【作者】 董春霞；

【导师】 安连祥； 王书田；

【作者基本信息】 河北工业大学 ， 控制理论与控制工程， 2004， 硕士

【摘要】 本文从实际出发，以唐山钢铁集团有限公司高线厂一车间蓄热室加热炉为研究对象，对受随机因素干扰的、具有大惯性、纯滞后的非线性分布参量的随机过程进行了研究。文中着重研究了神经网络与模糊系统和PID控制融合的可行性及融合方式，提出了一种复合式控制方案——神经网络模糊PID控制算法，并将此算法与PID控制、模糊自整定PID控制进行了仿真比较。本文的主要研究内容有如下几个方面： 1．研究了唐山钢铁公司高线厂的蓄热式加热炉的燃烧情况，得出了结论，即其过程为大惯性、纯滞后、受随机干扰的非线性分布参量过程，为后继的研究作了准备。 2．综述了PID控制、模糊控制、神经网络控制的发展和研究现状，并就这三种控制的研究提出了一些观点。 3．介绍了参量模糊化方法及基本辨识方法。 4．针对PID控制、模糊控制和神经网络控制的特点，提出了神经网络模糊PID控制算法。 5．介绍了MATLAB及Simulink仿真的应用，在MATLAB中对本文提出的神经网络模糊PID控制算法进行仿真试验，并与PID控制算法、模糊自整定PID控制算法进行仿真比较。验证本文所提出的算法在动、静态特性、鲁棒性和抗干扰性。结果表明，本文所提出的算法不仅获得强鲁棒性、强抗干扰性和满意的控制性能，而且系统设计方法简单，能方便地用于实际工业控制中。更多还原

【Abstract】 In view of practical use, taking the regenerative furnace of the first plant of high-speed-wire factory of Tangshan steel company as object, the thesis studies the casual process of nonlinear and distributed parameter with big inertia,pure lag and casual disturbance. The feasibility and the form of combining neural network and fuzzy system with PID control are discussed. The theory of compound control-fuzzy-PID control based on neutral network is put forward. Comparison among the compound control, the PID control and the fuzzy self-tuning PID control is made in simulation. The main research work and contribution of this dissertation are as follows:1. A study of combustion of the furnace in high-speed-wire factory of Tangshan steel company is made, from which the conclusion is acquired, that is the process is a casual process of nonlinear and distributed parameter with big inertia, pure lag and casual disturbance. This makes prepare for further studies.2. A survey of PID control, fuzzy control and neutral network control is made, and the problems existing in neutral network control and fuzzy control, PID control are also discussed.3. An introduction of the approach of parameter fuzzification and basic identification methods is followed.4. The fuzzy-PID control arithmetic based on neutral network is proposed, according to characteristics of neutral network control and fuzzy system and PID system.5. MATLAB and Simulink software are introduced at last. The simulation experiments are made for the arithmetic of fuzzy-PID based on neutral network in MATLAB. Comparisons are made among the arithmetic, PID control and fuzzy self-tuning PID control. It is showed that satisfied performance, strong robustness and disturbance-rejection are obtained by using the arithmetic put forward in this thesis. In addition, it is simple for system designing and convenient to use in industrial control.更多还原