《单神经元PID控制器设计毕业论文》.docVIP

单神经元PID控制器设计 摘 要 常规PID控制器具有结构简单、易于实现、鲁棒性强等优点，但实际的生产过程中，控制对象一般都具有延迟大、非线性强、干扰大的特点。因此当工况改变时，对象的动态特性也发生改变，控制品质就会随之下降，所以采用常规PID控制器很难获得令人满意的控制效果。 神经网络具有强鲁棒性、容错性、并行处理、自学习、逼近非线性关系等特点，在解决非线性和不确定系统控制方面有很大潜力，近年来已广泛应用于工业过程控制领域。由于单神经元模型具有自适应和自学习的能力，使得它可以作为一种很好的方法而得以应用，因此本文将单神经元模型与常规PID控制器相结合，形成了具有自适应能力的单神经元PID控制器。本文讨论了单神经元自适应PID控制器和多变量单神经元PID控制器的结构，控制算法，并MATLAB仿真软件给出了实例仿真。 MATLAB仿真结果表明，该控制系统既保持了常规PID控制的优点，又有自学习的智能特性，因而具有良好的控制品质和较强的自适应能力。 关键词：PID控制器；数学模型；自适应控制；单神经元；MATLAB仿真；多变量 The Design of Single Neuron PID Controller Abstract Conventional PID controller is simple in structure, easy to implement, robust and other advantages. However, in the actual production process control targets have delayed the general, non-linear strong, and the heavy characteristics of the disturbance, so when the situation changes, the object of dynamic change, quality control will be declined. Therefore, the conventional PID control method is difficult to obtain satisfactory performance. Neural network has stronger robust, fault-tolerant, parallel processing, self-learning, approaching the characteristics of non-linear relationship, and uncertainty in solving nonlinear control system there is great potential, in recent years has been widely used in controlled areas. In single-neuron model is self-adaptive and self-learning ability , it can be regarded as an effective intelligent way for application ,so this article will be use single-neuron model with the conventional PID controller combine to form the adaptive capacity of a single-neuron PID controller. This paper discusses the structure and?the control algorithm of the single neuron adaptive PID controller and the variable single neuron PID controller,besides,giving the simulation examples by ?MATLAB simulation software. The MATLAB simulation results indicated that the control system both maintained the conventional cascade PID control merit, and has from the self-learning intelligent characteristic, thus has the good control quality and the