【作者】 樊秀芬；

【导师】 付兴武；

【作者基本信息】 辽宁工程技术大学 ， 控制理论与控制工程， 2003， 硕士

【摘要】 本文回顾了PID控制器发展的历史及其参数整定方法。针对传统的PID参数整定方法在应用于高温力学实验机温度控制系统中存在的问题，本文将神经网络的自学习功能于PID控制相结合，提出了一种单神经元自适应PID，这种控制器实现了PID参数的在线自整定；另外为了克服单神经元PID控制器中控制器增益K不可以在线调整的缺陷，引入了PSD控制算法中增益K的调整方法，从而形成了单神经元自适应PSD控制算法。这种算法使K也具有了在线调整的能力，增强了控制器的适应性和鲁棒性。 单神经元PID和PSD控制算法在被控对象存在大时滞时会产生超调，甚至使系统失稳，为了解决这一问题，本文最后将Smith预估器引入了单神经元控制器中，从而很好地解决了时滞对系统的影响。 为了验证以上各种算法的控制效果，本文将其应用于高温力学实验机温度控制系统中，进行了仿真研究，仿真结果表明这类控制器结构简单，性能好，可望在控制领域得到广泛应用。更多还原

【Abstract】 In this paper, the development histroy of PID controller and the ripe methords of its parameters setting are reviewed.In order to solve the problem that the tradition methods of PID controller’ s parameters setting can’ t controll the temperature of the electrical furnance effectively, A neuro element PID controller with some adaptive ability is proposed .This controller realized the on-line self-tunning of parameters of PID controller. In addition, in order to overcome the shortcoming of neuro element PID controller,that is , the controller’ s gain K can’ t be tuned on-line , the tunning method of gain in PSD controller is introduced and the neuro element PSD controller is obtainted. The gain K can be tuned on-line in neuro element controlers, so it has stronger adaptability and robustness.In neuro-element PID and PSD controllers if the time delay of the controlled object , is too long, the overshoot of the system will be too high and system even will be unstable. In order to solve this problem, the neuro controller is combined with the Smith-predictor. Thus the influence of time-delay to control system is overcome effectively.These kinds of controll algorithom are applied in the temperature control system of electrical furnance and simulation research is done with Simulink. The simulation results show that neuro element PID and PSD controllers have good performance and can be used widely.更多还原