论文部分内容阅读
大纯时延、煤种多变、蒸汽负荷频繁变化是链条炉难以进行良好燃烧控制的原因。本文作者提出了改变神经网络输入样本区间,利用网络输出期望值与输出实际值之间的误差平方和产生的突变,辨识出非线性对象的延迟时间的方法,将神经网络大延迟系统的辨识与基于模型预测的神经网络控制策略相结合,可用于对具有变化参数或不确定性延迟时间的非线性大延迟系统的控制,同时,以10t/h链条炉作为研究对象进行仿真,仿真结果表明这种神经网络模型对非线性大纯时延系统的控制具有控制速度快,鲁棒性能好等优点。
Large pure delay, coal variety, frequent changes in steam load chain furnace is difficult to control the reasons for good combustion. In this paper, the author puts forward a method to change the input sample interval of neural network and make use of the squared error sum of network output expectation value and output actual value to identify the delay time of nonlinear object. The method of identifying and identifying the large delay system based on neural network Model predictive neural network control strategy can be used to control nonlinear large delay system with varying parameters or uncertain delay time. At the same time, the 10t / h chain furnace is used as the research object to simulate. The simulation results show that this The neural network model has the advantages of fast control and good robustness for the control of nonlinear large pure delay system.