论文部分内容阅读
针对过热汽温被控对象的特点,设计了相应的基于神经网络的自抗扰控制器,分析了自抗扰控制器与神经网络结合的方式以及神经网络训练样本提取与训练方法,将该控制器成功应用到过热汽温控制系统中。嵌入神经网络的自抗扰控制器的状态估计能力提高,适应能力增强。研究表明,针对大迟延、大惯性、非线性且动态特性参数随工况变化的过热汽温对象,该控制器不仅能保证较高的控制精度,而且具有很强的抗干扰性和鲁棒性。
According to the characteristics of superheated steam temperature controlled object, the corresponding ADRC controller based on neural network is designed. The method of combining ADRC with neural network and the method of neural network training sample extraction and training are analyzed. The control The successful application of superheated steam temperature control system. The self-disturbance rejection controller with embedded neural network can improve the state estimation ability and adaptability. The research shows that the controller not only guarantees high control accuracy but also has strong anti-interference and robustness against overheated steam temperature objects with large delay, large inertia, nonlinear and dynamic parameters changing with operating conditions. .