基于模糊神经网络的加热炉温度控制方法研究

发布时间：2018-03-21 19:21

本文选题：加热炉　切入点：模糊神经网络　出处：《青岛科技大学》2015年硕士论文　论文类型：学位论文

【摘要】：原油作为一种重要的物质资源,对国家经济发展有着至关重要的影响,考虑到我国原油的特点,含蜡高、凝固点高且粘度高,需要不断地给原油进行加热,防止在输送原油的过程中发生沉淀或凝固。而加热炉是这一过程中的主要设备之一,如何设计加热炉的控制系统,对提高系统自动化控制水平,提高系统控制效率和缩短系统控制时间有着非常重要的作用。加热炉是一个非线性、强耦合、多变量以及多干扰的复杂控制系统,其主要包含三个控制方向：调整引风量与送风量,保证炉膛内压力,保证系统的安全运行；保证空燃比保持在最佳范围内,已达到较高的燃烧效率,保证系统控制过程的经济性；保证原油温度输出在正常范围内,保证原油的顺利输送。而这些要求利用传统的控制方法已不能满足,结合当代先进的控制方法,模糊神经网络结合了模糊控制及BP神经网络的优点,而模糊系统不仅不需要对控制系统建立较为精确的数学模型,而利用BP神经网络可以将实际输出与期望输出之间的误差反向传播,以此来修正系统的隶属度函数及权值,以更短的时间使系统输出达到稳定。本文介绍了模糊系统与BP神经网络的特点,以及两种控制方法的结合点及特点。考虑到单纯使用BP神经网络易陷入局部最小的问题,遗传算法的全局搜索功能恰巧能解决这一问题,用GA训练模糊神经网络可以避免陷入局部最小,所以选用GA-BP网络,即遗传算法与BP神经网络相结合对系统的初始数据进行离线优化。最后利用MATLAB对模糊神经网络控制下的系统进行仿真,与传统的PID控制方法相对比,在多种角度分析该种控制方法下系统控制的优越性。
[Abstract]:Crude oil, as an important material resource, has a vital influence on the development of national economy. Considering the characteristics of crude oil in China, high wax content, high freezing point and high viscosity, it is necessary to continuously heat the crude oil. The furnace is one of the main equipments in this process. How to design the control system of the furnace and improve the automation control level of the system, It is very important to improve the control efficiency and shorten the control time of the system. The heating furnace is a complex control system with nonlinear, strong coupling, multi-variable and multi-disturbance. It mainly includes three control directions: adjusting the air intake and air supply, ensuring the pressure in the furnace and ensuring the safe operation of the system, and ensuring that the air-fuel ratio is kept in the optimum range and has achieved higher combustion efficiency. To ensure the economy of the control process of the system, to ensure that the crude oil temperature output is within the normal range and to ensure the smooth transportation of crude oil, and that these requirements can no longer be met by using the traditional control methods, combined with the modern advanced control methods, Fuzzy neural network combines the advantages of fuzzy control and BP neural network, and fuzzy system not only need not establish a more accurate mathematical model of the control system. By using BP neural network, the error between actual output and expected output can be propagated back to correct the membership function and weight of the system. In this paper, the characteristics of fuzzy system and BP neural network are introduced, as well as the combination points and characteristics of the two control methods. The global search function of genetic algorithm can solve this problem. Using GA to train fuzzy neural network can avoid falling into local minimum, so GA-BP network is chosen. That is, genetic algorithm and BP neural network are combined to optimize the initial data of the system offline. Finally, MATLAB is used to simulate the system under the control of fuzzy neural network, which is compared with the traditional PID control method. The superiority of system control under this control method is analyzed in various angles.
【学位授予单位】：青岛科技大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TE963;TP273

【相似文献】