究

发布时间：2018-03-09 23:35

本文选题：煤泥　切入点：滚筒干燥机　出处：《昆明理工大学》2015年硕士论文　论文类型：学位论文

【摘要】：随着煤炭行业产量和洗选比例的提高,煤泥产量也相应地会增加,要实现煤泥的大规模利用,就不得不对煤泥的含水率以及品质提出更高的要求和规范,除了其它一些办法,如制水煤浆、掺烧等以外,将煤泥进行滚筒干燥后再利用也是一个值得考虑的方向,这样不仅在经济上能取得更高的回报,而且对于资源的回收利用以及生态保护都有着长远的现实意义和历史意义。滚筒干燥是目前应用最为广泛的,但是,滚筒干燥的出口温度控制是一个非常复杂的时变过程,目前,因为滚筒干燥的内在结构,随机扰动对控制系统的影响,导致在理论上并没有精确数学模型能取得较好控制效果。因此,作为滚筒干燥控制系统中出口温度的自动化控制水平一直得不到明显的提高。本文主要对前馈以及前馈-负反馈的控制原理作了简要介绍,分析了RBF神经网络辩识的模型参考白适应控制,在MATLAB环境下进行仿真,可以看出基于RBF神经网络辨识的滚筒干燥机出口温度的单神经元模型参考自适应控制系统对出口温度控制模型跟踪吻合较好,与实际输出相差较小且误差调整时间小,系统响应时间短且控制过程比较稳定。通过飞升曲线法得到了滚筒干燥机出口温度的传递函数。设计了单神经元神经网络控制器和RBF神经网络控制器,并将传统的PID控制和上述两种控制单独进行了仿真。并分别两两做了比较并进行分析,进而设计了基于RBF神经网络辨识的自适应控制系统,仿真结果表明,该控制系统在模型的辨识以及跟踪方面具有更好的效果,而且还有控制误差小、鲁棒性较好、系统响应快等特点。
[Abstract]:With the increase of coal industry output and washing ratio, the output of coal slime will increase accordingly. In order to realize the large-scale utilization of coal slime, it is necessary to put forward higher requirements and norms on the moisture content and quality of coal slime, among other ways, For example, in addition to coal water slurry, mixed burning and so on, reusing the slime after drying the drum is also a direction worthy of consideration, so that it can not only achieve higher economic returns, And it has a long-term practical and historical significance for resource recycling and ecological protection. Drum drying is the most widely used at present, but the outlet temperature control of drum drying is a very complicated and time-varying process. At present, because of the inner structure of drum drying and the influence of random disturbance on the control system, there is no accurate mathematical model to obtain better control effect in theory. As the automatic control level of outlet temperature in drum drying control system, the control principle of feedforward and feed-negative feedback is introduced briefly in this paper. The model reference white adaptive control of RBF neural network identification is analyzed, and the simulation is carried out under the MATLAB environment. It can be seen that the single neuron model reference adaptive control system based on RBF neural network identification for the outlet temperature of the drum dryer is in good agreement with the outlet temperature control model, and the difference with the actual output is small and the error adjustment time is small. The response time of the system is short and the control process is relatively stable. The transfer function of the outlet temperature of the drum dryer is obtained by the flying curve method. A single neuron neural network controller and a RBF neural network controller are designed. The traditional PID control and the above two kinds of control are separately simulated and compared and analyzed, and then the adaptive control system based on RBF neural network identification is designed. The simulation results show that, The control system has better performance in model identification and tracking, and has the advantages of small control error, good robustness and fast response.
【学位授予单位】：昆明理工大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TD849

【参考文献】