摘要
受到风速变化特性的影响,在进行排烟风机转速变桨距控制时,通常会因缺乏对最优桨距角的动态预测而导致控制精度不佳。对此,提出基于DenseNet算法的排烟风机转速变桨距非线性控制方法。综合考虑风机的气动特性、机械结构等因素,对排风烟机进行动力学建模,建立描述排烟风机在不同工况下的转速与变桨矩之间的非线性关系。构建DenseNet网络模型,以风速变化率作为输入特征,通过特征向量拼接方式和回归预测层对最优桨距角进行动态预测。采用非线性状态误差反馈控制律,对跟踪微分器(TD)和扩张状态观测器(ESO)的输出误差及其微分进行非线性组合,对变桨矩调整控制量进行计算。
Affected by the wind speed variation characteristics,the control accuracy of smoke exhaust fan speed pitch control is usually poor due to the lack of dynamic prediction of the optimal pitch angle.A nonlinear control of smoke exhaust fan speed with variable pitch based on DenseNet algorithm is proposed.Taking the aerodynamic characteristics,mechanical structure,and other factors of the fan into account,a dynamic model of the exhaust fan is developed to describe the nonlinear relationship between the speed and pitch of the exhaust fan under different operating conditions.By building a DenseNet network with the wind speed change rate as input feature,it dynamically predicts the optimal pitch angle through feature vector concatenation and regression prediction layer.Using a nonlinear state error feedback control law,the output errors and their derivatives of the tracking differentiator(TD)and the extended state observer(ESO)are nonlinearly combined to calculate the pitch adjustment control variable.
作者
李平
LI Ping(School of Civil Engineering,Wanjiang University of Technology,Maanshan Anhui 243000,China)
出处
《桂林航天工业学院学报》
2025年第3期415-422,共8页
Journal of Guilin University of Aerospace Technology
关键词
排烟风机
转速
变桨距
控制方法
DenseNet算法
控制精度
smoke exhaust fan
rotational speed
variable pitch
control method
DenseNet algorithm
control accuracy
作者简介
李平,女,安徽广德人。讲师,硕士。研究方向:建筑消防安全。
